Skip to content

quidproquo

Tech, climbing, surfing, coffee, and everything else.

Quid pro quo /ˌkwɪd proʊ ˈkwoʊ/ — Latin for "something for something," a fair exchange.

This is where I document AI, tech, and product thinking, along with the process of building products — plus climbing, surfing, and coffee. Not just hoarding knowledge, but turning it into something useful for others, and then giving a little more.

Why the name →
置頂
tech guide

Conversation as Documentation: Turning Debug Sessions into Blog Posts with Claude Code

After finishing a debug session, just say 'write this up as a post' — Claude Code extracts content from the conversation, applies a template, generates frontmatter, and commits it to the repo. No extra writing required.

置頂
product project

Building a Low-Friction Blog from Scratch with Astro + Cloudflare Workers

To consolidate scattered notes and showcase diverse interests, I built a personal blog using Astro + Cloudflare Workers D1, paired with a Claude post skill for zero-friction writing.

置頂
tech guide

What Tools Power This Blog

Astro + the full Cloudflare suite — static-first, edge-computed, zero maintenance cost

Agentic Parsing: Letting Agents Decide How to Parse Documents

Traditional document parsing runs a fixed pipeline regardless of input, but contracts, financial reports, and technical manuals each need different strategies. Agentic Parsing lets LLM agents observe a document and dynamically choose tools — AgenticOCR parses only the regions that matter (70%+ visual token savings), and ParseBench shows even the best method scores only 84.9% across 2,000 enterprise pages. No silver bullet.

ColPali: Skip OCR, Retrieve Documents Directly from Images

ColPali renders each PDF page as an image, generates patch-level multi-vector embeddings with a vision-language model, and retrieves via MaxSim late interaction. On table-heavy financial PDFs, recall jumps from 62% to 84% — no OCR, no chunking. The tradeoff: ~100× storage, GPU required, no BM25.

Hierarchical Chunking + Auto-Merge: Small Chunks Search Well, Big Chunks Read Well

Small chunks give precise embeddings but lack context; big chunks have complete context but diluted embeddings. Hierarchical Chunking builds multi-level indexes (2048→512→128 tokens) with an Auto-Merge algorithm: leaf nodes match precisely, and when hit density exceeds a threshold the parent node is returned to the LLM instead. HiChunk shows a 12.7% evidence recall improvement; LlamaIndex and Haystack have it built in.

ai deep-dive

LLM Dev Workflow Landscape: When Verification Becomes the Bottleneck

AI boosted task output by 34%, but code review time surged 441% and measured delivery actually slowed 19%. A four-round research survey maps the current landscape: deterministic guardrails (hooks) vs probabilistic ones (prompts), clean-context review, self-improving feedback loops, specification-driven development, AI test quality crisis (high coverage but median 53% mutation score), and the Replit agent fabricating test results.

Multi-hop Retrieval: When Answers Are Scattered Across Documents

Standard RAG retrieves one set of documents per query, but real questions often need reasoning across 2-4 documents. IRCoT pioneered interleaved retrieval-reasoning, PAR²-RAG beats IRCoT by 23.5% accuracy on four benchmarks, and CompactRAG compresses LLM calls down to just two.

ai deep-dive

Self-RAG: Teaching the Model to Decide When to Retrieve

Self-RAG trains four reflection tokens (Retrieve / IsREL / IsSUP / IsUSE) into an LLM, letting it decide on-the-fly whether to retrieve, whether results are relevant, and whether its own output is grounded. ICLR 2024 Oral (top 1%), the 7B model beats ChatGPT and Llama2-chat + RAG on multiple QA benchmarks. The catch: it requires fine-tuning—no API-only models.

Table Serialization: How Format Choice Shapes RAG Retrieval for Tabular Data

Markdown-KV format achieves 60.7% LLM comprehension accuracy vs 44.3% for CSV — a 16-point gap from format alone. But retrieval and comprehension have different optimal formats: metadata prepend + row-wise key-value is the current best combination for table RAG.

learning guide

How to Practice Workplace English Speaking: Shadowing, Scenario Drills, AI Apps, and Meeting Phrases

30 minutes a day for 12 weeks — combine shadowing, scenario practice, and AI apps to go from 'I understand but can't speak' to holding your own in meetings.

tech debug

Claude Code Cloud Routines: When outcomes Pushes to a Feature Branch Instead of main

Cloud routine outcomes config creates a feature branch, so the agent commits and pushes there instead of main. Fix: remove outcomes + add explicit git checkout main in skills. Also hit a list API pagination bug (cursor never advances) along the way.

OMP agent loop: Why two while loops? What the outer 'stopped but woken by steering' layer actually does

omp's runLoopBody uses a double while loop: the inner loop drives the core model call → tool execution rhythm; the outer loop, when the agent would stop, drains queued steering / follow-up / asides to decide whether to run another turn. This design solves delivery timing for 'user typing while model streams' and 'background tasks quietly queueing messages'.

Learning Agent Design from Mature Coding Agents (4): Approval Grading and the Audit Trail

Looplane now grades effects as read/modify/modify_execute/execute and fails closed on unclassified tools. Native MCP tools default to execute unless trusted read-only metadata lowers them. Approval events still land in events.jsonl first and grants can scope to one change set or backend; general command rules and universal sandbox coupling remain unfinished.

Learning Design from Mature Coding Agents (37): Code Mode — Compiling Tool Calls into Batches of Executable Code

looplane now ships a bounded tool-program DSL: read-only programs support list/read/search/diff, repeat, and if_contains; modify/check transactions receive whole-transaction approval and roll back touched paths on failure. This is not arbitrary JavaScript/Python code mode, and transaction execution is not parallel.

Learning Design from Mature Coding Agents (26): Context Compression and Compaction — From Gap to Auditable Baseline

Mature-agent compaction must handle triggers, complete-turn cut points, and recovery. looplane now has an 85% high-watermark, automatic compaction, a deterministic native-loop fallback summary, persisted checkpoints, and workspace-context reinjection. Cross-runtime fallback, model-quality summaries, and live-provider long-session validation remain open.

Learning Agent Design from Mature Coding Agents (27): Cross-Session Memory — From Explicit Remembering to Semantic Recall

omp and claude-code provide cross-session memory while the other references mostly rely on instruction files. looplane now has an explicit remember/list/inject baseline: typed JSONL memories enter prompts across sessions, but retrieval is scope-and-recency only, with no semantic ranking, deduplication, forget command, or automatic extraction.

Learning Design from Mature Coding Agents (28): Dangerous Command Interception and Shell Escalation — Between Allowlists and Always Ask

All five projects combine allow/ask/deny decisions, compound-command inspection, and fail-closed behavior. looplane now has a deny-first classifier, critical floor, shell segmentation, timeout-deny, configured allow/deny rules, and visible policy reasons. Broader syntax coverage and live interactive validation remain open.

Learning Agent Design from Mature Coding Agents: Series Overview — Reading Five Codebases to Build My Own

I'm building my own Python coding agent called looplane. This series dissects the source code of five mature projects — pi, oh-my-pi, opencode, codex, and claude-code — topic by topic, while also comparing them with Looplane's current TUI, external CLI runtimes, local gateway, usage/OTel/session tooling, and Cloudflare slice. Every post follows a fixed five-part structure: design problem → how five projects do it → looplane's choice → academic grounding → improvement roadmap, with evidence cited at file#symbol level.

Hooks, Skills, Plugins: The Three-Layer Extension System of Mature Coding Agents

Hooks govern control flow, skills inject knowledge, and plugins package both. looplane now has opt-in deny-only project hooks, a bounded SKILL.md loader, exact enabled_skills selection, plugin manifests/install/list, and external-runtime projection. Input rewriting, full lifecycle coverage, remote registries, and a mature marketplace remain open.

Learning Design from Mature Coding Agents (36): LSP Integration — Pushing Compiler Diagnostics into Agent Context

looplane can now inject repository diagnostics and open-file state into the next model turn, push typed IDE context over WebSocket, package a VS Code bridge, and supervise long-lived LSP subprocesses through ManagedLspServer. Language-specific initialize/didOpen/didChange adapters and live-editor validation remain open.

Learning Design from Mature Coding Agents (30): MCP Integration — the Standard Socket for Tool Ecosystems

An MCP client must handle transports, tool refresh, approvals, and credential boundaries together. looplane now supports allowlisted stdio, Streamable HTTP/SSE, tools/resources/prompts, tools/list_changed, OAuth metadata/PKCE, and a 0600 credential store. A real authorization-server E2E and MCP-specific confirmation UX remain open.

Learning Design from Mature Coding Agents (35): Model Catalogs and Per-Role Routing — looplane's Role Aliases and Reviewer Lane

looplane now has static ModelRole/ModelRoute candidates, opt-in aliases such as --model @cheap, cross-provider fallback, and a no-tool reviewer lane that runs after verification. Role inheritance/override rules and automatic summarizer, parser, or scout routing remain open.

Learning Design from Mature Coding Agents (6): The ModelProvider Abstraction — Why Wrapping an SDK Is Not Enough

Wrapping an SDK directly buys you three walls within months: usage fields that don't agree, error semantics tied to SDK exception types, and tool-call formats that change per provider. All five reference projects separate 'wire protocol' from 'provider identity' as independent dimensions. Looplane goes further with pydantic canonical contracts (Message/ToolCall/Usage/ModelTurn) plus six protocol adapters, forces the OpenAI SDK's built-in retries to zero, and routes every failure through a classified ProviderErrorKind before any retry policy sees it. Its provider table is deliberately copied from pi's packages/ai — lineage, not coincidence.

Learning from Mature Coding Agents (29): OS-Level Sandboxing

An OS sandbox is the kernel boundary beyond path policy. looplane now ships a fail-closed CommandSandbox: sandbox-exec on macOS, Landlock plus seccomp on Linux, and exit 126 when containment cannot be proven. Coverage still focuses on verification commands, and external CI confirmation remains open.

Prompt Version Control: Changing One Word Can Drop an Eval from 5/5 to 0/5

Looplane's prompt is now `m3-exact-edit-v4`: the version persists into artifacts; core/tool/interaction/runtime/instructions/skills/workspace/memory are composed as stable or dynamic sections; and positive/negative examples cover replace_text, unified diffs, and direct replies. Unit tests pin the structure, while live-eval coverage still needs expansion.

Learning Design from Mature Coding Agents (7): Provider Retry Policy — From One 5xx to Bounded Retry and Fallback

Intermittent NVIDIA NIM 500s exposed Looplane's early gap: classified errors with no retry consumer. SDK retries are now disabled; the harness gives each candidate up to five attempts with jittered exponential backoff and capped Retry-After handling, then can move to an explicitly configured fallback model. Both model.retry and model.fallback enter the event log.

Learning Design from Mature Coding Agents (33): Session Recording and Replay — From Event Logs to Safe Forks

looplane now connects events.jsonl to a deterministic reducer, CLI timeline, canonical JSON, SDK replay, and safe event-point forks. Forking never replays prior tools or model calls; provider/live-runtime validation, redaction, and richer replay hooks remain open.

Learning Design from Mature Coding Agents (32): Subagents and Worktree Isolation — Teaching the Main Loop to Delegate

Mature subagents need roles, bounded fan-out, narrowed permissions, and a result contract. looplane now has native named-role schedules, parallel fan-out, child allowed_paths constrained by the parent, unsafe execution disabled by default, and parent-approved transaction proposals. Persistent background lifecycles, recursion trees, and automatic worktree merging remain open.

Learning from Mature Coding Agents (34): Telemetry and Cost Tracking — You Count Tokens, Then What?

looplane now has CostBreakdown, an explicitly estimated static GPT-5-family price table, per-lane usage/cost, and OTel cost fields. Unknown models still show tokens without invented dollars; broader pricing coverage, authoritative external-CLI bills, and live billing reconciliation remain open.

Learning Design from Mature Coding Agents (18): Toolset Design Philosophy — Drawing the Tool Surface Boundary

Looplane's core surface has grown from seven tools to nine with a read-only `tool_program` and rollback-capable `tool_transaction`; search prefers ripgrep and arbitrary shell remains absent. Native MCP tools join only from allowlisted servers and default to execute approval without trusted read-only metadata.

Learning from Mature Coding Agents (3): Workspace Isolation and Path Policy

Looplane's disposable clone and SafePathPolicy protect the source repo. `--sandbox-checks` can now wrap verification commands with macOS sandbox-exec, Linux bubblewrap, or Landlock, while Cloudflare provides a separate bounded Sandbox slice. Network policy, external-runtime coverage, and production hardening are not yet consistent across those backends.

Learning Design from Mature Coding Agents (38): Agent as a Service — Wrapping Your Loop in Something Other Programs Can Call

looplane now has a Cloudflare Durable Object run resource with async creation, status/cancel/artifacts, live NDJSON, and Last-Event-ID SSE; remote approvals use a separate short-lived capability. Python also provides an attach client and a stateful conversation WebSocket. Production deployment, cross-runtime parity, and full multi-tenant hardening remain unverified.

ai debug Ask AI in Practice

Why Ask AI Could Not List Its Course Maps: A Catalog Retrieval and Convergence Incident

The first observation of 'What course articles do you have?' was contaminated by an old cache entry. A real cache miss retrieved all four university maps but spent 51.169 seconds across three Writer and Critic passes; after catalog-specific retrieval and review fixes, one uncached production observation passed q21 in 26.821 seconds.

ai guide Ask AI in Practice

How Ask AI Finds Posts: Planner, Hybrid Retrieval, and Retry

Ask AI first extracts intent, complexity, and 1–4 search terms. It then routes across metadata, BM25, Vectorize, and RRF; a retry adds Critic gaps and disables the first-pass-only BM25 short circuit.

ai guide Ask AI in Practice

How Ask AI Indexes Posts: Chunks, D1 FTS5, and Vectorize

Ask AI indexing runs in two production stages: source-hash changes update D1, post chunks, and FTS5 first; embedding checkpoints and a delete queue then let Vectorize catch up asynchronously. The two stores do not share one transaction.

ai guide Ask AI in Practice

How a Question Moves Through Ask AI: UI, API, Agents, and Source Cards

Ask AI splits one question across the UI, `/api/chat`, Planner, Research, Writer, Validation, Critic, and Related stages. Answer text, displayed sources, and related-reading cards come from separate paths with separate gates.

ai guide Ask AI in Practice

Evaluating Ask AI Retrieval: Golden Contracts, Fixtures, Live Runs, and Evidence Boundaries

Ask AI keeps golden contracts, offline fixtures, live SSE output, and production observations separate. A passing fixture proves harness reproducibility; public sources can measure expected-source recall, but they do not expose hidden ranked chunks or establish model-graded faithfulness.

ai guide Ask AI in Practice

Debugging Ask AI in Production: SSE, Traces, Cache, Checkpoints, and Shadow Runs

One Ask AI request leaves five different evidence surfaces: public SSE, Langfuse traces, D1 logs, semantic cache, and a hidden shadow run. They expose different data, and no single surface reconstructs the complete retrieval context.

ai guide Ask AI in Practice

When Ask AI May Show Sources: Validation, Critic Review, Degradation, and the Source Gate

Ask AI finding a post does not mean the UI should display it as a source. An answer must pass deterministic Markdown and URL validation, then the Critic's relevance, intent, and grounding checks; if either gate fails, source cards are withheld.

ai guide Ask AI in Practice

How Ask AI Turns Evidence into an Answer: Writer Context and Citation Contracts

Writer sees the first 8 candidates for a factual query or 12 for a recommendation by default. Citations must use an exact `source_url` from that set, and weak or empty retrieval triggers an instruction to abstain rather than fill gaps from model knowledge.

How to Use Cloudflare Agent Memory: Keep Agent Memory Separate from RAG Documents

Agent Memory is a Cloudflare private beta service for letting agents remember users, teams, projects, and task context across conversations. It fits facts, events, instructions, and tasks; RAG documents, product data, files, and audit logs should still live in AI Search, Vectorize, D1, or R2.

How to Use Cloudflare Agents: Durable Runtime, Tools, and Real-Time Connections

Cloudflare Agents turns an agent session into a durable runtime: each agent instance has stable identity, local SQLite, WebSockets, scheduled work, recoverable execution, and tools. It is not just a chat example; it composes Workers, Durable Objects, AI models, Browser, Sandbox, AI Search, and MCP into a deployable agent app.

Where to Store Cloudflare AI App Data: D1, R2, and Durable Objects

An AI app should not put conversations, artifacts, memory, retrieval documents, locks, and eval traces into one store. D1 fits queryable product data, R2 fits large files and artifacts, Durable Objects fit named coordination and per-session state, and Agent Memory / AI Search / Vectorize handle memory and retrieval.

How to Use Cloudflare AI Gateway: Logging, Caching, Rate Limits, and Fallbacks

AI Gateway is the control plane for AI calls: one layer for logs, analytics, cache, rate limits, retry/fallback, BYOK, and Unified Billing. In Workers, use env.AI.run(..., { gateway }); with external SDKs, change the baseURL or provider-native endpoint.

Cloudflare AI Stack Guide: Building AI, RAG, and Agents on Workers

The Cloudflare AI Stack series covers the infrastructure around AI apps: where models run, how gateway control works, how RAG is built, how agents keep running, how memory is governed, and how browser, sandbox, secrets, data, and observability fit into a product.

How to Use Cloudflare Secrets Store: Worker Secret Reuse and AI Gateway BYOK

Secrets Store is Cloudflare's open beta account-level secret store, currently integrated with Workers and AI Gateway. It fits provider API keys, BYOK keys, and secrets reused across Workers; per-Worker secrets still work, but the governance scope is different.

How to Use Cloudflare Vectorize: Taking Control of RAG Retrieval

Vectorize is Cloudflare's vector database. AI Search is the right starting point for a managed RAG pipeline; Vectorize is the better fit when you need control over chunking, embeddings, metadata filters, hybrid retrieval, reindexing, and fallback behavior.

Looplane remote execution on Cloudflare: Worker, Sandbox, Capability DO, and durable RunSession

Looplane's old synchronous M6 path completed one real deployed coding run. It has since grown into an asynchronous control plane with RunSession, SSE, approvals, cancellation, and artifacts, but that newer path has not been live-revalidated. Audience-separated HMAC capabilities enter the Sandbox while provider credentials stay in the Worker; this is not production-traffic or SLO proof.

Looplane's disposable workspace and run bundle: why the source repository stays untouched

Looplane clones an exact Git commit into a detached-HEAD workspace inside the run directory before a runtime edits or verifies code. The source repository, execution workspace, and run artifacts therefore have distinct boundaries. This provides source isolation and an audit bundle, but it is not an OS sandbox.

Looplane's ExternalCodingRunner: why Codex and Claude Code CLI are external runtimes, not ModelProviders

`ExternalCodingRunner` is Looplane's second runtime lane. The external coding CLI owns its model loop and credentials; Looplane hands off a task and disposable clone, then treats the returned patch as untrusted input and reruns path audit, verification, and the source invariant. This is a capability-bounded handoff, not another `ModelProvider`.

Looplane's ModelProvider multi-gateway: multiple protocols, one canonical contract

Looplane collapses OpenAI-compatible, Responses, Anthropic, Gemini, Workers AI, scripted, and experimental Codex OAuth adapters into one `ModelProvider` contract. The Codex OAuth transport reads SSE but still reduces it inside the adapter into one canonical `ModelTurn`; AgentRunner does not consume token deltas.

Looplane's provider-neutral native loop: from one model turn to a verified terminal state

Looplane's native lane is controlled by AgentRunner: prepare a workspace, request a model turn, execute tool calls, append observations, and enter verification only when the model stops calling tools. Step, wall-time, repetition, token, and cancellation guards can terminate the run independently of the model. Protocol translation belongs to the next article.

Looplane's state-first event journaling: recovering between manifest commits and JSONL appends

Looplane maintains append-only `events.jsonl` and atomically replaced `session.json`, reconciling sequences before crash recovery. The same event contract now supports deterministic replay, canonical JSON replay, fork seeds at a selected sequence, and new workspaces without replaying old side effects; ambiguous `tool.started` or `verification.started` states still hard-fail.

Looplane's tool isolation: path allowlists, strict argv, process groups, and credential-free subprocesses

This article follows one Looplane tool call through its mechanical execution boundary: `SafePathPolicy` for paths and symlink escape, fixed argv with `shell=False`, a sanitized subprocess environment, read-version hashes plus atomic replace for writes, and process-group cleanup at timeout. Permission policy, OS containment, and tool programs are reserved for later articles.

Looplane's TUI and CLI: how a run becomes visible in the terminal

Looplane's TUI and plain CLI are two interfaces over the same runtime paths. The CLI selects a presentation mode from TTY state and flags, runners emit events, and the TUI projects those events into thinking, tool, approval, verification, and terminal states. The screen distinguishes native and external runtimes without treating UI entry points as proof of backend maturity.

How to Use Cloudflare Browser Run: Headless Chrome from Workers

Browser Run gives Workers access to Cloudflare-managed headless Chrome. Quick Actions fit one-shot tasks such as screenshots, PDFs, HTML, JSON, and crawls; Browser Sessions fit Puppeteer, Playwright, CDP, and Stagehand automation where you need full control.

Cloudflare Cache Rules: What to Cache and What Must Stay Dynamic

Cloudflare Cache Rules are zone-level cache policy: request expressions decide what is eligible for cache, how Edge TTL and Browser TTL behave, what dimensions enter the cache key, and how stale content, ETags, and purge interact. Use them for CDN cache policy; use the Worker Cache API for programmatic caching.

How to Use Cloudflare Containers: When Workers Need a Full Linux Runtime

Cloudflare Containers let a Workers app call on-demand serverless containers for workloads that need a full filesystem, a specific runtime, existing container images, or more CPU, memory, and disk. They do not replace Workers; Workers still handle entry, routing, and platform bindings while containers run the heavy runtime-specific work.

Cloudflare Edge Platform Guide: Running Websites and Apps on Cloudflare

The Cloudflare Edge Platform series answers one product question: how do Workers, D1, KV, R2, Durable Objects, Queues, Workflows, Cache, Images, Email, Turnstile, Observability, Browser Run, and Containers help you run a website or app cheaply and reliably?

Cloudflare Edge Platform Production Checklist: Custom Domains, Maintenance Pages, and Workers Limits

Before a Cloudflare app goes live, do not stop at a successful deploy. Check Custom Domains, Routes, www/root redirects, maintenance pages, CPU/memory/subrequest limits, log sampling, and fallback paths. This appendix turns the Edge Platform series into a production checklist.

How to Use Cloudflare Email Service: Sending, Routing, and Product Notifications from Workers

Cloudflare Email Service connects transactional email, magic links, notifications, and inbound routing to Workers. Arbitrary outbound sending currently requires Workers Paid; inbound routing is available on Free and Paid, with DNS, quota, message-size, bounce, and anti-spam limits still shaping the design.

How to Use Cloudflare Hyperdrive: Connecting Workers to Existing Postgres / MySQL

Hyperdrive solves the latency and connection-pooling problem when Workers connect to existing Postgres / MySQL databases. It uses edge connection setup, database-near pooling, and read query caching so a regional database works better with global Workers.

How to Use Cloudflare Images: Variants, Format Conversion, and Delivery Pipelines

Cloudflare Images has two paths: transform images stored in R2/S3/origin at the edge, or store images in Images and deliver named variants. The first is priced by unique transformations; the second also involves stored and delivered images.

How to Use Cloudflare Observability: Workers Logs, Traces, and Analytics Engine

Workers Observability is for debugging and request tracing; Workers Analytics Engine is for high-cardinality product events and custom metrics; GraphQL Analytics API is for querying existing Cloudflare product data. Keeping those roles separate prevents logs from becoming a database and keeps billing, monitoring, and product analytics from blending together.

How to Use Cloudflare Smart Shield: Reducing Origin Load

Smart Shield is Cloudflare's origin protection bundle: Smart Tiered Cache, connection reuse, Argo Smart Routing, Regional Tiered Cache, Cache Reserve, Health Checks, and Dedicated CDN Egress IPs reduce requests and connections reaching your origin.

How to Use Cloudflare Turnstile: Protect Forms and Public APIs Without Classic CAPTCHA

Turnstile is Cloudflare's CAPTCHA alternative: the client widget generates a token, and the server must validate it with the Siteverify API. Tokens expire after 300 seconds and are single-use; a widget without server validation is incomplete.

Cloudflare Workflows: Durable Multi-Step Execution on Workers

Cloudflare Workflows turns multi-step Workers processes into durable steps: each step can retry, sleep, wait for events, and register rollbacks, while instances can be inspected, paused, resumed, or terminated. Queues fit single-step background work; Workflows fit long processes that must remember progress.

tech deep-dive

How to Choose an Execution Environment and Sandbox: From Namespace and gVisor to Firecracker, E2B, and Lambda MicroVMs

A sandbox is not a single package but a spectrum—Namespace, cgroups, seccomp, gVisor, and Firecracker stacked by trust boundary; local OS sandboxes bound blast radius, cloud microVMs bound multi-tenancy, and the choice hinges on trust and ops cost.

A map of Looplane: how one coding-agent task crosses workspaces, runtimes, tools, and events

Looplane turns a coding-agent task into inspectable boundaries: native side effects cross Looplane tools, permissions, and sandboxing, while external runtimes retain their own loops and tools before returning a patch for Looplane audit. This article maps the planned 20-part series.

Looplane context pressure, compaction, and workspace reinjection

Near 85% context pressure, Looplane has two distinct paths: the native loop can apply one bounded deterministic history fallback, while a conversation runtime with native compaction can compact after a completed turn. Both paths re-anchor the next request with workspace context.

Looplane IDE/LSP Context: Diagnostics, Open Files, and the VS Code Bridge

Looplane normalizes up to 200 diagnostics and 32 visible files into bounded, repository-local, untrusted context. Its VS Code and managed-LSP paths supply signals rather than completion, rename, code actions, or full IDE RPC.

Looplane local OS sandboxes: fail-closed execution on macOS, bubblewrap, and Landlock

Looplane can wrap configured local commands and verification in macOS sandbox-exec, Linux bubblewrap, or Landlock/seccomp. A required unavailable backend stops with exit 126 instead of running bare, but external CLIs, MCP/LSP processes, and the entire Looplane process are outside this boundary.

Looplane model roles, fallback, cache hints, and estimated cost

Looplane uses a static model-role catalog and retries or falls back only after retryable provider errors. Cache data is a provider hint plus trace, while cost is a static-table estimate; neither is live routing intelligence or a bill.

Looplane Native MCP: transport, authorization, and approval boundaries

Looplane loads project MCP servers only through an explicit allowlist, projects stdio or Streamable HTTP capabilities into the existing ToolExecutor, and preserves hooks, approvals, timeouts, and cleanup.

Looplane permission layering: how dangerous commands become allow, ask, or deny

Looplane applies a non-bypassable critical floor, evaluates user, organization, and project denies before any allows, and keeps execute operations policy-gated even in dangerous mode. This decides authority; it is not an OS sandbox.

Looplane prompts, instruction precedence, and explicit memory: what the model actually sees

Looplane resolves user and root-to-leaf project instructions before rendering named prompt sections for runtime, skills, workspace state, and the latest 20 explicit memories. The pipeline is traceable and reloadable, but it is not semantic memory and repository text does not become system authority.

Embedding Looplane: SDK, ConversationController, and the WebSocket Boundary

Looplane exposes bounded-run and conversation contracts through a typed 0.x SDK facade. WebSocket attach wraps one prebuilt, controller-owned runtime session rather than providing conversation-ID resume or multi-client routing.

Looplane Skills, Blocking Hooks, and Plugin Packages

Looplane treats skills as bounded repository-local guidance, hooks as opt-in host commands that can only deny, and local plugin manifests as packages for skills and hooks; their authority is deliberately different.

Looplane Subagent Scheduling and Parent-owned Transactions

Looplane normalizes each subagent dispatch into dependency waves of at most four nodes, runs read-only children concurrently in isolated workspaces, then makes the parent repeat hooks, approval, and transaction execution for any modification.

Looplane tool programs, transactions, and safe concurrency

Looplane parallelizes calls only when they are read-only, concurrency-safe, and classified as READ. Tool programs provide bounded read-only repeat and branching, while transactions snapshot and restore possible workspace-file changes; external side effects are not rolled back.

Harvard CS50 AI Week 1: Knowledge — Propositional Logic, Model Checking, Inference Rules & Knowledge Representation

Week 1 shifts to knowledge representation: propositional logic syntax, model checking, Modus Ponens/Resolution inference, CNF conversion. Projects: Knights (logic puzzles) and Minesweeper (probabilistic inference).

Harvard CS50 AI Week 2: Uncertainty — Probability, Bayesian Networks, Markov Models & Genetic Inference

Week 2 shifts from deterministic to probabilistic: Bayes rule, Bayesian nets with D-separation, Markov chains, PageRank random walks. Projects: Heredity (genotype inference) and PageRank (web ranking).

MIT 6.7960 L03: Optimization Overview — SGD, Adam, LR Schedules & Scaling Rules

From SGD to Adam: pick the right optimizer and scale LR with batch size using scaling rules

Harvard CS50 AI Week 3: Optimization — Local Search, Simulated Annealing, CSP & Crossword Generation

Week 3 tackles optimization: hill climbing, simulated annealing escaping local optima, CSP framework with AC-3 arc consistency, backtracking with MRV/degree heuristics. Project Crossword builds a crossword puzzle generator.

MIT 6.7960 L04: Regularization in Practice — Weight Decay, Dropout, Batch Norm & Label Smoothing

Regularization isn't just anti-overfitting — mechanisms & combo strategies for WD, Dropout, BN, Label Smoothing

Harvard CS50 AI Week 4: Learning — Supervised Learning, k-NN, SVM, Reinforcement Learning Q-learning & Nim

Week 4 enters ML: supervised classification (k-NN, SVM, Perceptron), model evaluation, RL basics (MDP, Q-learning, ε-greedy). Projects: Shopping (purchase prediction with k-NN) and Nim (learning to play via Q-learning).

MIT 6.7960 PS1 Walkthrough: From NumPy MLP to PyTorch Autograd Backprop

Hand-write NumPy MLP + backprop → verify with PyTorch Autograd, fully reproducing OCW HW1 core concepts

Harvard CS50 AI Week 5: Neural Networks — Backpropagation, TensorFlow/Keras, CNN & Traffic Sign Classification

Week 5 enters deep learning: perceptron to multi-layer nets, backprop chain rule, loss functions, optimizers, TensorFlow/Keras modeling, CNN conv/pool. Project Traffic trains CNN to classify traffic signs.

MIT 6.7960 L05: CNN Architectures — From Convolution Kernels to Translation Equivariance

Lec 4 core: why CNN is the natural choice for grid data — convolution, translation equivariance, pooling, and classic architectures in one go

Harvard CS50 AI Week 6: Language — N-gram Language Models, TF-IDF QA, Parser & Attention

Week 6 processes natural language: N-gram conditional probability & smoothing, CFG syntax parsing with CYK, TF-IDF vector retrieval, attention mechanism & Transformer basics. Projects: Parser (syntactic generation) and Questions (TF-IDF QA system).

MIT 6.7960 L06: Modern CNN Architectures — ResNet, EfficientNet, ConvNeXt

ResNet's skip connections solve degradation, enabling 100+ layer nets; EfficientNet compound scales depth/width/resolution; ConvNeXt absorbs Transformer design to reclaim CV crown.

Harvard CS50 AI Synthesis (1): From Search to Language — The Complete Arc of Seven Weeks

Synthesis 1: Tracing how seven weeks form a deliberate knowledge arc from symbolic search to language models, revealing the design philosophy from classical AI to modern ML.

MIT 6.7960 L07: Scaling Rules for Optimization — Spectral View, Feature Learning, Hyperparameter Transfer

Optimization is not an isolated numerical problem: view SGD spectrally, the magnitude of weight updates determines feature learning; Maximal Update Parameterization transfers LR/init across width, and the critical batch size sets the marginal return of trading compute for convergence.

Harvard CS50 AI Synthesis (2): Project Portfolio — All 12 Projects Compared, Difficulty Tiered & Skill Mapped

Synthesis 2: Complete comparison of 12 projects — core algorithms, LOC estimates, difficulty tiers, check50 acceptance criteria, transferable skills. With difficulty grading and learning sequence advice.

MIT 6.7960 L08: Transformers — Tokens, Attention, Positional Codes, and How They Relate to MLPs/CNNs/GNNs

A Transformer is not an architecture from nowhere: tokens discretize data, attention does soft aggregation, positional codes restore order. Seen next to MLPs/CNNs/GNNs, all of them are special cases of 'weighted aggregation over neighbors'.

Harvard CS50 AI Wrap-up: What's Timeless, What's Changed, and Where to Go Next

Series finale: Retrospecting timeless core from 7 weeks/12 projects, gaps in 2020/2023 recordings vs 2026 reality, free OCW route completeness, and forward roadmap (Transformers, LLM fine-tuning, RAG, Agents, Evaluation).

MIT 6.7960 L09: Hacker's Guide to Deep Learning — Practical Know-How to Make Nets Actually Obey

Training neural nets is closer to engineering than magic: look at the data, overfit a mini-batch to prove capacity exists, then regularize back the generalization; learning rate is always the highest-leverage knob.

MIT 6.7960 L10: Memory and Sequence Modeling — RNNs, LSTMs, and Vanishing/Exploding Gradients

An RNN compresses the past into a hidden state, but recurrence makes gradients multiply over time — they either vanish or explode; LSTM decouples 'memory' from 'update' via input/forget/output gates so long-range information flows stably. Attention later replaced it because it reaches any history in O(1).

MIT 6.7960 L11: Representation Learning (Reconstruction-Based) — Autoencoders, VQ, Self-Supervision

Representation learning compresses raw data into a 'useful' vector: autoencoders force a meaningful latent space via reconstruction, VQ discretizes it into a codebook, and self-supervision turns 'mask-and-reconstruct' into free supervision.

MIT 6.7960 L12: Representation Learning (Similarity-Based) — Metric Learning, Contrastive, InfoNCE

Similarity-based representation learning does not reconstruct input; it directly shapes latent geometry: pull same-class representations together, push different ones apart. InfoNCE turns this into 'spot the positive among negatives', and alignment / uniformity give it interpretable metrics.

MIT 6.7960 L13: Theory of Representation — Inductive Biases, Gaussian Processes, and the NN–GP Correspondence

Take a net to infinite width and its random-init output becomes a Gaussian process (NN–GP); its training dynamics freeze into the Neural Tangent Kernel (NTK). This theory analyzes nets and, in reverse, guides us to design the 'right inductive bias'.

MIT 6.7960 L14: Generative Models Basics — Density/Energy Models, GANs, Autoregressive, Diffusion

Generative models learn the data distribution p(x). Density models model probability directly, energy models use an unnormalized potential + sampler, GANs let a discriminator force realistic samples, autoregressive predicts the next token step by step, and diffusion dodges tricky maximum-likelihood via 'add noise then learn to denoise'.

MIT 6.7960 Approximation Theory — Universal Approximation, Barron's Theorem, and Why Depth Matters

A single hidden layer can in principle approximate any continuous function (universal approximation), but width can blow up exponentially with dimension; Barron's theorem lets error decay as 1/sqrt(n) independent of dimension for a specific function class; and depth yields exponential width savings on compositional functions — that is the real reason deep beats shallow.

MIT 6.7960 Graph Neural Networks (GNN) — Message Passing, Permutation Equivariance, and the Expressiveness Ceiling

A GNN is essentially 'an MLP with local message passing on a graph' — it generalizes CNN's fixed-grid neighborhood to arbitrary topology. It must satisfy permutation equivariance/invariance. In theory, a first-order GNN's expressiveness is bounded by the Weisfeiler–Lehman graph isomorphism test: some structures it can never tell apart, which is exactly the gap GIN, positional encodings, and subgraph tricks later fill.

MIT 6.7960 L15: Variational Autoencoders (VAE) — ELBO, Reparameterization Trick, and Latent Representations

The core of VAE is ELBO + reparameterization: log p(x) is replaced with E_q[log p(x|z)] − KL(q(z|x)‖p(z)); the encoder outputs μ/σ and z = μ + σ⊙ε (ε ~ N(0,1)) makes sampling differentiable. Training = reconstruction + KL in tension, which gives rise to β-VAE, posterior collapse, VQ-VAE, and related fixes.

MIT 6.7960 L16: Conditional Generative Models — cGAN, cVAE, and Classifier-Free Guidance

The key to conditional generation is 'feed y into the model': cGAN concatenates y into G/D; cVAE passes y to both encoder and decoder; in diffusion, Classifier Guidance uses gradients from an external classifier to push samples toward a class, while Classifier-Free Guidance trains conditional + unconditional together and linearly combines them at inference — the latter is the standard weapon behind Stable Diffusion and Imagen.

MIT 6.7960 L01: Course Introduction — A Map of Deep Learning, Why Depth Works, and Your First Training Loop

Lecture 1 is the 6.7960 opener: deep learning took off because data + compute + algorithms matured together; the course threads from architectures (CNN/GNN/Transformer) through training, representation, generation, transfer, scaling, and LLMs; ends with a ~30-line PyTorch training loop to confirm your environment works.

MIT 6.7960 L17: Out-of-Distribution Generalization — Distribution Shift, Spurious Correlations, and Three Practical Remedies

OOD failure is not a bug, it's the i.i.d. assumption breaking: covariate shift (image style changes), label shift (class proportions change), concept shift (a word's meaning changes) each need different responses; the most common cause is the model latching onto spurious correlations (using grass as a cue for cows); IRM and domain randomization try to fix this in training data structure, test-time adaptation fixes it at inference.

MIT 6.7960 L18: Transfer Learning — Pretraining, Feature Extraction, and Fine-Tuning Strategies

Transfer learning's core insight is 'features learned on big data are good general-purpose representations': freeze the backbone and train only a linear head when downstream data is tiny; full fine-tune when data is plentiful; reach for LoRA / adapter when compute is tight. SimCLR and MAE removed the need for upstream labels and pushed downstream quality another notch.

ai guide

Should You Rent a GPU to Learn Model Training? GPUtw.ai, LoRA, Jupyter, and the First Experiment

GPUtw.ai makes sense as a short-rental GPU learning tool: start with Jupyter, Ollama, or ComfyUI, then try LoRA/QLoRA on a small model. It is not a large foundation-model training platform, and the first run should verify deployment, billing, and data retention with a small budget.

How AI Agent Search Infrastructure Is Changing: Keenable, Independent Indexes, and NEEDLE

Keenable.ai positions itself as search infrastructure for AI agents: a 100B+ document index, Search/Fetch APIs, MCP/CLI entry points, 100K free monthly requests, and keyless public endpoints. It is worth tracking, but the 100B+ index, latency, and quality claims are still mostly company-provided; NEEDLE is open, but needs external reruns and human review.

ai deep-dive

How screenshot-to-code Converts Screenshots to Code: Agent Loop, Asset Extraction, Visual Verification

screenshot-to-code is not a one-shot screenshot-to-HTML tool. Its core is a 30-step Agent Loop with 7 tools — extracting real assets from screenshots, self-verifying with Playwright, and running 4 models in parallel so users pick the best output. 74,500+ GitHub stars, MIT License.

ai deep-dive

How Agents Accumulate Team Judgment: Warp's Skill Feedback Loop

Warp's self-improving agent pattern is not about dumping every mistake into a prompt. A base skill does the work, humans leave feedback in GitHub or Slack, an improver skill turns repeated signals into a small diff, and humans review the PR before the next run inherits it.

TinyFish: Free Search and Fetch Infrastructure for AI Agents

TinyFish provides four web APIs for AI agents: Search, Fetch, Agent, and Browser. Search and Fetch are permanently priced at $0 with no credit card requirement, making them a practical default layer for RAG and document retrieval.

How Does A/B Testing Turn a Product Change Into an Estimable Effect?

A/B testing turns a product change into an estimate with uncertainty. A useful report covers effect size, confidence, guardrails, randomization, and launch risk.

Why Not Run Many t-Tests? What Is ANOVA Protecting?

ANOVA first checks whether three or more group means differ overall, so you do not inflate false-positive risk by running many pairwise t tests.

Why Do Large-Sample Approximations Work, and When Do They Fail?

Large-sample normal approximation describes the behavior of estimators, not raw data. It is useful, but dependence, boundaries, and distribution shift can make it unreliable.

How Does Bayesian Inference Connect Prior, Data, and Posterior?

Bayesian inference updates uncertainty about an unknown parameter by combining prior belief with the likelihood from observed data, producing a posterior distribution.

What Do Bias, Variance, and Consistency Check in Point Estimation?

Bias checks whether an estimator is centered correctly, variance checks sampling fluctuation, MSE combines both, and consistency asks whether the estimator approaches truth as sample size grows.

When the Formula Distribution Is Unknown, How Does Bootstrap Estimate Uncertainty?

Bootstrap estimates uncertainty by resampling from the observed sample with replacement, rebuilding many sample-like datasets, and watching the statistic fluctuate.

Causal Inference Basics: Why Prediction Accuracy Does Not Mean Real Effect

Causal inference separates prediction from effect. A model can predict who will buy without proving that an intervention will make them buy.

How Do You Tell Goodness-of-Fit From Independence in Chi-Square Problems?

Chi-square tests compare observed counts with expected counts. First decide whether the problem is goodness-of-fit for one categorical variable or independence for two categorical variables.

When Should Bernoulli, Binomial, Normal, and Poisson Appear?

Distributions are names for data-generating situations, not formula cards. Learn when Bernoulli, Binomial, Poisson, and Normal distributions fit a problem.

How Do You Write Confidence Intervals Without Only Memorizing Bounds?

A confidence interval puts a point estimate back inside sampling fluctuation. Computing bounds is only the first step; you also need to explain standard error, critical values, and coverage.

When You See a Dataset, What Statistics Should You Check First?

Data type determines the statistical tools you can use. Start with categorical, numeric, count, and time-ordered data, then choose summaries that fit the question.

How Does the Delta Method Estimate Uncertainty for F1 and Ratio Metrics?

The delta method transfers uncertainty through a smooth function: the local derivative expands or shrinks the estimator's original standard error.

What Makes an Estimator Good: Bias, Variance, or MSE?

An estimator is a rule for using samples to infer a population parameter. To judge whether it is good, look at bias, variance, and MSE together.

When a Mixed Problem Appears, How Do You Pick the Tool in 30 Seconds?

At the final review stage, train problem recognition: identify data type, unknown quantity, and decision goal before choosing a formula and writing a contextual conclusion.

What Do Expectation and Variance Mean in Exams and Model Evaluation?

Expectation describes long-run center; variance describes fluctuation. This post computes E[X], E[X^2], and Var(X), then connects them to average loss and model stability.

How Does Experimental Design Make Results Interpretable Rather Than Merely Correlated?

Experimental design decides whether a result can be interpreted. Randomization, control, blocking, replication, blinding, and pre-specified outcomes give inference a usable foundation.

How Does Fisher Information Tell You Whether a Parameter Is Stable?

Fisher information uses likelihood curvature to measure how well the data locate a parameter; larger information usually means a smaller standard error for the MLE.

Confidence Intervals Are More Than t-Tables: What Is the General Construction?

A confidence interval is built by defining the target estimate, describing its sampling error, and choosing a rule that turns uncertainty into a range.

How Does a GLM Choose Distributions and Link Functions by Data Type?

A generalized linear model starts from the response type, chooses a suitable distribution, and uses a link function to connect the mean to a linear predictor.

From H0 to p-Values, What Decision Is a Hypothesis Test Making?

A hypothesis test is a decision process under uncertainty: write H0/H1, choose alpha, compute a test statistic and p-value, then decide whether the data is strong enough to challenge H0.

How Do Estimation, Testing, Likelihood, and Bayes Fit on One Inference Map?

The inference map starts with the question type: point estimate, uncertainty interval, decision test, likelihood model comparison, Bayesian update, or resampling.

How Does the Likelihood Ratio Test Compare Nested Models?

The likelihood-ratio test compares the log likelihood of a restricted model with a full model; the usual chi-square reference only makes sense under nested-model and approximation conditions.

When OLS Assumptions Fail, How Can the Regression Line Still Be Used?

OLS is a useful baseline, but coefficient interpretation, inference, prediction, and diagnosis depend on assumptions about linearity, errors, independence, and variance.

How Does Logistic Regression Move From Probability to Thresholds and Error Costs?

Logistic regression estimates probabilities first. Classification decisions come later, when thresholds turn those probabilities into actions under real error costs.

Why Should Classification Start With Log Odds?

Logistic regression connects a linear score to a probability between 0 and 1. Understanding odds, log odds, and odds ratios prevents wrong coefficient interpretations.

Why Does MAP Turn Priors Into Regularization?

MAP maximizes the posterior. After taking logs, the prior becomes a penalty term, which connects Bayesian estimation to L1, L2, and regularized ML objectives.

How Do Matching and Weighting Make Observational Data More Experiment-Like?

Matching and weighting do not turn observational data into a true experiment. They try to make treatment and control comparable on observed variables.

Why Does MLE Ask Which Parameter Most Likely Generated the Data?

MLE fixes the observed data and compares which parameter values make that data most plausible; log likelihood turns products into sums and connects directly to negative log loss.

Why Does the Method of Moments Match Sample Moments to Population Moments?

Method of Moments matches sample moments to theoretical population moments, then solves for parameters. It is not always the most efficient method, but it builds the first intuition for parameter estimation.

Missing Data Is Not Just Blank Cells: How Does It Distort Statistics and Models?

Missing data can change representativeness, bias estimates, and mislead ML systems. The first question is why the data are missing.

How Do You Write an ML/AI Evaluation Report That Is More Than a Leaderboard Score?

A useful ML/AI evaluation report turns statistical evidence into a decision: ship, stage, roll back, or run more experiments.

What Do Residuals, Outliers, and Leverage Reveal About Model Failure?

Model diagnostics turn fitted errors into evidence: residual patterns, outliers, leverage, and influential points reveal how a model fails.

How Does Multivariate Analysis Organize Features That Move Together?

Multivariate analysis looks at features together. Covariance, correlation, and PCA reveal shared directions that univariate summaries miss.

What Kind of Optimal Test Is the Neyman-Pearson View About?

The Neyman-Pearson view treats a test as a decision rule: under a fixed Type I error rate alpha, choose the rejection region with the highest power.

What Assumptions Do Nonparametric Methods Relax, and What Do They Cost?

Nonparametric methods are not assumption-free. They relax fixed distributional forms, often gaining flexibility while paying in efficiency, interpretation, or overfitting risk.

How Should You Analyze NTU IM 114-115 Statistics Papers Without Memorizing Answers?

Past papers train question-analysis discipline, not fortune-telling. Each problem should return to data type, unknown quantity, statistical tool, calculation path, and contextual conclusion.

How Do You Avoid Missing Cells in Joint Distribution and PMF Transformations?

Joint PMF problems require listing every cell. Marginalization, conditional probability, and variable transformations are all sums or regroupings of the original cells.

Conditional Probability, Independence, and Bayes: What Viewpoint Is the Problem Switching?

Probability problems are often hard because the viewpoint changes. Define events first, then distinguish conditioning, independence, mutual exclusivity, and Bayes' rule.

How Do Samples, Statistics, and Sampling Distributions Differ?

A sample is the data, a statistic is a function of the sample, and a sampling distribution is the distribution of that statistic under repeated sampling.

How Do PMF, PDF, and CDF Turn Probability Into Computation?

Random variables turn uncertain outcomes into numbers. PMF, PDF, and CDF then let you compute discrete probabilities, continuous interval probabilities, thresholds, and model-score distributions.

How Should coef, SE, t, F, and R-Squared Be Read Together?

A regression table is not a p-value list: coef, SE, t, F, and R-squared answer effect size, uncertainty, single-coefficient tests, overall model signal, and in-sample explanation.

Why Do Ridge, Lasso, and Weight Decay Make Models More Stable?

Regularization adds a preference against extreme parameters. Ridge, Lasso, and weight decay trade some training fit for a model that generalizes more reliably.

How Can Statistics and ML Evaluation Be Rerun to Reach the Same Conclusion?

A reproducible workflow preserves the evidence chain from data to conclusion. Results need data versions, code, seeds, environment, metrics, and raw outputs.

Why Can a Sample Say Something About a Population or Model?

Sampling makes sample statistics fluctuate, and standard error describes that fluctuation. This post separates SD, SE, sampling distributions, and CLT, then connects them to benchmark uncertainty.

How Do Sampling Distributions Become Exam-Ready Reasoning?

A sampling distribution describes how a statistic fluctuates under repeated sampling. Means, proportions, and variances each connect to common distributions used in intervals and tests.

After 53 Posts, How Do You Connect Statistics to ML, Causality, and Mathematical Statistics?

The series does not finish all of statistics. It gives beginners a working map for exams, ML/AI evaluation, causality, Bayesian thinking, time series, and mathematical statistics.

How Does One Regression Line Become Prediction, Interpretation, and Error?

Simple linear regression uses one X to describe the average change in Y. Slope, intercept, residuals, and squared error form the smallest supervised learning model.

How Does Monte Carlo Use Repeated Simulation to Answer Hard Statistical Questions?

Monte Carlo repeats a data-generating process many times so sampling variation, power, coverage, and evaluation instability become visible.

Where Should You Start Statistics If You Need Exams and ML/AI?

Do not start statistics exam prep by memorizing formulas. Start with the sequence of data, probability, sampling, inference, regression, then connect those ideas to model evaluation, A/B testing, and uncertainty in ML/AI.

Why Should Time-Series Data Not Be Randomly Split?

Time-series data have order. Random splits can leak future information into training and make forecasting or monitoring results look better than they are.

Which Test Fits a Two-Group Mean or Proportion Difference?

Two-group comparisons start by classifying the outcome and the design: numeric or binary, independent or paired. That choice determines the standard error, test statistic, and conclusion.

How Does Variable Selection Avoid Memorizing the Training Data?

Variable selection is not only about choosing predictors. It is about avoiding noisy training-set wins that do not generalize.

Statistics Is Not Formula Memorization: What Is It Deciding?

The core of statistics is judgment: describe data, estimate unknowns, compare differences, inspect associations, and make decisions under uncertainty.

How to Use Cloudflare AI Search: Data Sources, Hybrid Retrieval, and Workers Bindings

Formerly AutoRAG, the managed search primitive: drop files into built-in storage or attach R2 and websites, auto-index with Markdown conversion plus vector and BM25, retrieve with hybrid, RRF, and reranking, and query from Workers via namespace or instance bindings, REST, or MCP.

tech deep-dive

What Is GPUtw.ai? Taiwan GPU Cloud, Short-Rental Compute, and Researcher Workflows

GPUtw.ai is a Taiwan-based short-rental GPU cloud. Its main value is not maximum scale, but Taiwan data centers, prepaid credits, Jupyter/ComfyUI/Ollama/vLLM templates, Vault storage, and team billing. Public information is enough for a service introduction, not enough for procurement or production endorsement.

Why Did 'I Want a Beginner AI Course' Return Zero Results? Debugging Chinese Tokenization and a Broken RAG Data Path

Entering '我想找入門的ai課程' in Ask AI showed zero searched posts and triggered a refusal, while Related Reading recommended exactly the right article. The first fix addressed Chinese tokenization, the LIKE fallback, and the Vectorize data path. A second pass added short Han-and-number tokens, post metadata retrieval, and a rule that exposes sources only after both Validation and Critic pass.

Harvard CS181 HW0: Do These 4 Problems First — They Tell You What to Patch

HW0 checks CS181 prerequisites in four problems — y=Xw solvability, optimizing an objective, reasoning about randomness, and OLS in Python. The problem that slows you down most is the gap to patch before HW1.

Harvard CS181 HW1: Ice Core Regression — Linear, Kernel, and Neural Nets in One Assignment

HW1 uses an 800k‑year ice‑core temperature dataset to implement three regression models (OLS, RBF kernel, MLP) and compare them on the same data, laying the groundwork for later classification and deep‑learning assignments.

Harvard CS181 Machine Learning: Your 2026 Roadmap Through 7 Homeworks (With a 4-Year Comparison)

CS181 2026 is A3 with hw0–6 as the weekly clock (no public recordings); 2025 adds a practical, 2024 has two midterms, 2023 was taught by Weiwei Pan. Start with HW0, then follow hw1→hw6.

Harvard CS50 AI Week 0: Search — From DFS, BFS, A* to Minimax and Alpha-Beta Pruning

Week 0 opens with search algorithms: BFS for shortest paths, Minimax for adversarial play, Alpha-Beta for pruning. Two projects: Degrees (BFS) and Tic-Tac-Toe (Minimax).

tech deep-dive

Choosing a Mobile Client for Claude Code: Moshi's SSH Terminal, moshi-hook, and Pricing

Moshi is an iOS/Android terminal app (plus a free Moshi Desktop web UI) that connects over SSH/Mosh straight to your own machine to drive Claude Code, Codex, and other coding agents. The free tier is a complete terminal; Pro ($7.99/mo and up) unlocks Mosh's connection resilience, deep tmux integration, and the diff viewer.

ADE Workspace Showdown: ADE vs Superset vs Herdr vs Orca

Four philosophies of ADE workspaces: arul28/ADE's Brain+Lane, Superset's 100-agent IDE, Herdr's Rust-native runtime, and Kadro/Orca's pane and Fleet angles — compared in one table with a decision tree for when to pick a workspace over an Omnigent-style control plane.

Governance Deep Dive: Policies, Omnibox, Spend Controls and Credential Brokering

Omnigent moves governance off prompts into a Server-side Policy engine: Python functions returning allow/deny/ask, a three-layer stack with cost budgets and tool caps, plus Omnibox OS-native isolation via bwrap/seatbelt and egress credential brokering — compared with five peer governance stacks.

How to Read 2026 Coding Benchmarks: SWE-bench, Terminal-Bench, DeepSWE, Aider Explained

The same model can score 20 points apart on different harnesses, 32% of SWE-bench Pro verifier judgments were found to be wrong, and DeepSWE's 113 tasks make most models score zero. This guide decodes six major coding benchmarks — what they test, which are easy to game, and which ones you should care about.

Harvard CS50 AI Guide: Seven Weeks, Twelve Projects, and How to Follow a Course Filmed in 2020

CS50 AI's OpenCourseWare edition publishes seven weeks of lectures, slides, notes, and twelve Python projects with autograder feedback, plus a free CS50 Certificate if you score at least 70% on every project. The catch: weeks 0–5 still use the Spring 2020 recordings; only Week 6 (Language) was re-recorded, in 2023.

LLM API Routing: Direct, Aggregator, or Cloud — A Price Comparison

The same model can cost 2-5× more depending on the channel. Direct API is simplest, aggregators (OpenRouter) are most flexible, cloud platforms (Bedrock/Vertex) suit enterprises. This post compares actual August 2026 prices across six channels with a decision tree.

Same Name, Different Layer: meta-harness, ACP, HarnessAgent and Flue

meta-harness means two things: Databricks' control plane and Stanford's outer-loop optimizer. This post uses a four-layer model (MCP/ACP/Runtime/meta-harness) to place Omnigent, Zed ACP, Vercel HarnessAgent and Cloudflare Flue.

ai guide Reading MIT 6.7960

Reading MIT 6.7960: One Course, Two Official Editions — Complete the OCW 2024 Package, Read the 2025 Decks for What's New

MIT 6.7960 Deep Learning (Fall 2025) publishes all 21 lecture decks as public Dropbox PDFs, and most required readings map to free textbook chapters; but the five problem sets are released only through Gradescope, and solutions plus recordings live behind Canvas login. This guide covers how the three instructors split the course, a topic map of all 21 lectures, textbook-based substitutes for lectures, and where outside self-learners realistically stop.

Why MoE Wins: The Architecture Behind Every 2026 Frontier Model

Nearly every frontier open-source model in 2026 is MoE: Ornith 35B activates only 3B to beat 31B dense models, MiniMax M3 uses 456B total but 45.9B active to hit SWE-bench Pro 59%, DeepSeek V4 runs 1.6T total with 49B active. This post explains why MoE dominates coding and agentic benchmarks using four case studies.

One Multi-Agent Task, Four Implementations: Omnigent YAML vs LangGraph vs CrewAI vs Goose

The same Polly task — parallel git worktrees plus cross-vendor review — implemented four ways: Omnigent YAML governs at the Server layer, LangGraph controls flow with a StateGraph, CrewAI assembles roles quickly, and Goose ships a desktop Recipe, compared on tokens, latency, and maintainability.

ai deep-dive

OLMo: The Only Language Model Family That Open-Sources Its Training Data

Allen AI's OLMo is the only language model family that fully publishes weights, training data (Dolma, 9.3T tokens), training code, all intermediate checkpoints, and evaluation tools. OLMo 3's 32B Think model hits 96.1% on MATH — and you can use OlmoTrace to trace any output back to the exact training data that produced it.

Managing Multiple Agents Together: Omnigent's Meta-Harness, Policies, and Cross-Device Sessions

Databricks' open-source Omnigent wraps Claude Code, Codex, Cursor, Pi and custom agents in a Runner/Server + Omnibox sandbox, adding three-layer Policies and shareable persisted Sessions so you can swap models and harnesses with one-line changes — 9.3k stars, still alpha.

Open-Source AI Licensing Guide: What MIT, Apache 2.0, and Llama License Actually Allow

'Open-source' in AI doesn't mean what it means in software. MIT and Apache 2.0 let you do almost anything; the Llama License requires a separate deal above 700M MAU; old Gemma terms let Google change rules unilaterally (Gemma 4 switched to Apache 2.0). This guide maps what you can and can't do by license type.

ai guide 認識 AI 模型

Self-Hosting Open-Source LLMs: Framework Choice, Hardware Math, and When It Beats APIs

Open-source models now match closed-source on coding benchmarks, but self-hosting isn't just picking a model — vLLM handles high-concurrency production serving, SGLang is 29% faster on prefix-heavy workloads, Ollama is the local dev default, and llama.cpp runs on the least hardware. A100 cloud rentals run ~$1.4-2.2/hr; self-hosting breaks even at roughly 100M tokens/month.

Three RL Post-Training Playbooks: How Ornith, Nous Research, and MiniMax Built Dark Horse Models

Three non-big-lab teams used different RL post-training strategies to produce benchmark dark horses in 2026: Ornith's self-improvement loop (GRPO), Nous Research's DataForge + Atropos execution-reward RL, and MiniMax's massive-scale RL across 200K real environments. Different strengths, but one shared proof point: post-training RL matters more than pretraining scale.

Tokens, Context Windows, and Inference vs Training: Three Things to Know Before Using AI Models

Models don't read words — they read tokens. A Chinese character is typically 1-2 tokens; an English word is 1-3. The context window is the token limit per request. Inference is using a model; training is teaching one. What you do every day is inference.

Embeddings: How Models Turn Words Into Computable Vectors

Models don't understand text — they only understand numbers. Embeddings map each token to a vector of several hundred dimensions, where semantically similar words end up close together in vector space. This is the shared foundation behind search, RAG, and classification.

How to Read a Model's Report Card: Benchmarks, Arena Elo, and the Traps Behind the Numbers

Benchmark scores in model releases have three common traps: cherry-picking (only showing wins), contamination (test data leaking into training), and saturation (when everyone scores 90%+, the benchmark stops being useful). The most manipulation-resistant signal is Chatbot Arena's Elo ranking — real humans, blind voting, uncontrolled questions.

Fine-tuning vs RAG: When to Teach the Model vs When to Look Things Up

Data changes often and you need citations → RAG. Need consistent style or want to run on a small device → fine-tuning. In practice, many production systems use both: fine-tune a small model that speaks your domain language, then use RAG to supply up-to-date facts.

How Models Improve Themselves: Gradient Descent and the Training Loop

A model uses loss to know how wrong it is and gradients to know which direction to adjust. Gradient descent repeats three things: compute loss, compute gradients, update parameters. The learning rate controls step size — too large and you overshoot, too small and training takes forever.

How a Model Knows It's Wrong: Loss Functions and Cross-Entropy

Every time a model predicts the next token, it assigns a probability to every candidate word. A loss function measures how far that probability distribution is from the correct answer — the further off, the higher the loss, the more the model knows it got it wrong. Cross-entropy is the standard formula; perplexity is its human-readable translation.

Quantization & Inference Optimization: Running a 70B Model on Your Laptop

A 70B model needs ~140GB VRAM in FP16, but 4-bit quantization shrinks it to ~35GB. With llama.cpp's partial CPU offloading, it can run on consumer hardware. GGUF naming conventions (Q4_K_M, Q5_K_S) tell you the precision-size tradeoff. KV cache is why long conversations slow down.

Scaling Laws: How Big Should a Model Be, and Why Bigger Isn't Always Better

Scaling laws show that loss decreases predictably with more parameters, data, and compute — following power-law relationships. The Chinchilla paper's key finding: most models were too large and undertrained. Given the same compute budget, training a smaller model on more data produces better results. This reshaped the entire industry's training strategy.

ai guide 認識 AI 模型

Understanding AI Models: 18 Articles from Tokens to Self-Hosting

You don't need to become a researcher to understand AI models systematically. This series starts from what you can see (tokens, context windows) and works up to self-hosting open-source models — 18 articles covering everything you need to choose models, read benchmarks, and estimate costs.

Tokenization: The BPE Algorithm, and Why Chinese Costs More Than English

Models charge by tokens, not characters. The BPE algorithm starts from individual bytes and repeatedly merges the most frequent adjacent pair to build a vocabulary. English 'understanding' might be 1-2 tokens, but Chinese '理解' could take 2-3 — same meaning, higher cost.

Pre-training, SFT, RLHF: Three Stages That Turn a Text Predictor into a Useful Assistant

Every LLM goes through three training stages: pre-training reads the internet to learn language, SFT uses example conversations to learn the format, and RLHF uses human preferences to learn what a good answer looks like. The gap between a base model and a chat model is what the last two stages do.

Transformers and Attention: How Models Decide Which Words to Look At

The core of the Transformer is self-attention: for each token, the model computes how relevant every other token is, then takes a weighted sum. This lets the model reach across distance to figure out that 'it' refers to 'cat' not 'mat' — and is the foundation for how it handles long documents.

Ahead of AI: How a Scholar Built 200K Subscribers by Publishing Monthly, Not Daily

Computational biology PhD turned UW-Madison professor Sebastian Raschka launched Ahead of AI on Substack in 2022, publishing monthly deep dives into LLM papers and architectures. Four years later: 200K+ subscribers, zero sponsorships, and a book-newsletter flywheel that proves low frequency and high depth can win in a crowded AI newsletter market.

ByteByteGo: From a Self-Published Book to a Million-Subscriber System Design Empire

Former Twitter/Apple/Zynga engineer Alex Xu self-published System Design Interview in 2020 and hit the Amazon bestseller list. In 2022, he and ex-Discord engineer Sahn Lam launched a Substack newsletter that crossed 26K subscribers in month one and hit one million in two and a half years. From book to newsletter, YouTube, and paid platform, ByteByteGo reached $3.5M ARR in 2024 with a 26-person team — all fully bootstrapped.

Daily Dose of Data Science: From a Cancelled Master's to 200K Newsletter Subscribers

Former Mastercard AI engineer Avi Chawla turned a cancelled US master's admission into a daily Substack newsletter — 150-word visual posts on data science. 10K subscribers in 5 months, income exceeding his full-time job, 200K+ subscribers and a paid course platform four years later.

Dense Discovery: 403 Issues of Deliberately Staying Small

Berlin-born, Melbourne-based designer Kai Brach spun his indie print magazine Offscreen into Dense Discovery, a weekly curated newsletter. Eight years, 403 issues, 36,000 subscribers, 63% open rate — deliberately not scaling, sustained by a single $649+ sponsor slot per issue and a Friends membership program. Proof that 'enough' is a viable business model.

Lenny's Newsletter: How an Ex-Airbnb PM Built the Most Influential Product Management Newsletter

Former Airbnb product lead Lenny Rachitsky left in 2019, wrote a viral Medium post, moved to Substack, and built a 1.2M-subscriber newsletter empire through obsessive quality (50+ revision cycles per post), a 40K-member Slack community, a top-ranked podcast, and an annual summit — all without hiring a single full-time employee.

Morning Brew: From a Michigan Dorm Room to a $75M Media Empire

Alex Lieberman started a PDF called Market Corner in his Michigan dorm, rewriting Wall Street Journal-style news in a casual, conversational tone. Five years later, Morning Brew hit 4 million subscribers and $13M in revenue, then sold to Insider Inc. for $75M. Their referral program accounted for up to 75% of new signups at its peak — one of the most successful growth engines in newsletter history.

Not Boring: When Writing Itself Becomes the Deal Flow

Former investment banker and startup VP Packy McCormick turned a COVID-era social club pivot into Not Boring, a long-form business analysis newsletter. In two years he hit 100K subscribers and $1M in sponsorship revenue, then extended the flywheel into three venture funds totaling $68M+ across 200+ investments — proving that writing can literally be deal flow.

One-Person Media Company: Ten Newsletter Cases and Four Revenue Playbooks

From Stratechery proving in 2014 that one person can make a living writing analysis to TLDR hitting $10M+ ARR in 2024 — ten newsletter cases distilled into four revenue playbooks, three content models, and one universal rule: format choice determines the ceiling.

The Pragmatic Engineer: From Uber Payments Lead to Substack's #1 Tech Newsletter

After six years managing Uber's payments infrastructure and witnessing pandemic layoffs, Gergely Orosz launched The Pragmatic Engineer on Substack. It hit #1 in tech within four months, crossed one million subscribers in three and a half years, and generates $1.5M+ annually — entirely from reader subscriptions, with zero ads or sponsors.

Stratechery: The Paid Newsletter Pioneer Who Proved the Model from a Taipei Apartment

Ben Thompson launched Stratechery full-time from his Taipei apartment in 2014 with a three-tier subscription model. Twelve years later, he has 40,000+ paid subscribers, $5M+ annual revenue, and created Aggregation Theory — the most influential business framework in tech analysis since Clayton Christensen. Substack's seed-round pitch was literally 'Stratechery-in-a-box.'

The Hustle: From a Hot Dog Stand to a $27M SaaS Acquisition

Sam Parr went from selling hot dogs in Nashville to building a 1.5M-subscriber business newsletter, then sold it to HubSpot for eight figures. The buyer didn't want the content — they wanted the mailing list as a SaaS lead funnel.

TLDR: How a Basement Side Project Became an Eight-Figure Ad-Only Newsletter Empire

Dan Ni — Yale math-econ grad, former Jane Street quant trader — left Wall Street after a rare medical condition, and launched TLDR from his parents' basement in Missouri with $50/day in Reddit ads. Eight years later: 13 verticals, 7.2 million subscribers, 22 fully remote employees, and eight-figure annual revenue — all from advertising, not a cent from readers.

FLUX: The Image Model Family Built by Stable Diffusion's Original Team, from 12B to a Self-Flow World Model

FLUX is Black Forest Labs' image-model family. The Stable Diffusion team launched it in August 2024 with a 12B rectified-flow transformer. Two years later it spans klein 4B ($0.014 and the only current Apache-2.0 model) / 9B, pro ($0.03), flex ($0.05), max ($0.07 with live web grounding), and open-weight 32B dev. FLUX 3 extends Self-Flow to video, synchronized audio, and robot actions. This guide covers the FLUX.1-to-FLUX 3 evolution, three-tier licensing, and model selection.

Speech and Audio Models: Four Years from Whisper Rewriting Open ASR to ElevenLabs Consolidating Voice APIs

Speech models split into two lines. Whisper led ASR: its MIT-licensed 1.55B model drove transcription cost toward zero in September 2022; after v2, v3, and turbo cut the decoder from 32 layers to four, OpenAI moved to closed gpt-4o-transcribe. In TTS, ElevenLabs grew to an $11B valuation and $500M ARR, while Kokoro (82M, Apache-2.0) and Chatterbox preserved self-hosting. Speech-to-speech Realtime APIs are now rewriting live conversation.

Video Generation Model Families: Sora Exits after a Two-year Arms Race Dominated by Veo 3.1, Kling 3.0, and Gen-4.5

Sora's February 2024 preview shocked the industry, but the landscape reversed in two and a half years: OpenAI closed the consumer Sora app in April 2026 and scheduled its API for retirement on September 24; Veo 3.1 became the narrative default with native audio and Flow; Kling 3.0 became a unified multimodal model with $240M annualized revenue; and Runway Gen-4.5 briefly led Artificial Analysis in a November 2025 snapshot while defending the professional market through enterprise workflows. This guide compares four families by generation, specifications, pricing, and use case.

When Search Returns Only 10 Results: Fixing CJK Recall in Cloudflare D1 FTS5 Hybrid Search

Querying “認證” returned only ~10 hits while 149 files (509 occurrences) matched; 41 posts and 76 chunks were found via LIKE, but D1 chunks_fts had 0 rows and unicode61/trigram both returned 0 for 2-char CJK terms. Fix: LIKE fallback with char-level OR first, then trigram migration + pnpm sync, then pagination beyond the hard limit of 12.

MiniMax: The Chat App Company That Built a Coding Model to Rival Frontier Labs

MiniMax started as a consumer chat app company, then M2.5 scored 80.2% on SWE-bench Verified at 1/10-1/20 the cost of Claude Opus; M3 (456B total / 45.9B active) became the first open-weight model to clear 59% on SWE-bench Pro, with 1M context powered by their novel Sparse Attention mechanism.

Nous Research: From Research Collective to Open-Source AI Ecosystem Rebel

Nous Research doesn't pretrain — they fine-tune and do RL. Hermes 4 scores 96.3% on MATH-500, NousCoder-14B improves Qwen3-14B's coding ability by 7% using only 24K training samples. But the real moat is Hermes Agent: 236K GitHub stars, #19 globally, 3,000 contributors.

Ornith: The Open-Source Coding Dark Horse Built on Self-Improvement RL

DeepReinforce's Ornith 1.5 family, trained with self-improvement RL: the 397B flagship scores 86.0 on SWE-bench Verified, matching Claude Opus 4.8; the 35B-A3B activates only 3B parameters per token yet leads every coding benchmark in its class; the 9B runs on phones. MIT-licensed, fully open-source.

Managing Multiple Claude Code Sessions: Agent View, Dispatch, State Monitoring, and Cross-Session Messaging

`claude agents` gives you one screen listing every background session, with Needs input, Ready for review, Working, Completed, and related states managed in one table. Press Space to peek, Enter to attach. Combined with cross-session messaging (ListAgents / SendMessage, v2.1.224+), sessions can also message each other; same-machine delivery uses a local socket and never touches Anthropic servers.

The .claude Directory, Explained: settings, rules, skills, and auto memory

Claude Code splits its configuration across the project `.claude/` folder and your home directory — 20+ file locations. Only two mental models matter: settings merge across layers and are enforced; CLAUDE.md and rules concatenate into context as guidance. And every file is committed, gitignored, or Claude-written.

How Claude Code Reviews Your PRs: Multi-Agent Analysis, REVIEW.md, and ultrareview

After GitHub PR review is configured, a fleet of agents reviews PRs according to the repo's trigger mode — 20 minutes on average, about $15–25 per review, with findings posted as inline comments on the offending lines. For larger changes, /code-review ultra launches a cloud deep review that reports independently verified bugs in 5–10 minutes at roughly $5–25 per run; Pro/Max plans include 3 free runs.

Managing Claude Code Costs: Token Tracking, Model Choice, Effort, and Team Analytics

Claude Code costs accumulate with context size: enterprise deployments average ~$13 per developer per active day and $150–250 per month. This post covers /usage and /insights tracking, six token-saving tactics, and a systematic answer to 'which model should I use': provider-dependent model aliases, effort levels, fast mode ($10/$50 per MTok for Opus 5/4.8), and the advisor tool.

Claude Code Config Not Taking Effect: Diagnosing with /context, /doctor, /mcp, and Error References

When CLAUDE.md rules are ignored, hooks never fire, or an MCP server shows no tools, the file usually didn't load, loaded from an unexpected location, or got overridden. This guide covers what the diagnostic entries (/context, /memory, /skills, /doctor, /mcp, and more) actually show, safe-mode bisection, and a table of six high-frequency error messages with fixes.

How to standardize Claude Code dev environments: devcontainer.json, CI consistency, and team rollout

Add Anthropic's official Dev Container Feature (`ghcr.io/anthropics/devcontainer-features/claude-code:1.0`) to `.devcontainer/devcontainer.json` and three steps—write the config, rebuild the container, run `claude` to sign in—put every teammate's Claude Code behind the same container definition. The same definition feeds GitHub Codespaces and CI; a five-step rollout gets the whole team there.

How Claude Code Orchestrates Subagents at Scale: Dynamic Workflows, ultracode, and Rerunnable Scripts

Dynamic workflows let Claude write multi-agent orchestration as a JavaScript script that a runtime executes in the background — up to 1,000 agents per run, savable as a /<name> command. This piece covers trigger methods, the save-and-rerun flow, three fit scenarios (codebase audit, large migration, cross-checked research), and where they differ from Agent Teams.

How Claude Code Works: The Agentic Loop, Built-in Tools, and Two Safety Rails

Claude Code runs an agentic loop — gather context, take action, verify results — until the task is done. This entry to the series breaks down its five tool categories, the model/harness split, and the two safety rails: checkpoints and permission modes.

How to Choose Claude Code's Multi-Agent Options: Subagents, Agent View, Agent Teams, Dynamic Workflows

The official docs split Claude Code's parallel work into 4 approaches: subagents delegate inside one session, agent view lets you supervise background sessions yourself, agent teams coordinate workers through a lead, and dynamic workflows run scripted fleets of subagents with cross-checks; file collisions are always handled by worktrees. Includes a translated comparison table and a three-question decision guide.

Claude Code in the Cloud: on the web, --cloud/--teleport, and Steering from Mobile

Claude Code on the web runs tasks in cloud environments, Anthropic-managed VMs by default or self-hosted environments when routed there: authorize GitHub, dispatch from browser or mobile, start cloud sessions with --cloud, and pull them back local with --teleport. Research preview on Pro/Max/Team; no separate compute charge, but rate limits are shared.

How Much Autonomy to Give Claude Code: Permission Modes, the Auto Mode Classifier, and Allow/Deny Rules

Claude Code ships six permission modes; day to day you cycle Manual, Accept edits, Plan, and Auto with Shift+Tab. On Pro/Max/Team plans, eligible interactive terminal and VS Code sessions start in auto mode by default, with a background classifier reviewing most actions and blocking force pushes, `curl | bash`, production deploys, and more by default. This post covers the four-mode spectrum, permission rule syntax, and organization-level trust config.

How prompt caching shapes Claude Code's speed and bill: prefix matching, invalidation triggers, and hit rate

Claude Code's prompt caching works by exact prefix matching: a cache read bills at roughly 10% of the standard input rate, but switching models, changing effort, enabling fast mode, toggling MCP servers, or denying an entire tool forces the next turn to reprocess everything. The TTL defaults to five minutes; the main conversation and a few helper requests on a subscription get one hour.

How to manage Claude Code sessions: --continue, --resume, /branch, and JSONL transcripts

Claude Code writes every session line by line to a JSONL file under ~/.claude/projects/, kept for 30 days by default. This post breaks down --continue vs --resume, session naming rules, /branch fork semantics, and transcript export and cleanup settings.

Claude Code install and login troubleshooting: PATH, install sources, proxy, OAuth callback

Work through install and login failures in five steps: verify PATH on your OS, confirm there is only one installation, test downloads.claude.ai for a 200, recover failed OAuth callbacks by pasting the login code or using claude auth login, then finish with claude doctor.

Claude Code Runtime Troubleshooting Guide: CPU/Memory, Session Hangs, Auto-Compact Thrashing, Tables, Search Failures

Five classes of runtime fixes: diagnose high memory with /compact plus /heapdump; recover a hung session with Ctrl+C and claude --resume; write large tables to files instead of forcing terminal output; escape autocompact thrashing by reading files in chunks or running /compact with a focus; fix broken search by installing system ripgrep and setting USE_BUILTIN_RIPGREP=0.

Agentic / Reasoning RAG: From Search-R1's RL Multi-Turn Search to Deep Research and MCP's Reasoning × Retrieval Paradigm

In 2025 RAG stopped being 'retrieve once, generate once.' Search-R1 trains models to search autonomously in multiple turns with RL, REX-RAG/AlignRAG add policy and alignment branches, OpenAI Deep Research productizes the loop, and MCP generalizes retrieval into unified tool invocation. This post unpacks the design philosophy, trade-offs against ten generations, and when to adopt the new paradigm.

ai deep-dive

Apple Opens Free Private Cloud Compute Access: AFM 3 and What Developers Need to Know

Apple is giving App Store Small Business Program developers free access to AFM 3 models on Private Cloud Compute if their apps have fewer than two million first-time downloads. The five-model family includes the sparse 20B-parameter AFM 3 Core Advanced, which activates only 1–4B parameters on-device, and AFM 3 Cloud Pro on Google Cloud NVIDIA GPUs, refined with outputs from Gemini.

ai deep-dive

BytePlus ModelArk Coding Plan: ByteDance's AI Coding Subscription

BytePlus ModelArk Coding Plan offers Lite ($10/month) and Pro ($50/month) subscriptions covering models such as DeepSeek-V4, GLM-5.2, and Seed-2.0 in tools including Claude Code and Cursor. Lite includes about 24,000 requests per month; Pro includes five times as many.

Learning Agent Design from Mature Coding Agents (2): The Shape of the Agent Loop — Event Streams, Checkpoints, Resume

pi's loop is a double while-loop wrapped in an EventStream; claude-code's source openly says stop_reason is unreliable and uses tool_use blocks observed during streaming as the sole continue signal; codex models a turn as a cancellable SessionTask and records sessions with a dedicated rollout crate. looplane chose an ordering — manifest first, JSONL second — that turns Ctrl-C into verified resumption instead of a rerun. All evidence cited at file#symbol level.

Learning from Mature Coding Agents (13): CLI Ergonomics — Make New Tools Feel Already Familiar

Mature coding-agent CLIs have converged on the same conventions: positional prompt, -p means print, exec is headless, resume is a first-class command, -C changes directory; looplane inherits this vocabulary directly, driving learning cost close to zero.

Learning Design from Mature Coding Agents (10): Edit Tool Trade-offs — unified diff, exact edit, hashline, and whole-file

LLMs break unified diffs on bookkeeping: wrong hunk counts, hallucinated context lines. The five reference projects split into two camps — simplify the diff grammar (Codex drops line numbers), or drop diffs entirely (Claude Code/Pi/OpenCode exact replace); OMP goes further by binding read state into the format via hash anchors. looplane took the minimal-intervention path: keep the guarded apply_patch, add a zero-fuzzy replace_text, and its qwen3:4b eval went from stable failure to 5/5.

Learning Agent Design from Mature Coding Agents (9): External CLIs as a Backend — Where Does the Security Boundary Go?

Every mature coding agent ships a machine interface: codex has `exec --json` plus a full app-server JSON-RPC protocol, claude-code has `-p` with stream-json, and pi/opencode/omp each expose a JSON event stream. Wrapping these CLIs as your backend is the fastest path to subscription-backed coding — but they own their agent loop, their login, and their permission model. looplane's answer: let the external CLI fully own its loop while looplane holds only three things — an isolated working copy, patch audit, and final verification. One runtime never impersonates another.

Learning Design from Mature Coding Agents (22): The Gateway Pattern — Turning Any Provider into an OpenAI-Compatible Endpoint

The ecosystem treats /v1/chat/completions as the lingua franca, but your providers don't all speak it. The five reference projects split into three camps: pi and OpenCode make the client speak every dialect natively so no gateway is needed; OMP builds a real protocol translator (foreign wire → neutral context → provider adapter, no raw passthrough); Codex and Claude Code run proxies that translate nothing and exist purely to force traffic through a controllable path. Looplane copies OMP's boundary but narrows it to one wire in, one out: strictly parse OpenAI Chat into a canonical contract, then dispatch to any ModelProvider — and along the way hit a cross-event-loop client-close bug whose lesson is that provider lifecycles belong to the ASGI lifespan, not the signal handler.

Learning Design from Mature Coding Agents (21): Headless Mode and CI Usage — When Nobody Can Click Approve

The biggest problem when an agent enters CI is approval: no TTY, nobody to click approve. The five reference projects converge on two strategies — delegate permission decisions to the calling program (claude-code's control protocol), or replace approval semantics entirely (codex defaults to Never plus sandboxing, opencode auto-rejects). looplane keeps one AgentRunner loop and injects a different ApprovalPolicy: headless uses HeadlessApprovalPolicy, which never reads stdin so it cannot hang the pipeline, and denies EXECUTE by default — fail closed.

Learning from Mature Coding Agents (14): Onboarding Design — Provider-Aware Init and Instant Verification

A blank config file drives people away; a bad credential discovered too late drives them away faster. All five mature agents treat setup as a first-class state, and looplane adds the step most of them skip: verify the key right after saving it.

Learning Design from Mature Coding Agents (20): The Run Artifacts Contract—What Makes a Run Auditable After It Ends?

After an agent run finishes, 'the model said it's done' is not evidence. Codex splits traces into a manifest + JSONL + payloads bundle, omp mirrors on-disk files into SQLite, pi indexes native session files with runs.jsonl. Looplane picked the strictest option: six fixed files per run, the run is incomplete if any is missing, and patch review reads changes.patch—not anyone's verbal claim.

Learning from Mature Coding Agents (16): Runtime Abstraction and Capability Handshake

Five external CLIs expose five different machine interfaces: JSONL event streams, JSON-RPC handshake, HTTP API, ACP, stream-json. The right way to support them is not one interface that pretends they're identical — it's a narrow runtime boundary plus an honest capability matrix. Availability means installed, not authenticated; protocol drift fails closed.

Learning from Mature Coding Agents (11): Sandboxes and Remote Execution — Deploying on Cloudflare Sandbox

A local sandbox limits the blast radius of an agent on your machine; a cloud sandbox is about moving code safely onto someone else's machine. All five mature projects solve the first problem; only looplane actually deployed the second. Lessons from production: mocks can't catch SSE framing, green CI can't catch a stale wheel, and cleanup paths deserve timeouts just as much as success paths.

Learning Agent Design from Mature Coding Agents (19): Session Persistence and Crash Recovery — Rescuing State After the Agent Dies

All five agents store sessions as append-only JSONL plus some form of single-writer protection, but crash recovery lives in the details: pi repairs torn tails, codex reopens and retries after write failures, and looplane picked a 'manifest first' ordering that reduces the only crash window to one repairable slot. This post dissects each project's write ordering and fail-closed conditions, all cited at file#symbol level.

Learning from Mature Coding Agents (12): Can Small Models Code? — Capability Boundaries and Eval Discipline

Small models don't fail at reasoning first — they fail at format stability: tool-call JSON, diff hunk arithmetic, and context budgets all break. The mature harnesses build evals on real model behavior (pi's model-backed evals, OMP calibrating benchmarks from real session logs, Codex even relaxing its parser for weaker models). looplane picks the narrowest but hardest path: one fixture, five real Ollama runs, a manifest declaring exactly which files and patch fragments count as success — and M2's failure kept verbatim as evidence. Never pass mock off as E2E; never spin partial success into full passes.

Learning Design from Mature Coding Agents (17): Startup Performance and Engineering Discipline — It Was Never the Language

A CLI tool pays its startup cost on every invocation, and performance optimization without a baseline means no regression protection. codex uses daemon reuse and skill snapshot caches; claude-code splits its entrypoint into dynamic imports plus a built-in startup profiler; opencode and omp each maintain lazy-loading discipline; pi does none of it and leans on Bun being fast. looplane is Python — slow by birth — so it applies the full discipline: lazy imports, single-flight disk cache, background controller prewarming, and hyperfine paired benchmarks wired to a CI gate that fails on >10% regression.

Learning Design from Mature Coding Agents (8): The Right Way and the Wrong Way to Use Subscriptions — OAuth and Credential Boundaries

The five reference projects split into three camps on subscription auth. Codex and Claude Code implement OAuth only for their own official clients and store tokens in the OS keyring. pi and OMP directly reuse Claude Code's client ID to implement Pro/Max OAuth — technically feasible, but Anthropic's docs explicitly bar third parties from offering claude.ai login without approval. OpenCode removed its bundled Pro/Max plugins entirely, the cleanest policy precedent in the ecosystem. Looplane's rules: own your grant, never scrape another CLI's credentials, accept third-party OAuth only when the provider clearly supports it, and never copy or forward credentials.

Learning Agent Design from Mature Coding Agents (24): Testing a Moving Agent — fake-CLI Contracts, Recorded Streams, TUI Pilot

An agent's two dependencies — the LLM and external CLIs — are both non-deterministic, but mature projects separate 'the moving parts' from 'the shape of the boundary': codex fakes the Responses API with wiremock plus a scripted SSE server and pins its TUI with insta snapshots; opencode built a VCR-style http-recorder package; pi splits model-backed evals from unit tests into two vitest configs; omp wraps its edit benchmark itself in unit tests. looplane stacks four layers against external CLIs: unit tests, fake-CLI contract tests, recorded-stream integration proofs, and Textual pilot TUI tests. The methodology in one line: record real non-deterministic output, then make deterministic assertions about it.

Learning Design from Mature Coding Agents (15): From Full-Screen TUI to Semantic Transcript

Mature coding agent TUIs never print the event stream directly — they build a typed projection layer first and update it in place. looplane took three steps (full-screen composition, runtime-first dual modes, removing the Ask/Agent split) before two old constraints — non-streaming output and resume-without-replay — were truly lifted.

Learning Agent Design from Mature Coding Agents (5): The Verification Gate — Changed Files Isn't Success, Verified Is

None of the five reference projects enforces 'all declared verification commands pass' at the harness level: pi leaves verification to the model, OpenCode and Codex put it in the system prompt, Claude Code uses a separate adversarial verifier subagent but as a soft contract, and only OMP's cleanse actually runs checks from harness code. looplane takes the hardest path: if files changed, every declared verification command must pass before terminal_reason=verified; with no changes, checks don't rerun (no_changes). Whether to verify is decided by code, not by the model.

Why Python: The Cost and Compensation of Language Choice for Coding Agents

None of the five mature coding agents use Python — pi/opencode/claude-code run on TypeScript, codex rewrote TS into Rust, omp bolted ~80k lines of Rust native crates onto its hot path. looplane still chose Python; the costs are startup performance and packaging, compensated by lazy imports, uv, and Cloudflare Sandbox.

Which Graph RAG to Choose: GraphRAG v3.1.2 vs LightRAG vs HippoRAG 2 — Design, Cost, and Selection

Same 'knowledge graph + retrieval' label, three different bets: Microsoft GraphRAG v3.1.2 pays indexing cost for global summarization, LightRAG cuts cost with dual-level retrieval and incremental updates, HippoRAG 2 turns RAG into growing associative memory via PPR — this guide splits the trade-offs by component with four query modes, indexing pipelines, and a selection matrix.

Late Chunking vs Contextual Retrieval: Encode-First Zero-Cost Context vs LLM-Prefix Precision and Cost

Anthropic Contextual Retrieval uses an LLM to prefix each chunk with 50-100 tokens, cutting failure rate from 5.7% to 1.9% with rerank at ~$1.02/1M tokens; Late Chunking encodes the full 32K-window document first then mean-pools by chunk boundaries for zero extra LLM cost — the trade-off is window, latency, and update shape.

ai guide

The Complete Unsloth Guide: Fine-Tune and Run LLMs Locally, Faster

Unsloth is the fastest, most VRAM-efficient local LLM fine-tuning tool — 2× training speed and 70% less VRAM. In 2026 it added a Desktop app that bundles inference, training, image/video generation, web search, and agent integration into a complete local AI workstation.

Apple Foundation Models: Privacy-first Ecosystem AI with a 20B Sparse Model on Phones

Apple Foundation Models (AFM) is Apple's closed-ecosystem AI family. It evolved from a 3B dense model with LoRA adapters in 2024 into five models in 2026. AFM 3 Core Advanced runs a 20B IFP sparse architecture on phones while activating only 1–4B parameters; Cloud Pro runs on Google Cloud NVIDIA GPUs and is refined through Gemini distillation. There is no public API price or third-party benchmark, and access is limited to Apple's Foundation Models framework.

tech guide

Choosing Mac Remote Desktop: Tailscale, Jump Desktop, RustDesk, and Built-in Screen Sharing

No budget needed: Tailscale plus built-in Screen Sharing is free and the most reliable stack for daily remote work on Mac — upgrade to Jump Desktop only if you want it smoother. This guide compares four options and fixes lid-close and sleep pitfalls in 5 minutes.

Self-Hosted Inference Overview: When Running Your Own Models Makes Sense

The key question in self-hosted inference isn't how fast the engine is — it's your GPU utilization. A fully saturated A100 costs ~$0.70 per million output tokens; at 10% utilization that becomes $7, more than most cloud APIs. This overview maps seven tools across three layers to help you decide which layer you need.

TensorRT-LLM: The Compile-for-Performance NVIDIA-Only LLM Inference Engine

TensorRT-LLM is NVIDIA's open-source LLM inference library (Apache 2.0). It offline-compiles model weights and compute graphs into optimized TensorRT engines, then serves them with custom CUDA kernels, in-flight batching, and multi-dimensional parallelism. The cost: NVIDIA GPUs only, compilation takes tens of minutes, and switching models or quantization means rebuilding.

TGI: HuggingFace's LLM Inference Server, and Why It Entered Maintenance Mode

Text Generation Inference (TGI) is HuggingFace's own LLM inference server, built in Rust and Python. It pioneered continuous batching and Flash Attention in open-source inference engines. The GitHub repository was archived on March 21, 2026, and HuggingFace recommends migrating to vLLM or SGLang. TGI still matters: it defined the architectural baseline that successor engines inherited, and many HuggingFace Inference Endpoints still run it.

2021 AI Conference Guide: Computer Vision

2021 was the year Transformers decisively entered computer vision: Swin Transformer won the ICCV Best Paper award, DINO showed that a self-supervised ViT could learn object segmentation without labels, and NeRF grew from one paper into an entire subfield. CVPR and ICCV both moved fully online because of the pandemic, yet the work published that year shaped architectural choices across computer vision for years to come.

2021 AI Conference Guide: Machine Learning

2021 was the year diffusion models surpassed GANs, self-supervised learning made theoretical breakthroughs, and reinforcement learning confronted weaknesses in its evaluation methodology. NeurIPS received a then-record 9,122 submissions, ICLR’s Score-Based Generative Modeling paper became a theoretical foundation for the diffusion ecosystem, and ICML delivered substantial work on optimization theory and the dynamics of self-supervised learning.

2021 AI Conference Guide: Natural Language Processing

2021 marked NLP’s shift from fine-tuning an entire model to adapting only a small fraction of its parameters. Prefix-Tuning at ACL, LoRA on arXiv, and Prompt Tuning at EMNLP all appeared that year; ACL Rolling Review launched; and the Findings track established itself as a second publication channel.

What AI Conferences Published in 2021: Transformers Spread, Self-Supervised Learning, and the Start of Diffusion

2021 was a dividing line for major AI conferences. Transformers spread from NLP throughout computer vision and time-series research, self-supervised learning became the most common cross-conference theme, and a diffusion model won an ICLR Outstanding Paper award before anyone realized it would displace GANs. Meanwhile, GNNs and federated learning reached historic peaks in paper volume before beginning to decline.

2022 AI Conference Guide: Computer Vision

2022 marked computer vision’s turn from recognition toward generation. Latent Diffusion Models appeared at CVPR and led to Stable Diffusion; NeRF research jumped from 25 papers in 2021 to more than 50 at CVPR alone; ConvNeXt mounted a compelling counterattack for CNNs; and ECCV in Tel Aviv set a record with 157 oral papers.

2022 AI Conference Guide: Machine Learning

2022 was the year diffusion models took center stage, Chinchilla scaling laws rewrote large-model training, and Chain-of-Thought turned reasoning into an ability that prompts could elicit. NeurIPS passed 10,000 submissions; three of its 13 Outstanding Papers directly concerned diffusion; and Chinchilla and data pruning both challenged the belief that bigger was always better. On the eve of ChatGPT’s release, every required piece fell into place at that year’s conferences.

2022 AI Conference Guide: Natural Language Processing

2022 marked NLP’s shift from demonstrating model capabilities toward aligning and controlling them. InstructGPT brought RLHF into the mainstream, Chain-of-Thought showed that prompts could unlock reasoning, and Flan 2022 matured instruction-tuning methodology. ACL and NAACL adopted ARR as their sole review path, exposing infrastructure and reviewer-load problems. ChatGPT launched at year-end and rewrote the rules of NLP research.

What AI Conferences Published in 2022: The Diffusion Boom, Chain-of-Thought, and the Eve of ChatGPT

2022 was a turning point at major AI conferences. Diffusion models moved from emerging to mainstream, with two NeurIPS Outstanding Papers; Chinchilla rewrote scaling laws; Chain-of-Thought showed that large models could reason; and InstructGPT used RLHF to teach language models to follow instructions. When ChatGPT launched at year-end, these academic topics instantly became global news.

A Guide to the Top AI Conferences of 2023: Computer Vision

In 2023, computer vision moved from seeing images to understanding, generating, and controlling them. Segment Anything turned segmentation into a general zero-shot capability, ControlNet made diffusion models precisely controllable, and 3D Gaussian Splatting challenged NeRF with real-time rendering. CVPR received more than 9,000 submissions and ICCV more than 8,000 as both conferences returned to in-person events.

A Guide to the Top AI Conferences of 2023: Machine Learning

In 2023, LLMs took over the machine-learning conference agenda. NeurIPS received more than 12,000 submissions; both Outstanding Papers addressed large models, while runner-up DPO became a practical alternative to RLHF within two years. DreamFusion opened the text-to-3D field, ICML spotlighted LLM watermarking and learning-rate adaptation, and the Mamba preprint emerged as the first serious architectural challenger to the Transformer.

A Guide to the Top AI Conferences of 2023: Natural Language Processing

2023 was the first full academic year after ChatGPT, and LLMs rewrote the NLP conference agenda. ACL's Best Papers examined humor understanding and the propagation of political bias; an EMNLP Best Paper explained in-context learning through information flow; and the HackAPrompt competition paper also won an EMNLP Best Paper award, signaling that security research had entered the mainstream. The year's largest shift was from asking how to make models more accurate to asking how we can tell when a model is misleading us.

What Topics Dominated the Top AI Conferences of 2023? The Year LLMs Rewrote the Research Agenda

2023 was the first year in which LLMs comprehensively rewrote the AI research agenda. DPO received a NeurIPS Outstanding Paper Runner-Up award, ReAct became an ICLR Oral, and hallucination grew from a marginal term into a major track at every conference. Meanwhile, 3D Gaussian Splatting swept through computer vision after its SIGGRAPH debut, Mamba emerged at the end of the year to challenge the Transformer attention monopoly, and publication volume for traditional NLP pipelines began a clear decline.

2024 AI Conference Review: Computer Vision

In 2024, 3D Gaussian Splatting took over 3D reconstruction, video generation moved from research toward products, and vision-language models spread into specialized domains. CVPR received a record 11,500-plus submissions; its Best Papers were Google Research's Generative Image Dynamics and the UCSD/Google collaboration Rich Human Feedback for Text-to-Image Generation. ECCV gave its Best Paper award to Columbia's Minimalist Vision with Freeform Pixels, an unconventional return to the physics of optics.

2024 AI Conference Review: Machine Learning

ML conference submissions exploded in 2024: NeurIPS received a record 15,671 papers, while ICML and ICLR passed 9,000 and 7,000. Research shifted from training ever-larger models toward spending inference compute more intelligently, making test-time compute scaling the year's defining new direction. VAR beat diffusion with next-scale image prediction, Rectified Flow became the theoretical foundation for Stable Diffusion 3, and ICLR gave its inaugural Test of Time Award to the original VAE paper.

2024 AI Conference Review: Natural Language Processing

NLP conferences redefined themselves under LLM dominance in 2024. ACL made open science its annual theme, and four of its seven Best Papers probed fundamental limits of language models. EMNLP turned toward multilingual and cross-cultural work, with Best Papers spanning speech representations and gradient interpretability. ACL and EMNLP received more than 10,000 submissions combined, but the deeper anxiety was what remains of NLP when LLMs can perform nearly every traditional NLP task.

What Top AI Conferences Accepted in 2024: The Year of Agents and the Scaling Debate

The defining conference keywords of 2024 were agents, alignment, multimodal LLMs, and inference-time compute. The LLM share at five major conferences doubled again after its sharp 2023 rise; agent-related terms grew 4.3 times; and diffusion models graduated from an emerging topic to a second generative-AI pillar alongside LLMs. Traditional task-oriented NLP continued to contract, while GANs almost disappeared from top venues.

2025 AI Conference Review: Computer Vision

2025 was a two-conference year for computer vision, with CVPR and ICCV both taking place. CVPR received a record 13,008 submissions; Best Paper VGGT turned 3D reconstruction from iterative optimization into feed-forward inference. ICCV's Marr Prize went to BrickGPT, which generates brick structures from text that can actually be assembled. 3D Gaussian Splatting displaced NeRF, video generation moved toward products, and flow models began replacing diffusion, completing several paradigm shifts in one year.

2025 AI Conference Review: Machine Learning

ML conferences broke every submission record in 2025 and pushed peer review to its limit. NeurIPS received 21,575 papers and used more than 20,000 reviewers; ICML passed 12,000 for the first time, and ICLR reached 11,565. Reasoning and agents were the strongest trends. One NeurIPS runner-up, the conference's only perfect-score paper, challenged whether RLVR creates new reasoning ability. Awards for Alibaba Qwen's Gated Attention and a mechanistic theory of neural scaling laws showed a community moving from scaling at all costs toward understanding why scaling works.

2025 AI Conference Review: Natural Language Processing

NLP conference submissions nearly doubled in 2025: ACL received 8,360 papers and EMNLP 8,174. China-based first authors exceeded 51% at ACL, and DeepSeek's Native Sparse Attention won Best Paper. The deeper story was an identity crisis: an ACL president said 'ACL is not an AI conference,' a quantitative study asked 'Has ACL Lost Its Crown?', and EMNLP faced questions about what still distinguished it from ACL or NAACL.

What Top AI Conferences Accepted in 2025: The Agent Breakout and Reasoning Revolution

The two strongest signals at AI conferences in 2025 were reasoning papers jumping from 47 to 216, a 4.6-fold rise, and agent-related terms exceeding 150 papers with 4.3–11-fold growth. Diffusion moved from breakout topic to infrastructure; RAG became a mainstream enterprise architecture with unusual coverage across all five conferences; state-space models and world models began tracing the early 2020–2021 path of Vision Transformers. Pure prompt-engineering papers encountered reviewer fatigue.

Submitting to Top AI Conferences as an Independent Researcher: A Reality Check

Publishing at a top conference as an independent researcher is possible, but the numbers are harsh: single-author papers have fallen to a single-digit share, the average author count has risen from 3 to 5, and the top 20 institutions account for 35-50% of authorships. Andreas Madsen spent eight months working without pay, earned an ICLR Spotlight, and still ended up returning for a PhD. This article examines real cases, evidence of review bias, and viable paths for researchers without a large lab behind them.

ai guide AI 頂會導讀

What Happens to a Top-Conference Paper from Submission to Publication

An AI conference paper passes through anonymized submission, format screening, reviewer bidding and assignment, independent scores from 3-4 reviewers, an author rebuttal, AC/SAC/PC decisions, and camera-ready revision—a process lasting about 4-5 months. ACL-family conferences add ARR's rolling-review model, in which review comes before the author commits the paper to a venue.

Main Track, Findings, and D&B Track: Three Publication Routes at Top AI Conferences

A paper submitted to a major conference can follow three very different routes: the Main Track is the highest-threshold formal publication, Findings is the ACL family's companion venue for solid work that misses the main program, and NeurIPS created the D&B Track specifically for datasets and evaluation methodology. Their review standards, prestige, and career signals differ enough that understanding the route matters before writing the paper.

Who Submits to Top AI Conferences: Labs, Companies, and the Global Map

The institutional map of top AI conferences is being rapidly redrawn. Industry labs dominate frontier model development—nearly 90% of notable models came from industry in 2024—but academia remains the largest source of highly cited research. Chinese universities went from challengers to nearly half of NeurIPS paper volume in five years, while OpenAI and Anthropic have nearly vanished from conference author lists. The decoupling of publication volume from research capability is the defining signal.

ai deep-dive

How Marin Trains 535B: Scaling Ladder, MoE Expert Parallel, Harrier Data and Live W&B

Stanford Marin pre-registers a paloma macro-loss of 2.04 with a 5-rung Scaling Ladder at 1% cost, then trains 535B-A23B on 11×GB200 in public with live W&B telemetry — 847 training buckets already show the most teachable frontier run.

tech guide

AI Model Evaluation Sources: How to Judge Whether a Model Is Actually Good

You cannot take model vendors' self-reported scores at face value. This guide covers the most important independent evaluation platforms, domain benchmarks, adoption indicators, and official sources in 2026: what each measures, how to read it, where it is biased, and which figures matter for different use cases.

Claude——From AI Safety Lab to SWE-bench Champion, the Strongest Closed-Source Agent Choice

Claude is Anthropic's closed-source LLM family, known for Constitutional AI training, agent capabilities, and coding performance. In July 2026, Opus 5 scored 96% on SWE-bench Verified to claim the coding crown, while Fable 5 led general capability at 83% on LiveBench. Four tiers (Fable / Opus / Sonnet / Haiku) span $1–$10, making this the only family in the series with zero open weights.

Cohere — The RAG-Native Outlier: How Command, Embed, Rerank, and Aya Fit Together

Cohere is the only family that ships generation, retrieval, reranking, and multilingual as distinct products. Command A runs 256K context on two GPUs at 111B, Embed v4 does mixed image-text retrieval, Rerank v4 handles 32K semi-structured data, and Aya covers 101 languages — a four-piece stack built for RAG. This post breaks down each pillar's positioning, licensing, and selection guide.

DeepSeek: From an MoE Lab to OpenRouter's Most-used Open Model

DeepSeek used MLA and MoE innovations to drive inference costs to an industry low. V4 Flash activates only 13B parameters while approaching frontier-model quality and ranks first by OpenRouter usage. This guide traces V1 through V4, the R1 reasoning branch, and how to choose each version.

Gemini——Google's Native Multimodal Flagship: 1M Context and Scientific Reasoning Champion

Gemini is Google DeepMind's native multimodal LLM family, famed for a 1M-token context window and native video/speech input plus scientific reasoning. 3.1 Pro tops GPQA Diamond 94.1% and ARC-AGI-2 77.1% to claim science-reasoning dual crowns, at $2/$12—1/6 of Claude. 3.7 Flash delivers near-Pro agent capability for $0.75/$3.75.

GLM——From a Tsinghua Lab to a 744B Open-Source Flagship, and GLM-5.3's Cybersecurity Surge

GLM is Zhipu AI (Z.ai)'s open LLM family from Tsinghua's KEG Lab. GLM-5.3 (2026/08) lifts coding +50% over the previous generation, hits 84.5% on CyberGym ahead of Anthropic Mythos 5 and OpenAI GPT-5.6 Sol, and scores 60 on the Artificial Analysis Intelligence Index tied with Kimi K3 for open-source #1. The only frontier open model trained entirely on Huawei Ascend.

GPT——Closed API for Revenue, Open GPT-OSS for Ecosystem: the Unified Routing Platform Behind the World's Largest AI Service

GPT is OpenAI's LLM family, from 117M parameters in 2018 to the three-tier GPT-5.6 Sol/Terra/Luna lineup in 2026, serving 1B+ users and 2M enterprise customers. GPT-5.6 Sol leads LiveBench 81.1%, Terminal-Bench 2.1 88.8%, and Artificial Analysis Coding Agent Index 80 across multiple agentic benchmarks, while OpenAI's first open-weight model GPT-OSS ships under Apache 2.0.

Grok — From a 314B Open-Source Bet to Grok 4.6/Build/Imagine, xAI's Distribution-Driven Catch-Up

Grok is xAI's LLM family: founded July 2023, opened with a 314B MoE under Apache 2.0 in March 2024, and two and a half years later spans Grok 4.6 (500K, $2/$6, four reasoning levels), Grok 4 Fast (2M), Imagine for image/video, and Grok Build for terminal coding — its moat is distribution (X / grok.com / Tesla / Bedrock), not single-model supremacy. This post traces Grok 1→4.6, sub-line positioning, pricing, and licensing traps.

Kimi——From a 200K Long-Context Tool to a 2.8T Open-Source Frontier, and K3's Architectural Leap

Kimi is Moonshot AI's LLM family, born from ultra-long context. Kimi K3 (2026/07) is the world's first open 3T-class model—2.8T params, 104B active, 1M context, scoring 60 on the Artificial Analysis Intelligence Index tied with GLM-5.3 for open-source #1. Its Kimi Delta Attention brings a 2.5× scaling efficiency gain.

Llama——From Open-Source Experiment to the Most Deployed Open LLM, and Meta's Closed-Source Pivot

Llama is Meta's open-source LLM family, with the largest enterprise deployment footprint and the most mature ecosystem. Llama 4 Scout (10M context) and Maverick (17B active / 400B total MoE) are the current open multimodal benchmarks, but Meta pivoted to closed-source Muse Spark in April 2026—Llama 4 is likely the last major open Llama, and its license is not truly open (Llama 4 Community License, separate license required above 700M MAU).

Mistral——Europe's Open AI Challenger: Smaller Models and European Sovereignty as a Different Bet

Mistral is Europe's most successful AI startup, cutting through the market with a 'smaller, faster, cheaper' strategy and European data-sovereignty positioning. Mistral Large 3 is Europe's strongest commercial LLM, Small 4 is the 24B efficiency king, and Medium 3.5 is the open Modified-MIT model optimized for agentic coding. Its moat is not technical scale but the 'European compliance' card.

Qwen: Open Weights at Every Size from 0.8B to 2.4T — How HuggingFace's Download Champion Runs a Two-Track Play

Qwen is the most-downloaded model family on HuggingFace, spanning sizes from 0.8B to 2.4T. In August 2026, Alibaba open-sourced a Max-tier flagship for the first time (Qwen3.8-2.4T-A95B) — but swapped the customary Apache 2.0 license for custom terms. Meanwhile the other new release, Qwen3.8-27B, runs native vision on laptop-class hardware and is the only one shipping under Apache 2.0. This post traces the family from 2023 through generation 3.8, explains how the open line and the commercial line split apart, and helps you pick the right model at each tier.

tech guide AI 模型家族

AI Model Landscape: The 2026 Map You Need

In 2026, AI models span seven major categories and more than 20 subcategories. This introduction to the AI Model Families series maps use cases to models and models to families, with current rankings and selection advice for each use case.

Antigravity CLI: Google Replaces a 100K-Star Open-Source Tool with a Closed-Source Go Binary

At Google I/O 2026, Antigravity CLI (agy) replaced Apache 2.0 Gemini CLI with a closed-source Go binary. Technical upgrades — multi-agent orchestration, native sandbox, millisecond startup — but free tier cut 98%, open-to-closed source, 28-day transition window. Community reaction was sharp.

Grok Build: xAI's Rust Coding Agent That Uploaded Your Repo Before Going Open Source

Grok Build is xAI's Rust coding agent — 845K LOC, 8 parallel sub-agents, Arena Mode. May 2026 beta, July open-sourced (Apache 2.0) — but the direct trigger for open-sourcing was a privacy incident: it silently uploaded entire repos (including SSH keys, .env files) to Google Cloud Storage at a 27,800x traffic ratio. The exfiltration code remains in the binary, disabled only by a server-side flag.

Muse Code: Meta's First Coding Agent, Trading Training Rights for a 20x Discount

In August 2026, Meta Superintelligence Labs released Muse Code beta. Closed-source static binary, Muse Spark 1.2 model, parallel persistent sub-agents with worktree isolation. The biggest controversy is pricing: Standard at $1.25/$4.25 per M tokens, or Contributor at $0.10/$0.20 — 20x cheaper, but your code enters Meta's training pipeline.

How to Pick a Self-Hosted Inference Server: From Ollama to Xinference, Six Tools and Their Trade-Offs

Self-hosted inference servers fall into three layers: execution engine (llama.cpp), serving engine (vLLM, SGLang), and model management platform (Ollama, Xinference, Triton). Picking the right layer matters more than picking the right tool — ask where your bottleneck is before deciding where to add complexity.

Xinference: One Platform to Manage LLM, Embedding, Speech, and Image Models

Xinference wraps vLLM, SGLang, llama.cpp, Transformers, and MLX under a single management layer, using a Web UI and OpenAI-compatible API to manage LLMs, embedding, rerank, speech, and image models — suited for self-hosted deployments that need multiple model types to coexist. But the management layer's parsing logic also creates a larger attack surface than pure serving engines (CVE-2026-61539 is a case study).

What Is an AI 'Top Conference': Why CCF, CORE and h5-index Disagree

There's no official certificate for being an 'AI top conference.' It's a community consensus built from four independent signals — CCF-A, CORE-A*, a high Google Scholar h5-index, and a low acceptance rate — and those four signals frequently disagree. ICLR being completely absent from CCF's list is a live example.

tech deep-dive

Agent Platform Deep Dive (8) — Context/Memory and Cloudflare Deployment: Seamless Migration from Local Development to Production

Agent Platform uses a Cloudflare-first architecture: local `npm run dev` runs Node-based simulations, while production maps to Workers + Workers Assets + D1 + KV + R2 + Vectorize + Queues + Workflows + Durable Objects + Workers AI. The Runtime interfaces stay the same (InMemory → Cloudflare implementations), so upper layers migrate without noticing. Deployment requires only `wrangler login` → create resources → fill in IDs → `wrangler secret put` → `wrangler deploy`. CI/CD watches the main branch and runs typecheck + build + dry-run + migration + deploy.

tech deep-dive

Agent Platform Deep Dive (VII)—Evaluation & Quality Gates: Comprehensive Evaluation, Regression Prevention, and an Immune System for Skill Releases

Evaluation is Agent Platform's quality immune system: instead of collecting statistics only after a run, it enforces checks throughout Pre-run, In-run, and Post-run execution. Seven eval categories cover Flow → Step → Skill → Artifact → Evidence → Policy → Regression. A Skill release must pass five gates—Trigger, Functional, Policy, Regression, and Human Review—and any failure blocks it. The Learning Loop moves from Run signals through Proposal, Human Review, Sandbox Eval, Quality Gate, and Publish, under one strict rule: agents propose, humans review, and eval gates decide whether a change can ship.

tech deep-dive

Agent Platform Deep Dive (Part 2) — Flow Runtime: Versioned Flows, Checkpoints, and Resume/Retry Mechanisms

Flow Runtime is the heart of Agent Platform: a Flow becomes immutable when published, each Run is bound to a specific version and preset, Steps move through a DAG according to edge conditions, every boundary saves a checkpoint, and resume/retry-step preserves the complete trace history.

tech deep-dive

Agent Platform Deep Dive (Part 6) — Observability, Evidence, and Artifacts: Structured Traces, Claim-to-Source Lineage, and Versioned Outputs

Observability is a first-class capability, not logging added after the fact: a structured trace connects FlowRun→StepRun→SkillInvocation→ProviderCall→ToolInvocation→GuardResult→EvidenceItem→ArtifactVersion. The Evidence Store traces every claim back to its source, excerpt, citation, confidence, and conflicts. Artifact versioning supports approve/reject/regenerate without deleting history. Context Snapshots allocate token budgets by category and record automatic compression when a block exceeds its budget. Procedural, episodic, and semantic memory can be written only through proposals reviewed by a human.

tech deep-dive

Agent Platform: An In-Depth Look at an Open-Source AI Workflow Control Plane (Part 1)—Architecture and Positioning

Agent Platform turns AI agents from a blank chat window into a structured workflow platform whose behavior can be defined, versioned, observed, verified, and improved. Its built-in Deep Research seed flow demonstrates the complete feedback loop.

tech deep-dive

Agent Platform Deep Dive (Part 5) — Policy Engine: Runtime Guards, Budget Control, Human Approval, and Loop Protection

The Policy Engine acts as the Agent Platform's constitution and enforcement layer: policies are versioned and bound to flows and presets; four guard layers enforce rules at step boundaries; budgets cap cost, tokens, runtime, iterations, and tool calls; external writes require human approval; loop detection trips circuit breakers; and escalation records provide an auditable trail. Rules are configuration-driven, so adding one means changing JSON rather than hard-coded logic.

tech deep-dive

Agent Platform Deep Dive (Part 4) — Provider Router & MCP: Multi-Provider Routing, Fallback Chains, and an OpenAI-Compatible Proxy

The Provider Router is Agent Platform's model and tool gateway: it unifies 30+ providers, MCP tool discovery, step-local permission control, fallback chains with RRF fusion, and an OpenAI-compatible Proxy that existing SDKs can use without code changes. It is configuration-driven rather than hard-coded, with provider-health-aware routing.

tech deep-dive

Agent Platform Deep Dive (3) — Skill System: Versioned Capability Packages, Explicit Binding, and the Learning Loop

A Skill is a versioned, installable, and auditable capability package. Its dual-file architecture separates metadata from instructions, explicit binding replaces model-driven routing, and every invocation is recorded. The Learning Loop turns run signals into proposals, sandbox evaluations, human review, and publication while enforcing the principle: agents propose, humans review, and evals serve as the gate.

tech deep-dive

Groundlane Series Part 1: Why AI Agents Need a Controlled Web Access Layer

Groundlane is an open-source TypeScript remote MCP server (v0.1.0) giving AI agents web_search, web_fetch, and web_extract through a single stable contract, with auth, provider routing, and resource limits kept at the operator boundary.

tech deep-dive

Groundlane Series Part 2: Actual Calls, Response Structures, and Error Boundaries for the Three MCP Tools

Hands-on parameter choices and response structures for web_search (ten adapters, RRF merge, dual-provider default), web_fetch (format/render strategies, finalUrl provenance), and web_extract (CSS selector determinism, no implicit LLM step), with verifiable error boundaries.

tech deep-dive

Groundlane Series Part 3: Comparing with Traditional Approaches — WebFetch, stealth_fetch, puppeteer, and requests

A four-dimension comparison (determinism, replaceability, identity boundary, operational cost) between Groundlane's controlled remote MCP contract and traditional local approaches (WebFetch, stealth_fetch, puppeteer, requests), with verifiable scenario recommendations.

tech deep-dive

Groundlane Series Part 4: In-Site Application — Verified Workflow with Existing MCP Tools and Usage-Mode Rules

Based on the in-site groundlane skill (mcp__groundlane__*) and usage-modes rules, this part describes reproducible steps for reference verification and digest data collection — without assuming unimplemented features or using deprecated stealth_fetch.

tech deep-dive

Groundlane Series Part 5: Pitfalls and Best Practices — timeout, selector, render mode, version-change risk, and security boundaries

A reproducible checklist of the most common operational pitfalls: truncated results from fixed caps, selector errors tied to DOM stability, render-mode cost/determinism tradeoffs, version-change verification (v0.1.0 preview), and the non-negotiable security boundaries (URL policy, auth, concurrency, budget).

tech deep-dive

Building a Taiwan Stock Research Agent (Part 1): Why Taiwan Needs Its Own Research Agent

US-stock LLM agents have attracted nearly 100,000 GitHub stars, yet no Taiwan-stock project has even passed 10. I consolidated three side projects into a Taiwan-stock research agent where every conclusion must first survive a backtest; this article explains why.

tech deep-dive

Building a Taiwan Stock Research Agent (Part 2): LangGraph Parallel Architecture—Five Analysts Working at Once

Five analysts fan out in parallel within one superstep, so latency is max rather than sum; backtesting and reflection stand before synthesis, restricting the LLM to explaining evidence that already exists.

tech deep-dive

Building a Taiwan Stock Research Agent (Part 3): Tiered LLMs and a Degradation Chain—API, Local CLI, and Dictionary Fallbacks

Only two roles call an LLM; every other analyst remains fully programmatic. Each call follows an Anthropic API → local Claude CLI → rules-based degradation chain, and cost accounting trusts only provider-reported values—unknown cost is never treated as $0.

tech deep-dive

Building a Taiwan Stock Research Agent (Part 4): Backtest Accountability—Why Backtests Lie

This project has one core rule: every LLM conclusion must first pass a historical backtest of the same signals. When expectancy is negative, synthesis cannot issue an optimistic verdict. Each of the four traps that make backtests lie has a programmatic countermeasure.

tech deep-dive

Building a Taiwan Stock Research Agent (Part 5): Walk-Forward Evaluation, Run Cards, and an Honest 50% Baseline

I do not measure whether the agent ‘feels accurate.’ I freeze parameters in walk-forward OOS tests, record a hash of every input in run cards, and keep the honest 5/10 = 50% golden-eval baseline so the agent has to admit that it is not accurate yet.

tech deep-dive

Building a Taiwan Stock Research Agent (Part 6): Making Every Number in an LLM Report Auditable

Numbers are the easiest part of an LLM report to hallucinate. I therefore put every trusted number into a SHA-256-addressed evidence manifest and let the LLM cite only {{fact.id}} placeholders. If it writes a bare number, the entire output is discarded and replaced with a deterministic template.

tech deep-dive

Building a Taiwan Stock Research Agent (Part 7): The Copilot Loop—Plan Contracts, Verifiable Sources, and Human Review

A research request first becomes a ResearchPlan that requires human approval. External documents must be fetched in full, and verbatim quotes must be verified before they can enter a report. Quant review is always append-only, and free-text feedback never flows back into a prompt. This is the complete M5 Copilot loop.

tech deep-dive

Building a Taiwan Stock Research Agent (Part 8): The Boundary Between Research and Paper Orders—Content-Addressed Execution Contracts

Three frozen Pydantic contracts weld the boundary between a research artifact and order-placement authority shut: content addressing, eight hard gates, and paper-only execution, while the agent never touches credentials.

tech deep-dive

Building a Taiwan Stock Research Agent (Part 9): Deployment Boundaries—from Docker to a Public API on Cloudflare Containers

The full deployment path for a Python agent, from local uv run to Docker to a public API on Cloudflare Containers: the Worker enforces authentication, the Container runs FastAPI, secrets never enter the image, and the service sleeps automatically after 10 idle minutes—the right way for a side project to save money.

Building an Academic Search Pipeline: The Roles of arXiv, OpenAlex, Crossref, Semantic Scholar, and PubMed

An academic-search pipeline cannot simply concatenate five APIs: use arXiv or PubMed for domain discovery, align OpenAlex and Semantic Scholar records through DOI, PMID, and arXiv IDs, then use Crossref and PubMed relationships to check the version of record, corrections, and retractions.

ai deep-dive

AG2: Organizing Multi-Agent Collaboration with Conversations and GroupChat

AG2 continues AutoGen's ConversableAgent model: agents collaborate through messages, while GroupChatManager selects the next speaker by round robin, manual choice, randomness, or an LLM.

ai deep-dive

Choosing an Agent Framework in 2026: LangGraph, CrewAI, MAF, AG2, Mastra, Pydantic AI, and DSPy

These seven tools are not one product category: LangGraph, MAF, and Mastra emphasize durable workflows; CrewAI and AG2 emphasize multi-agent collaboration; Pydantic AI emphasizes typed Python agents; DSPy optimizes AI programs against data and metrics. Choose the control model first.

Writing Search Queries for Agents: Keywords, Semantic Descriptions, Decomposition, and Rewriting

An agent should not send the user's sentence unchanged to every search service. Classify the need as exact lookup, keyword, semantic, or fielded search; move source, date, language, and field constraints into native provider parameters; then rewrite according to zero-result, overbroad, stale, or source-mismatch symptoms.

ai deep-dive

Amazon Bedrock Deep Dive: Putting Model APIs Inside the AWS Governance Boundary

Amazon Bedrock is more than a reseller for multiple model APIs. It brings model invocation, IAM, Regions, Knowledge Bases, Guardrails, and CloudWatch into one AWS control plane. It fits teams already on AWS that value governance overhead more than the lowest token price.

ai deep-dive

Arize Phoenix: Turning Traces into Datasets, Experiments, and Evaluators

Phoenix is an MIT-licensed open-source LLM observability and evaluation platform. It collects traces with OpenTelemetry and OpenInference, turns production failures into versioned datasets, compares prompt, model, or RAG changes in experiments, then writes code, human, and LLM evaluator scores back as annotations. It is not Arize AX, and self-hosting defaults require security work.

Giving an Agent Access to Logged-In Websites: Sessions, Permissions, and Automation Boundaries

Authenticated browser state is not a convenience setting; it is a credential that can impersonate its owner. Use a dedicated low-privilege account and isolated profile, separate reading from reversible writes and high-risk transactions, and leave MFA plus final submission to a human.

ai deep-dive

Baseten: The Model Inference Lifecycle from Truss Packaging to Autoscaling

Baseten puts custom-model packaging, GPU deployment, inference engines, autoscaling, and release workflows on one platform. Its value is not another OpenAI API, but retaining runtime control while operating less GPU orchestration.

ai deep-dive

Braintrust: Closing the LLM Evaluation Loop from Datasets Back to Production

Braintrust connects versioned datasets, immutable experiments, scorers, and production traces into one evaluation loop. Its value is not another score but the ability to turn production failures into offline tests. The company announced an $80 million Series B in February 2026; its customer list is company-reported.

ai guide

Brave Search API Complete Guide: An Independent Search Index for Agents

Brave Search API exposes five endpoint families—Web, News, Images, Videos, and LLM Context—backed by Brave's own Web index and ranking models. Its core search is not merely a Google SERP wrapper.

ai deep-dive

Browserbase: Turning Agent Browsers into Operable Infrastructure

Browserbase combines remote Chromium, persistent Contexts, proxies, and a Session Inspector in one control plane. It operates browser fleets; it does not decide an agent's next action. As of August 2026, the company reports more than 35 million monthly browser sessions and over 10,000 customers.

ai deep-dive

Cartesia Deep Dive: From Sonic Streaming TTS to a Real-Time Voice Agent Pipeline

Cartesia's core is Sonic real-time TTS, Ink STT, and streaming inference. Although it offers the Line voice-agent platform in 2026, buyers must still separate the model layer from telephony orchestration and design consent, retention, and fallback for cloned voices.

ai deep-dive

Cerebras Inference: Know the Bottleneck Before Putting Wafer-Scale Speed in an Agent Loop

Cerebras can dramatically accelerate generation on supported models, but agent latency still depends on prefill, tool I/O, model quality, and platform compatibility.

ai deep-dive

Chroma Vector Database: From Local RAG to Distributed Retrieval

Chroma manages embeddings, documents, and metadata through collections; it embeds into Python locally, uses HNSW on a single node, and separates compute from storage with object storage, SSD caches, and SPANN in distributed deployments.

ai deep-dive

Claude Code Startup Playbook: Five Operating Principles from Anthropic's Guide

Anthropic interviewed 15 startups and distilled five Claude Code operating principles: everyone ships, automate the tedium, trust but verify, build for rebuilding, prototype to productionize. ClickHouse shipped 30% more features, Clay automated 100% of bug triage, Artemis Security hit 6,000+ PRs per week.

ai deep-dive

Cloudflare Kitesurf: An Agent Browser That Is Not Chromium—and What It Trades for Scale

Kitesurf is a non-Chromium browser backend in Browser Run that remains in beta. It trades pixel compatibility, persistent authenticated sessions, WebGL, and full anti-bot behavior for low CPU and memory through Workers isolates, Rust/Wasm, and stateless components.

Cloudflare Sandboxes Deep Dive: How Workers, Durable Objects, and Containers Form an Agent Runtime

Cloudflare Sandboxes uses a Worker as the entry point, a named Durable Object as the control plane, and a Container inside an isolated VM as the execution plane. It fits Cloudflare-native fleets of ephemeral Linux workspaces, but persistence, security boundaries, and three layers of billing remain your responsibility.

ai guide Reading CMU 07-280

Completing CMU 07-280: What You Know, What Is Missing, and What Comes Next

Finishing 07-280 means more than reading 24 guides: produce a search engine, supervised-model comparison, CNN/GPT-2 experiments, and a small RL-plus-MCTS system before choosing 07-380, 10-301, or a specialist course.

Reading CMU 07-280: Why Search, GPT-2, and AlphaZero Belong in One Course

07-280 is CMU's new Spring 2026 AI+ML core: 24 lectures and 12 main assignments move from heuristic search and CSPs to AlexNet, GPT-2, and AlphaZero. Its public material supports self-study, but complete recordings, Canvas checkpoints, Gradescope, and staff feedback remain unavailable.

CMU 07-280 Lecture 1: The Shared Problem Behind AI, ML, and Representation Learning

Lecture 1 uses an alien autoencoder, the scope of AI and ML, and AI history to establish the course's coordinate system: an intelligent system turns inputs into representations and decisions under uncertainty.

CMU 07-280 Lecture 2: Heuristic Search from UCS and Greedy to A*

Lecture 2 decomposes search into a problem, frontier, and priority: UCS uses paid cost, Greedy uses estimated remaining cost, and A* combines them as `f=g+h`; tree and graph search require different optimality conditions.

CMU 07-280 Lecture 3: Minimax, Alpha-Beta, and Expectimax

Lecture 3 turns a single path into a contingent plan: minimax faces an optimal opponent, alpha-beta skips branches without changing the root value, and expectimax replaces worst-case choice with probability.

CMU 07-280 Lecture 4: CSPs, AC-3, and Search Order

Lecture 4 exposes structure through variables, domains, and constraints, then upgrades DFS with backtracking, forward checking, AC-3, MRV, and LCV; the goal is to prove failure earlier.

CMU 07-280 Lecture 5: Defining Machine Learning with Loss, Risk, and ERM

Lecture 5 formulates machine learning through `X → Y`, loss, risk, and empirical risk minimization: a training set only gives average observed loss, while the real objective remains generalization over an unknown distribution.

CMU 07-280 Lecture 6: How Decision Trees Split Data with Mutual Information

Lecture 6 recursively grows a tree from decision stumps, measures label uncertainty with entropy, and selects splits by `I(Y;W)=H(Y)-H(Y|W)`; this is computationally practical greedy ERM, not a global optimal-tree guarantee.

CMU 07-280 Lecture 7: Linear Regression and the Normal Equation

Lecture 7 applies ERM to linear functions and squared loss, moves from a one-dimensional slope to `argmin ||y-Xθ||²`, and derives the normal equation when `XᵀX` is invertible.

CMU 07-280 Lecture 8: Gradient Descent, SGD, and Learning Rate

Lecture 8 moves from a one-dimensional parabola to vector gradients and compares batch GD, SGD, and mini-batches; the learning rate determines whether updates converge, oscillate, or diverge.

CMU 07-280 Lecture 9: Logistic Regression as Probability Estimation

Lecture 9 models P(y=1|x) with a sigmoid instead of directly predicting 0 or 1, learns parameters with cross-entropy and convex optimization, and extends naturally to softmax regression.

CMU 07-280 Lecture 10: Trading Expressiveness for Stability with Features and Regularization

Lecture 10 uses φ(x) to let linear models express nonlinear functions, then controls the resulting overfitting with train/validation/test separation, L1/L2 regularization, and model selection.

CMU 07-280 Lecture 11: Building a Neural Network from Logistic Regression

Lecture 11 expands a logistic unit into a multilayer network: linear layers produce z, activations produce a, and multiple neurons jointly learn a feature transform trained through a final loss.

CMU 07-280 Lecture 12: How Backpropagation Reuses the Chain Rule

Lecture 12 treats a network as a computation graph: the forward pass stores intermediates, the backward pass propagates upstream gradients, and local linear, activation, and softmax rules compute every parameter gradient efficiently.

CMU 07-280 Lecture 13: From Reward Hacking to Auditable AI Scientists

Lecture 13 separates alignment into specification, distribution shift, oversight, and corrigibility, then uses benchmark selection, leakage, and post-hoc selection experiments to show why a final paper cannot audit an autonomous research workflow.

CMU 07-280 Lecture 14: Encoding Image Structure with Convolutional Networks

Lecture 14 replaces dense image models with local connectivity and parameter sharing, moving from convolution, stride, padding, and pooling to AlexNet, GPU data parallelism, ResNet skip connections, and BatchNorm.

CMU 07-280 Lecture 15: Separating Pretraining, Transfer Learning, and Fine-Tuning

Lecture 15 splits a pretrained model into representation g and task head h: freeze g and train only the head, or fine-tune some or all parameters at a smaller learning rate depending on data volume and source-target distance.

CMU 07-280 Lecture 16: Unifying Logistic and Linear Regression with Maximum Likelihood

Lecture 16 starts from likelihood p(D|θ), uses i.i.d. to factor the joint probability and logs to turn products into sums; Bernoulli MLE yields sample proportions, conditional Bernoulli yields logistic cross-entropy, and Gaussian noise yields squared error.

CMU 07-280 Lecture 17: From Tokenization to N-gram Language Models

Lecture 17 first decides how text becomes tokens, then uses N-grams to turn sequence probability into conditional probabilities estimated from corpus counts. Tokenization is the first design decision about what a model can see.

CMU 07-280 Lecture 18: How N-grams Train, Sample, and Fail

Lecture 18 truncates the chain rule with an N-gram Markov assumption, estimates probabilities from corpus counts, and contrasts greedy, categorical, and temperature sampling. The real bottlenecks are zero probability for unseen contexts and a fixed window.

CMU 07-280 Lecture 19: Turning Next-token Prediction into Geometry

Lecture 19 builds a minimal next-token model from two embedding matrices, dot-product similarity, softmax, and cross-entropy. Shared vector parameters replace the isolated count cells of an N-gram table.

CMU 07-280 Lecture 20: From Position Encoding to Causal Self-Attention

Lecture 20 expands one-token embeddings into sequences, adds positional information, derives Q/K/V scaled dot-product attention and causal masking, and assembles multi-head blocks into a GPT-2 skeleton.

CMU 07-280 Lecture 21: How Bellman Equations Solve Markov Decision Processes

Lecture 21 formulates stochastic sequential decisions as an MDP with known dynamics, defines value and Q-values through Bellman backups, and solves for an optimal policy with value or policy iteration.

CMU 07-280 Lecture 22: Q-learning When Dynamics Are Unknown

Lecture 22 keeps the MDP structure but removes known transitions and rewards. TD learning updates value from one sample, and Q-learning uses an off-policy target to learn optimal action values directly.

CMU 07-280 Lecture 23: From Approximate Q-learning to DQN

Lecture 23 replaces a huge Q-table with Qθ(s,a): first derive a gradient update for linear features from squared TD error, then add replay data and a fixed target network to form DQN.

CMU 07-280 Lecture 24: How Monte Carlo Tree Search Connects to AlphaZero

Spring 2026 Lecture 24 is MCTS, not Fall 2026 LLM post-training. It allocates simulations through selection, expansion, rollout, backup, and UCB, then connects policy/value heads and self-play to AlphaZero.

CMU 07-280 Stage Review I: From Search Problems to Supervised Learning

Lectures 1–12 form one decision pipeline: define states, moves, and objectives, then use heuristics, losses, regularization, and backpropagation to control an otherwise intractable search space.

CMU 07-280 Stage Review II: Building AlexNet and GPT-2 as Working Systems

Stage II uses HW8 and HW11 to test whether representation, computation graphs, training, transfer, and generation actually connect, rather than treating CNNs and Transformers as diagrams to memorize.

CMU 07-280 Stage Review III: From MDPs and Q-learning to AlphaZero

Stage III connects value, policy, bootstrapping, function approximation, and MCTS into AlphaZero: a network supplies priors and estimates, search improves decisions, and self-play creates the next training set.

CMU 11-785 Lecture 1: Introduction

Spring 2026 Lecture 1 focuses on neurons, perceptrons, connectionism, and the problem framing of deep learning. This guide follows the official slides and recording and adds a small self-check that does not depend on the enrolled-course grader.

CMU 11-785 Lecture 2: Neural Nets as Universal Approximators

Spring 2026 Lecture 22 focuses on latent variables, the ELBO, the KL term, and the reparameterization trick. This guide follows the official slides and recording and adds a small self-check that does not depend on the enrolled-course grader.

CMU 11-785 Lecture 3: Training I: Learning and Empirical Risk Minimization

Spring 2026 Lecture 3 focuses on data distributions, hypotheses, losses, empirical risk, and their roles in generalization. This guide follows the official slides and recording and adds a small self-check that does not depend on the enrolled-course grader.

CMU 11-785 Lecture 4: Training II: Gradient Descent

Spring 2026 Lecture 4 focuses on gradients, learning rates, parameter updates, and the training of a linear neuron. This guide follows the official slides and recording and adds a small self-check that does not depend on the enrolled-course grader.

CMU 11-785 Lecture 5: Training III: Backpropagation

Spring 2026 Lecture 5 focuses on computational graphs, the chain rule, local derivatives, and gradient reuse. This guide follows the official slides and recording and adds a small self-check that does not depend on the enrolled-course grader.

CMU 11-785 Lecture 6: Training IV: Convergence, Loss Surfaces, and Momentum

Spring 2026 Lecture 6 focuses on non-convex loss surfaces, curvature, saddle points, and momentum's accumulated direction. This guide follows the official slides and recording and adds a small self-check that does not depend on the enrolled-course grader.

CMU 11-785 Lecture 7: Training V: SGD and Second-order Methods

Spring 2026 Lecture 7 focuses on the tradeoffs among full-batch, mini-batch, stochastic gradients, and second-order information. This guide follows the official slides and recording and adds a small self-check that does not depend on the enrolled-course grader.

CMU 11-785 Lecture 8: Training VI: Optimizers and Regularization

Spring 2026 Lecture 8 focuses on AdaGrad, Adam, regularization, BatchNorm, Dropout, and loss selection. This guide follows the official slides and recording and adds a small self-check that does not depend on the enrolled-course grader.

CMU 11-785 Lecture 9: CNNs I

Spring 2026 Lecture 9 focuses on local connectivity, weight sharing, convolution kernels, and feature maps. This guide follows the official slides and recording and adds a small self-check that does not depend on the enrolled-course grader.

CMU 11-785 Lecture 10: CNNs II

Spring 2026 Lecture 10 focuses on stride, padding, receptive fields, and multi-channel convolution. This guide follows the official slides and recording and adds a small self-check that does not depend on the enrolled-course grader.

CMU 11-785 Lecture 11: CNNs III

Spring 2026 Lecture 11 focuses on stacked convolutional architectures, feature hierarchies, and design tradeoffs. This guide follows the official slides and recording and adds a small self-check that does not depend on the enrolled-course grader.

CMU 11-785 Lecture 12: CNNs IV

Spring 2026 Lecture 12 focuses on CNN training, architecture selection, and the end-to-end assembly of a vision model. This guide follows the official slides and recording and adds a small self-check that does not depend on the enrolled-course grader.

CMU 11-785 Lecture 13: RNNs I

Spring 2026 Lecture 13 focuses on sequence state, temporal unrolling, parameter sharing, and recurrent computation. This guide follows the official slides and recording and adds a small self-check that does not depend on the enrolled-course grader.

CMU 11-785 Lecture 14: RNNs II

Spring 2026 Lecture 14 focuses on backpropagation through time, gradient stability, and LSTM-style gated memory. This guide follows the official slides and recording and adds a small self-check that does not depend on the enrolled-course grader.

CMU 11-785 Lecture 15: Seq2Seq and Connectionist Temporal Classification

Spring 2026 Lecture 15 focuses on variable-length input/output, unknown alignment, and the CTC objective. This guide follows the official slides and recording and adds a small self-check that does not depend on the enrolled-course grader.

CMU 11-785 Lecture 16: CTC Blanks and Beam Search

Spring 2026 Lecture 16 focuses on blanks, collapse rules, prefix probabilities, and approximate decoding. This guide follows the official slides and recording and adds a small self-check that does not depend on the enrolled-course grader.

CMU 11-785 Lecture 17: Language Models and Translation

Spring 2026 Lecture 17 focuses on autoregressive factorization, conditional language models, and translation decoding. This guide follows the official slides and recording and adds a small self-check that does not depend on the enrolled-course grader.

CMU 11-785 Lecture 18: Attention and Transformers

Spring 2026 Lecture 18 focuses on queries, keys, values, scaled dot-product attention, and the Transformer block. This guide follows the official slides and recording and adds a small self-check that does not depend on the enrolled-course grader.

CMU 11-785 Lecture 19: Transformers and Newer Architectures

Spring 2026 Lecture 19 focuses on encoder/decoder structures, masks, residual paths, and architecture variants. This guide follows the official slides and recording and adds a small self-check that does not depend on the enrolled-course grader.

CMU 11-785 Lecture 20: Large Language Models

Spring 2026 Lecture 20 focuses on scaled autoregressive models, training stages, inference, and capability boundaries. This guide follows the official slides and recording and adds a small self-check that does not depend on the enrolled-course grader.

CMU 11-785 Lecture 21: Representations and Autoencoders

Spring 2026 Lecture 21 focuses on bottleneck representations, reconstruction objectives, dimensionality reduction, and representation quality. This guide follows the official slides and recording and adds a small self-check that does not depend on the enrolled-course grader.

CMU 11-785 Lecture 22: Variational Autoencoders

Spring 2026 Lecture 22 focuses on latent variables, the ELBO, the KL term, and the reparameterization trick. This guide follows the official slides and recording and adds a small self-check that does not depend on the enrolled-course grader.

CMU 11-785 Lecture 23: Diffusion Models

Spring 2026 Lecture 23 focuses on forward noising, reverse denoising, score or noise prediction, and sampling. This guide follows the official slides and recording and adds a small self-check that does not depend on the enrolled-course grader.

CMU 11-785 Lecture 24: Generative Adversarial Networks

Spring 2026 Lecture 24 focuses on the generator, discriminator, minimax objective, and training instability. This guide follows the official slides and recording and adds a small self-check that does not depend on the enrolled-course grader.

CMU 11-785 Lecture 25: Graph Neural Networks

Spring 2026 Lecture 25 focuses on message passing, aggregation, node representations, and permutation symmetry. This guide follows the official slides and recording and adds a small self-check that does not depend on the enrolled-course grader.

CMU 11-785 Lecture 26: Reinforcement Learning

Spring 2026 Lecture 26 focuses on states, actions, rewards, returns, values, and policy learning. This guide follows the official slides and recording and adds a small self-check that does not depend on the enrolled-course grader.

CMU 11-785 Lecture 27: Hopfield Networks

Spring 2026 Lecture 27 focuses on associative memory, energy functions, fixed points, and pattern retrieval. This guide follows the official slides and recording and adds a small self-check that does not depend on the enrolled-course grader.

CMU 11-785 Lecture 28: Boltzmann Machines

Spring 2026 Lecture 28 focuses on energy-based probability models, stochastic units, the partition function, and learning difficulty. This guide follows the official slides and recording and adds a small self-check that does not depend on the enrolled-course grader.

A Complete Guide to CMU 11-785: 28 Public Lectures, but an Incomplete Assignment Chain

CMU 11-785 Spring 2026 publishes official slides and YouTube recordings for all 28 content lectures, plus extensive bootcamps and recitations. Its HW1–HW4 specifications, starters, and evaluation still depend on Autolab, Piazza, and Kaggle.

ai deep-dive

Python Coding Agent M11: Why an Exec Loop Cannot Reproduce the Claude Code Conversation Experience

A Claude Code- or Codex-style TUI depends on long-lived sessions, typed transcripts, and approval at tool boundaries—not a screen full of color.

ai deep-dive

Cognee Complete Guide: Turning Documents into Graph Memory for Agents

Cognee is a data-to-memory pipeline: a relational store preserves sources and provenance, a vector store finds semantically similar content, and a graph store represents entity relationships, exposed through remember, recall, improve, and forget.

CS124 Week 1 Introduction and Setup: Turning Language Problems into Computable Components

CS124 Winter 2026 opens by mapping a ten-week path from tokenization and classification to retrieval, speech, networks, and LLMs, while PA0 establishes the Jupyter environment used throughout the quarter.

CS124 Week 10 PageRank and Social Networks: From Anchor Text and Centrality to the Course Wrap-Up

Week 10 models the Web with anchor text, PageRank, and centrality; post-training, multilinguality, and speech belong only to a public final-deck outline labeled 2025, not the 2026 live narration.

CS124 Week 2 Words, Tokens, Edit Distance, and N-grams: Decide What the Model Sees First

Week 2 builds three layers: a token vocabulary with BPE, sequence comparison with dynamic-programming edit distance, and probability approximation with n-grams; PA1 turns regex and BPE into executable work.

CS124 Week 3 Logistic Regression and Text Classification: From Features to Probability and Loss

Week 3 connects text features, sigmoid probabilities, cross-entropy loss, and gradient descent, producing a classifier whose feature contributions remain inspectable.

CS124 Week 4 Information Retrieval: The Indexing and Ranking Layer Beneath RAG

Week 4 builds candidates with an inverted index, ranks them with tf-idf and cosine similarity, and then connects retrieved evidence to generation; PA3 exposes RAG's inspectable retrieval half.

CS124 Week 5 Embeddings and Social NLP: Context Vectors and the Public-Evidence Boundary

Week 5's public materials support the distributional hypothesis, word embeddings, and cosine similarity; the paired Social NLP lecture is unrecorded and restricted, so concrete audit methods are labeled as author extensions.

CS124 Week 6 Neural Networks and LLMs: From Units and Backpropagation to Decoder-Only Models

Week 6 uses public neural-network slides for weighted sums, nonlinearities, loss, and backpropagation, then a public LLM/Transformer deck labeled 2025 for decoder-only architecture without treating it as the 2026 live transcript.

CS124 Week 7 Transformers and Speech Processing: Causal Attention, Generation, and an Unrecorded Lecture

Week 7's public path is PA6a: implement causal self-attention, train a small Shakespeare Transformer, sample text, and compute perplexity; the live speech lecture remains an explicit source gap.

CS124 Week 8 Speech and the PA7/Git Lab: Auditing Information Loss in a TTS-to-STT Pipeline

Week 8 sends text through TTS and back through STT, requiring error classification, formatting-loss analysis, and accent stress tests, while Lab 4 prepares Git collaboration for the team agent project.

CS124 Week 9 Collaborative Filtering and LLM Agents: From Movie Similarity to Search and Memory Tools

Week 9 builds movie recommendations with item-item collaborative filtering, then packages recommendation, web search, databases, and memory as agent tools under API-budget and team constraints.

CS224N Lecture 3: Matrix Calculus and Backpropagation

Lecture 3 decomposes neural-network training into computation graphs, local derivatives, and the chain rule: the forward pass computes a result; backprop accumulates gradients from the output so every parameter knows how to move.

CS224N Lecture 11: Why LLM Benchmarks Expire

Lecture 11 divides evaluation into what to test, how to measure it, and when the result stops being trustworthy. Benchmarks saturate or leak, prompts change scores, and an LLM judge remains a biased model.

CS224N Lecture 9: Prompting, LoRA, and Parameter-Efficient Adaptation

Lecture 9 compares prompting, pruning, LoRA, prompt tuning, and adapters. Each asks the same question: how many parameters must change, and how much task-specific state must be stored, to adapt a large pretrained model?

CS224N Lecture 6: Turn a Final Project into a Testable Question

Lecture 6 completes the Transformer picture with encoders, decoders, and cross-attention, then breaks the final project into formats, assessment, research topics, and data. A viable topic needs one explicit baseline and metric.

CS224N Lecture 1: Four Paradigm Shifts in NLP

Winter 2026 Lecture 1 divides NLP into four eras: early exploration, symbolic systems, statistical machine learning, and deep/self-supervised learning. The point is not the dates but how each era redefined the language problem.

CS224N Lecture 15: Reading Agentic Interpretability Without Public Slides

Lecture 15 is Been Kim's interpretability guest session, but the Winter 2026 site publishes no slides or agenda. This article does not invent lecture content; it maps the five official readings across concept discovery, agentic investigation, and new vocabulary.

CS224N Lecture 17: An Official Reading Map for Multimodality

Lecture 17 is Luke Zettlemoyer's multimodality guest session, but the site publishes no slides or agenda. Its official readings establish three routes: visual reasoning workspaces, early-fusion token models, and text autoregression with image diffusion.

CS224N Lecture 19: How Small Models Can Move Beyond Brute-Force Scaling

The final lecture frames Open Questions in NLP 2026 as smart scaling: prolonged RL, Prismatic synthetic data, RL as pretraining, and open collaboration seek reasoning gains beyond adding parameters.

CS224N Lecture 8: From Instruction Tuning and RLHF to DPO

Lecture 8 explains how instruction tuning, preference data, and RLHF turn a pretrained model into an assistant, then derives DPO from winner–loser pairs. Every step converts human judgment into signal—and imports its biases.

CS224N Lecture 7: Pretraining, Subwords, and In-Context Learning

Lecture 7 decomposes pretraining into scalable data, subword tokenization, three model objectives, and in-context learning. A general self-supervised objective yields reusable representations; downstream signals specify their use.

CS224N Lecture 10: Six Components of RAG and Language Agents

Lecture 10 moves from question answering and RAG into language agents, then decomposes them into reasoning and planning, memory, tools, data, and evaluation. An agent is an inspectable loop between a model and external state.

CS224N Lecture 12: Decoding, DeepSeek-R1, and Reasoning Training

Lecture 12 shows that output policy is not a detail: greedy, beam, and sampling produce different text. It then moves from R1-Zero/R1 into PPO, GRPO, and DAPO, asking when longer reasoning actually helps.

CS224N Lecture 13: Speculative Decoding and Test-Time Scaling

Lecture 13 moves from inference efficiency to inference capability: speculative decoding drafts with a small model and verifies with a large one; on-policy distillation addresses drift; long context and test-time scaling spend inference resources.

CS224N Lecture 4: Language Models, RNNs, and Vanishing Gradients

Lecture 4 defines a language model as a next-word probability distribution, then uses an RNN to compress an arbitrarily long prefix. It also exposes recurrence's central cost: information and gradients travel one time step at a time.

CS224N Lecture 16: Hallucination, Creativity, Work, and Alignment

Lecture 16 divides NLP's social impact into four questions: why models hallucinate, why AI-assisted creativity may homogenize output, how work is reorganized, and why value alignment cannot be reduced to one reward.

CS224N Lecture 18: Material-Gap Record for Tinker and LoRA Without Regret

Lecture 18 is a John Schulman guest session. The official page gives only the title Tinker and LoRA Without Regret, date, and speaker—no slides, agenda, or readings—so this article records confirmed facts and unknowns only.

CS224N Lecture 14: How Tokenization Creates Multilingual Cost Gaps

Lecture 14 moves from word, character/byte, and subword segmentation to BPE failures and cross-lingual fairness. A tokenizer determines sequence length, compute cost, and the units a model sees; it is not neutral preprocessing.

CS224N Lecture 5: From Recurrence to the Transformer

Lecture 5 moves from the long-range and sequential bottlenecks of RNNs to self-attention and the Transformer. It shortens information paths and enables parallel computation, at the price of quadratic attention and separately encoded position.

CS224N Lecture 2: How word2vec Turns Meaning into Vectors

Lecture 2 moves from word2vec's prediction task, objective, and gradients to count-based vectors and evaluation. Meaning becomes a high-dimensional position learned from context, not a label retrieved from a dictionary.

Stanford CS224V Lecture 10: How SPINACH Explores Wikidata and Builds SPARQL

SPINACH does not guess complete SPARQL in one shot. It searches entities and properties, inspects Wikidata entries and examples, executes small queries, and composes a final query under explicit action and stopping rules.

Stanford CS224V Lecture 12: CHURRO Makes Multilingual Historical Documents Searchable

CHURRO represents full-page text, layout, and metadata in HDML, unifies multilingual historical data for a page-level VLM, and connects extraction to HistoryGenie for searchable, conversational archives.

Stanford CS224V Lecture 14: Scaling Language Models When Data Is the Bottleneck

The final lecture is not a complete LLM-training tutorial. It studies data efficiency under fixed data and abundant compute, revisiting epochs, batches, ensembles, self-training, and conditions for synthetic continued pretraining.

Stanford CS224V Lecture 5: WikiChat's Seven-Stage Defense Against Hallucination

The [WikiChat paper](https://aclanthology.org/2023.findings-emnlp.157/) expands RAG into query formulation, retrieval, filtering, generation, claim extraction, renewed retrieval and verification, and removal of unsupported content—and evaluates retrieval separately from factuality.

Stanford CS224V Lecture 1: Turning Hallucinating LLMs into Dependable Assistants

Fall 2025 opens with computational thinking: reliability comes from decomposing retrieval, formal representation, verification, and generation into testable algorithms, not from one heroic prompt.

Stanford CS224V Lecture 2: STORM, Co-STORM, and Knowledge Curation

STORM uses perspective-guided questions, simulated interviews, and outlines to broaden research; Co-STORM keeps a person in the loop so discovering unknown questions and co-editing become part of the system.

Stanford CS224V Lecture 8: SLIDERS Turns Long-Document Sets into Queryable Tables

SLIDERS induces a question-specific schema, applies semantic chunking and contextualized extraction, reconciles duplicate rows, and answers with SUQL instead of feeding every long document directly to one model.

Stanford CS224V Lecture 13: ReactGenie Gives Voice and Native GUIs Shared State

ReactGenie annotates React components to expose data, actions, and views, parses composite voice commands into a DSL, and renders native graphical output against shared UI context.

Stanford CS224V Lecture 11: Translate Trial Criteria into SMT Instead of Asking an LLM to Decide

The lecture parses patient records and trial criteria into SMT, retrieves candidates through a weaker propositional projection, and runs a solver on the reduced set. Reasoning is inspectable, but NL-to-SMT remains the main error boundary.

Stanford CS224V Lecture 9: Why Automated Qualitative Coding Still Needs Expert Review

Automated qualitative coding defines event types and arguments in a codebook, then separates document classification, structured extraction, and entity linking. Constrained JSON fixes form, not expert judgment.

Stanford CS224V Lecture 6: Why Database Agents Begin with Semantic Parsing

Reliable database agents map language to executable queries, resolve schemas and enumerated values, and evaluate execution separately from answer generation; hybrid questions additionally require explicit source routing.

Stanford CS224V Lecture 7: SUQL Unifies SQL and Free-Text Retrieval

SUQL adds answer and summary functions over text to SQL. A semantic parser emits one hybrid query, while an optimizing compiler applies predicate pushdown, top-k pruning, and lazy evaluation.

Stanford CS224V Lecture 4: Task-Agent Evaluation Beyond Human-Like Answers

CS224V splits task-agent evaluation into state updates and complete interaction: isolate the semantic parser, then test task completion, grounded queries, and valid actions with real users.

Stanford CS224V Lecture 3: Building Task-Oriented Agents with Genie Worksheets

Genie Worksheets declare task capability as a form-like specification. A contextual semantic parser updates formal dialogue state while the runtime controls queries, actions, and responses.

CS336 Lecture 3: Transformers Have Many Variants but Few Stable Defaults

Lecture 3 does not turn its survey of modern LLMs into a single best recipe. It finds a conservative consensus—pre-norm, RMSNorm, no biases, SwiGLU, and RoPE—plus a small set of deviations justified by inference cost or stability.

CS336 Lecture 4: Attention Has Alternatives, and MoE Does Not Scale for Free

Lecture 4 studies two kinds of sparsity: linear/recurrent attention reduces sequence-length cost, while MoE activates only part of a model for each token. Both turn saved FLOPs into routing, balancing, communication, and kernel problems.

CS336 Lecture 14: Filtering, Deduplication, and Mixing Turn Raw Web Data into Training Data

Lecture 14 moves raw documents through language, quality, and safety filtering; exact and near deduplication; and source mixing. Each stage reshapes model behavior, while synthetic instruction and agent trajectories extend the pipeline into executable environments.

CS336 Lecture 13: Data Does Not Fall from the Sky, and Every Source Has Access and License Costs

Lecture 13 traces training sources through Common Crawl, Wikipedia, GitHub, arXiv, books, and open datasets. Technically accessible is not the same as licensed, and raw data is not training data; provenance must precede cleaning and mixing.

CS336 Lecture 12: There Is No Single True LLM Score, Only Different Games

Lecture 12 moves from perplexity to exams, chat preferences, agents, reasoning, and safety. Every benchmark changes the capability definition, scaffold, judge, and contamination risk, so evaluation must first say whether it compares a method, model, or complete system.

CS336 Lecture 5: GPUs Win by Moving Data Less, Not by Making Each Thread Fast

Lecture 5 explains GPUs through SMs, warps, and the memory hierarchy, then unifies common optimization under low precision, fusion, recomputation, coalescing, and tiling. FlashAttention combines those principles for attention.

CS336 Lecture 10: LLM Inference Is About Reading Weights and KV Cache Less Often

Lecture 10 separates prefill from decode: prefill parallelizes and is often compute-bound, while decode is sequential and commonly bandwidth-bound. GQA/MLA, quantization, speculative decoding, continuous batching, and PagedAttention reshape that cost.

CS336 Lecture 6: Benchmark and Profile Before Writing a Triton Kernel

Lecture 6 turns GPU principles into kernels: benchmark scaling across shapes, profile actual calls and time, then implement GeLU, softmax, reductions, and tiled matrix multiplication in Triton. Speed begins with measuring correctly.

CS336 Lecture 17: Multimodal Models Turn Images into Tokens, Then Reconcile Semantics with Detail

Lecture 17 organizes CLIP/SigLIP, LLaVA, Qwen-VL, and Chameleon into three paths: contrastive encoders learn semantics, vision-encoder/projector/LM stacks provide understanding, and discrete image tokens enable generation. Resolution, token budgets, and modality balance constrain them all.

CS336 Lecture 1: From Bytes to a Tokenizer—and What Deserves to Scale

CS336's first lecture does not treat building a language model from scratch as reenacting every old technique. It separates mechanics, mindset, and intuitions, then uses BPE to show how raw bytes become trainable tokens.

CS336 Lecture 7: Build Data, Tensor, and Pipeline Parallelism from Collectives

Lecture 7 starts below FSDP APIs, building a communication language from broadcast, all-reduce, all-gather, reduce-scatter, and all-to-all before assembling data, tensor, and pipeline parallelism.

CS336 Lecture 8: Align ZeRO, FSDP, and 3D Parallelism with Hardware Topology

Lecture 8 moves from parallel primitives to system design: ZeRO progressively shards optimizer state, gradients, and parameters; TP, PP, SP, and EP split width, depth, sequence, and experts. Their composition must follow topology and dynamic activation memory.

CS336 Lecture 2: Count FLOPs and Memory Before Asking Whether a Model Fits

Lecture 2 reduces model training to tensors, FLOPs, bytes, and time: use einops to track dimensions, arithmetic intensity and roofline analysis to identify bottlenecks, then trade compute for memory with gradient accumulation and activation checkpointing.

CS336 Lecture 16: RLVR Scales Reasoning with Verifiable Rewards, but GRPO Is Not Free PPO

Lecture 16 moves from PPO to GRPO and RLVR. Math, code, and environment outcomes provide scalable rewards and avoid some preference-model overoptimization, but group-normalized advantages introduce difficulty and length bias while rollout infrastructure becomes the dominant cost.

CS336 Lecture 9: Scaling Laws Are Extrapolation Tools, Not Crystal Balls

Lecture 9 begins with log-log linear relationships between data and error, then uses scaling laws to compare architectures, optimizers, batches, and model-data allocations. The Chinchilla dispute shows how fitting methods, observed ranges, and deployment objectives change the answer.

CS336 Lecture 11: Scaling Laws in Practice Must Scale Learning Rate and Batch Too

Lecture 11 reads public recipes from MiniCPM, DeepSeek, Qwen, and Llama 3: hold most architectural ratios fixed, sweep learning rate and batch at small scale, then choose model/data allocation with IsoFLOPs. μP helps, but normalization, optimizers, and weight decay can break transfer.

CS336 Lecture 15: SFT Teaches Imitation; RLHF Begins Direct Preference Optimization

Lecture 15 divides post-training into imitation and optimization. SFT extracts pretrained capabilities from instruction-response data; RLHF uses pairwise feedback to bridge demonstrations and preferences. PPO and DPO both inherit data bias, reward overoptimization, and mode collapse.

ai deep-dive

Daytona Agent Sandbox: A Forkable Computer for Every Agent

Daytona treats a sandbox as a long-lived computer that can start, pause, snapshot, and fork. It raised a $24 million Series A in 2026, while a Laude Institute case study reports 37,000 sandboxes in one week. It fits parallel evaluations and coding agents, but its core open-source repository is no longer maintained.

ai deep-dive

Deepgram Voice Agent API: From Streaming STT and Turn Detection to TTS

Deepgram combines streaming STT, LLM orchestration, turn detection, barge-in, and streaming TTS over one WebSocket while preserving paths for standalone speech models and bring-your-own LLM or TTS.

ai deep-dive

Dify as a Low-Code Agent Platform: From a Working Workflow to a Published AI App

Dify puts models, Knowledge, visual Workflows, Agents, Plugins, and application APIs in one workspace; this guide builds a minimal Workflow that can be tested, published, and called through the API, then explains when an Agent is actually warranted.

ai deep-dive

DSPy: Compiling AI Programs with Signatures, Metrics, and Optimizers

DSPy replaces handwritten prompt strings with task Signatures, execution Modules, and Optimizers that compile better instructions and examples against a dataset and metric.

ai deep-dive

E2B Agent Sandbox: Put Model-Generated Code in a Resumable microVM

E2B combines Templates, Firecracker microVMs, and process, file, and network APIs into an agent execution layer. Its real selection advantage is preserving memory and processes across pause and resume, not merely providing another code interpreter.

ai deep-dive

ElevenLabs ElevenAgents: The Lifecycle from Realtime Speech to Phone Agents

ElevenLabs has expanded from a TTS vendor into the ElevenAgents platform: Scribe Realtime listens, Flash speaks, and the platform connects the LLM, turn-taking, tools, and telephony. The key choice is whether you need a voice model or the whole agent control plane.

ai deep-dive

Fireworks AI: From Serverless APIs to Custom Model Deployments

Fireworks AI puts open-weight model evaluation, dedicated GPU deployments, and LoRA customization behind one API surface. Serverless fits low-volume starts, On-demand fits sustained traffic and custom models, while reserved capacity adds enterprise capacity guarantees.

ai deep-dive

Flowise Deep Dive: From Assistant, Chatflow, and Agentflow to an EOL Migration Decision

Flowise uses Assistant, Chatflow, and Agentflow to cover simple assistants, single-agent systems, and multi-agent orchestration; however, its repository was archived in August 2026 and official EOL is scheduled for August 31, so new projects should not adopt it without a maintained fork and migration plan.

ai deep-dive

Galileo Deep Dive: Experiments, Evaluators, and the Agent Observability Loop

Galileo connects dataset experiments, LLM/code/Luna evaluators, production traces, and runtime guardrails. It fits enterprises that need observability plus intervention, but the old Protect surface is deprecated and evaluator models do not replace human calibration or application security.

The H2 2026 Harness War: Eight Frameworks Rewriting, Three Model Makers Entering, 110+ CLIs — How to Make Sense of It

In August 2026, it's not just five frameworks moving. Beyond OMP 2, Pi v2, Opencode 2, dsh, and Claude Code, three model makers — Google (Antigravity CLI), Meta (Muse Code), and xAI (Grok Build) — are building coding agents directly. Add Amp, Cline 2.0, and the Codex CLI Rust rewrite, and eight-plus frameworks are undergoing architecture-level changes simultaneously. Factor in 110+ total CLI tools, and H2 2026 is a divergence period for harness methodology. This article analyzes four architectural approaches, one shared direction, and one emerging trust crisis.

ai deep-dive

Haystack Deep Dive: Testable RAG with Components and Pipelines

Haystack turns indexing, retrieval, generation, and evaluation into replaceable Components connected by directed-multigraph Pipelines; it fits Python teams that want RAG flows to be tested, versioned, and deployed as code.

ai deep-dive

Helicone Deep Dive: LLM Gateway, Request Tracing, and Cost Analytics

Helicone is an open-source LLM gateway and observability platform: requests sent through its compatible endpoint automatically capture model, latency, tokens, cost, and custom properties, while managed credits or BYOK enable routing and fallbacks.

ai deep-dive

Hugging Face Is More Than a Model Download Site: Hub, Datasets, Spaces, and Inference

Hugging Face Hub is a collaboration layer for versioned models, datasets, and applications. Datasets handles data, Spaces runs demos, while Inference Providers and Endpoints provide managed inference.

ai deep-dive

Hyperbrowser Deep Dive: Browser-as-a-Service Infrastructure for Agents

Hyperbrowser packages Chrome sessions, proxies, stealth, profiles, and recordings behind managed Playwright and Puppeteer APIs. It fits agents that need to scale real-browser work quickly, while profile credentials, anti-bot compliance, and proxy bandwidth costs remain application responsibilities.

Jina Reader Guide: Turn Web Pages into Agent-Readable Markdown

Jina Reader turns a known URL into LLM-friendly Markdown; production use still requires explicit rendering, scope, token-budget, validation, and fallback decisions.

ai deep-dive

LanceDB Deep Dive: Embedding Vector Search in Arrow Data Workflows

LanceDB stores vectors, metadata, and multimodal source data in the Lance columnar format. Its OSS edition embeds in Python, TypeScript, or Rust processes; distributed Enterprise becomes relevant when the data or service outgrows one machine.

ai deep-dive

LangChain v1 Agents: create_agent, Middleware, and the LangGraph Runtime

LangChain v1 provides a high-level agent loop through create_agent, runs it on LangGraph, and treats tools, structured output, and middleware as its extension boundaries.

ai deep-dive

LangSmith Deep Dive: From Agent Traces to Offline and Online Evaluation

LangSmith structures LLM applications as projects, traces, runs, and threads, then uses datasets, evaluators, and experiments to turn production failures into offline regression tests. It observes any LLM application and does not require LangChain.

ai deep-dive

Letta and MemGPT Complete Guide: Memory Inside a Stateful Agent Runtime

Letta extends MemGPT's operating-system analogy but is not a standalone memory API. The runtime persists agent state, editable in-context blocks, conversation history, and external archival memory, while the model can actively curate memory through tools.

ai deep-dive

LiteLLM: From a Python SDK to a Self-Hosted AI Gateway

LiteLLM is not a model provider. It is a Python SDK and self-hosted proxy that normalizes 100+ LLM APIs, then centralizes routing, fallbacks, virtual keys, budgets, and observability at the gateway layer.

ai deep-dive

LiveKit Voice Agents: From WebRTC Rooms to Interruptible Voice Pipelines

LiveKit models a voice agent as a server participant in a realtime media room, with AgentSession orchestrating STT, turn detection, LLM, TTS, and interruption. It raised a $100 million Series C at a $1 billion valuation in 2026. It fits products needing WebRTC, multiple client platforms, telephony, and swappable models, but self-hosting the media server does not self-host the entire AI pipeline.

ai deep-dive

Mastra: Agents, Workflows, Memory, and Evals in TypeScript

Mastra is a TypeScript agent framework that combines agents, typed workflows, memory, MCP, tracing, and scorers in one Node.js development environment.

ai deep-dive

Mem0 Complete Guide: Controlled Long-Term Memory for AI Agents

Mem0 sits between an agent and storage: it extracts durable facts from interactions, scopes them by user, agent, or run, and searches them before a later generation. Its appeal is a small API; its risks are extraction errors, stale memories, and authorization boundaries.

ai deep-dive

Milvus Vector Database Deep Dive: Segments, Indexes, and Distributed Operations

Milvus separates real-time ingestion, historical queries, index building, and persistence into independently scalable components. It fits large, continuously updated retrieval services, but smaller projects often pay too much operational complexity for that architecture.

ai guide Reading MIT 6.S191

MIT 6.S191 Lecture 1: The Minimal Structure of Deep Learning

Lecture 1 of the 2026 course builds the vocabulary shared by the rest of the course: perceptrons, forward propagation, loss, and gradient descent.

ai guide Reading MIT 6.S191

MIT 6.S191 Lecture 2: Sequence Modeling: From RNNs to Attention

Lecture 2 of the 2026 course addresses data where order changes meaning—text, audio, and time series—and connects directly to music generation in Lab 1.

ai guide Reading MIT 6.S191

MIT 6.S191 Lecture 3: Computer Vision: How Convolution Preserves Spatial Structure

Lecture 3 of the 2026 course moves from image tensors, convolution, and pooling to recognition systems, preparing for MNIST and face detection in Lab 2.

ai guide Reading MIT 6.S191

MIT 6.S191 Lecture 4: Generative Modeling: From Latent Spaces to Diffusion

Lecture 4 of the 2026 course separates generative from discriminative tasks, organizes VAE, GAN, and diffusion objectives, and leads into Lab 2’s DB-VAE.

ai guide Reading MIT 6.S191

MIT 6.S191 Lecture 5: Reinforcement Learning: Learning from Return Instead of Labels

Lecture 5 of the 2026 course connects agent, environment, state, action, reward, and policy into an interaction loop, introducing credit assignment and exploration.

ai guide Reading MIT 6.S191

MIT 6.S191 Lecture 6: New Frontiers: Choosing the Problem Beyond the Model

Lecture 6 of the 2026 course places deep learning in emerging applications and real constraints, emphasizing data, outputs, evaluation, and failure conditions.

ai guide Reading MIT 6.S191

MIT 6.S191 Lecture 7: The Three Laws of AI: Safety Through Observability and Evaluation

Lecture 7 of the 2026 course starts from Asimov’s literary laws and examines modern safety protocols through traces, test data, and continuous evaluation.

ai guide Reading MIT 6.S191

MIT 6.S191 Lecture 8: AI for Science: Putting Domain Structure into Learning

Lecture 8 of the 2026 course uses the scientific-discovery loop to show how simulators, AI emulators, and experiments cooperate instead of reducing science to generic prediction.

ai guide Reading MIT 6.S191

MIT 6.S191 Lecture 9: Massively Parallel Training: Memory and Communication Set the Boundary

Lecture 9 of the 2026 course starts with GPU memory pressure and moves through checkpointing, offloading, ZeRO, FSDP, and multiple forms of parallelism.

ai guide Reading MIT 6.S191

MIT 6.S191 Lab 1: Generate Music with PyTorch and an LSTM

In the 2026 lab, students cover tensors, autograd, and modules before turning ABC notation into character sequences for LSTM music generation.

ai guide Reading MIT 6.S191

MIT 6.S191 Lab 2: From MNIST to Facial Debiasing with a DB-VAE

In the 2026 lab, part 1 classifies MNIST with dense and convolutional networks; Part 2 learns a facial latent distribution with a DB-VAE and changes training sampling.

ai guide Reading MIT 6.S191

MIT 6.S191 Lab 3: LoRA Fine-Tuning and LLM-as-a-Judge Evaluation

In the 2026 lab, students build chat templates and generation with LFM2-1.2B, adapt style through LoRA, and combine OpenRouter with Opik for a judge workflow.

ai deep-dive

n8n Deep Dive: From Triggers and AI Agents to Human Review and Operations

n8n is automation-first: a webhook, schedule, or application event starts a workflow, then an AI Agent may choose tools inside it; production still requires deliberate memory, approvals, credentials, execution data, and scaling architecture.

ai deep-dive

OpenRouter: One API Key for Multi-Model, Multi-Provider LLM Routing

OpenRouter exposes many models and inference endpoints through an OpenAI-compatible API, with provider ordering, failover, BYOK, and zero-data-retention controls in one routing policy.

ai deep-dive

OpenViking: Agent Memory as a Virtual Filesystem

Volcano Engine's open-source OpenViking stores agent memory, knowledge, and skills as a viking:// virtual filesystem — browsable with ls, tree, and find. Three-tier loading (L0/L1/L2) averages just 550 tokens per retrieval, boosting LoCoMo memory accuracy from 24–57% to 80–83%.

ai deep-dive

Parallel Web Systems: Search, Extraction, and Deep Research for Agents

Parallel Web Systems separates Search, Extract, and Task APIs into web-access layers with different latency and cost profiles, while Basis maps citations, excerpts, and confidence to output fields.

ai deep-dive

Patronus AI Deep Dive: From Evaluators and Experiments to Production Monitoring

Patronus AI treats evaluators as reusable scoring units, then applies them to offline experiments and production traces. It suits teams that want managed hallucination, safety, and multimodal evaluators, but judge scores cannot replace human labels, deterministic tests, or real security validation.

ai deep-dive

pgvector Deep Dive: Bringing Vector Search Back into PostgreSQL

pgvector is a PostgreSQL extension, not a standalone vector database. It adds exact and approximate vector search to the same data model, transactions, and operations stack, while leaving index tuning and horizontal scaling as PostgreSQL concerns.

ai deep-dive

Portkey: Put LLM Routing, Observability, and Governance Behind One AI Gateway

Portkey sits between applications and model providers: one OpenAI-compatible endpoint adds routing, fallbacks, request logs, budgets, and guardrails, with an open-source gateway available for self-hosting.

Securing Private-Corpus Queries: ACLs, Deletion Propagation, and Freshness

Authorization must take effect before candidate generation, while ACLs, deletion events, and source versions must propagate to every derived index; freshness needs measurable event-time SLOs too.

Private Corpus Search Boundaries: Decide Where Data May Go First

The first private-corpus decision is not which vector database to buy. Define data classes, trust zones, policy enforcement points, and freshness SLAs so every index, model, and observability system receives only the minimum data it is allowed to process.

Private-Corpus Retrieval Eval: Turning a Traditional Chinese Query Set into a Reproducible Benchmark

The repository has a 20-query Traditional Chinese/English golden dataset, but no document-level qrels, retrieval runs, raw latency data, or executable benchmark script. Reporting Recall@k, MRR, or nDCG as measured results would therefore be dishonest; this article defines the contract needed to run them reproducibly.

From Source to Index: Sync and Incremental Updates for Private Corpora

Private-corpus sync is not periodic refetching. It requires stable canonical IDs, source versions plus checksums for change detection, idempotent upserts, and tombstones that propagate deletion through every index.

ai deep-dive

Promptfoo Deep Dive: Local-First LLM Evaluation and Red Teaming

Promptfoo combines prompts, providers, test cases, and assertions in YAML to produce repeatable local and CI evaluation matrices, with red teaming against the same targets. It lowers the testing barrier but does not remove output variance, LLM-judge bias, or hosted data-flow concerns.

ai deep-dive

Pydantic AI: Building Python Agents with Types, Dependencies, and Validation

Pydantic AI models an agent as Agent[Deps, Output]: dependencies, tool inputs, and final outputs are typed, and model results must pass Pydantic validation.

ai deep-dive

R2R Deep Dive: Ingestion, Hybrid Search, and RAG behind an API

R2R packages document ingestion, hybrid search, knowledge graphs, RAG, Agents, and access controls behind a REST API; it fits teams that already own their product frontend and backend and need a retrieval service.

ai deep-dive

Choosing a RAG Framework: LlamaIndex, Haystack, RAGFlow, Dify, and R2R Operate at Different Layers

LlamaIndex and Haystack are code-first frameworks; RAGFlow and Dify are managed application platforms; R2R packages retrieval as an API service. Choose how much control your team needs over ingestion, retrieval, and operations before choosing a tool.

ai deep-dive

RAGFlow Deep Dive: From Document Parsing and Chunk Review to Cited Answers

RAGFlow puts document parsing, human chunk review, retrieval tests, chat, and citations in one platform; it fits layout-heavy PDFs and tables, but carries more deployment weight and platform state than a Python library.

ai deep-dive

Runloop: Devbox Infrastructure Built for Coding Agents

Runloop combines isolated microVMs, reproducible images, disk branching, credential proxies, and evals in one coding-agent platform; an official case study reports more than 10,000 concurrent Devboxes in one workload.

ai deep-dive

Sail Research: Trading Latency for Cost in Long-Horizon Agent Inference

Sail Research lets each inference request declare a completion window, scheduling patient background agents on cheaper capacity, while Sailboxes provide persistent long-running execution environments.

From Search Results to Reliable Citations: URL Deduplication, Source Tiers, and Claim-Source Mapping

Reliable citation is not appending URLs to an answer. Separate URLs, content copies, and source independence, then connect atomic claims to quote spans and snapshots through a rerunnable claim-source matrix.

ai guide

SerpAPI Complete Guide: Multiple Engines, Structured SERPs, and Async Queries

SerpAPI primarily manages search-results-page retrieval and parsing: select an engine, receive a structured SERP, then handle location, pagination, asynchronous polling, and validation in your application.

ai guide

Serper Search API Guide: Turn Google Results into Agent-Ready JSON

Serper is a third-party Google SERP API: one POST request returns structured JSON such as organic, knowledgeGraph, and peopleAlsoAsk, but production code still needs optional-field validation, URL checks, retries, and source verification.

ai deep-dive

Slack Code: Multiplayer AI Coding and the Agent Control Plane Landscape

Slack Code moves AI coding agents from individual terminals into shared Slack channels where teams can see diffs, previews, and plans in real time. But it solves management's visibility anxiety, not engineers' productivity bottleneck — the real battle is over who becomes the agent control plane.

CS221 Lecture 1: Overview: Defining Intelligence Under Resource Constraints

Lecture 1 of Stanford CS221 Autumn 2025 develops operational representations and algorithmic intuition through Overview: Defining Intelligence Under Resource Constraints.

CS221 Lecture 2: Learning I: From Computation Graphs to Linear Regression

Lecture 2 of Stanford CS221 Autumn 2025 develops operational representations and algorithmic intuition through Learning I: From Computation Graphs to Linear Regression.

CS221 Lecture 3: Learning II: Linear Classification, Features, and Cross-Entropy

Lecture 3 of Stanford CS221 Autumn 2025 develops operational representations and algorithmic intuition through Learning II: Linear Classification, Features, and Cross-Entropy.

CS221 Lecture 4: Learning III: Deep Networks as Composable Computation Graphs

Lecture 4 of Stanford CS221 Autumn 2025 develops operational representations and algorithmic intuition through Learning III: Deep Networks as Composable Computation Graphs.

CS221 Lecture 5: Search I: Define the State Before Choosing the Algorithm

Lecture 5 models search with states, actions, successors, and costs, then uses acyclic dynamic programming to show that an efficient algorithm still solves the wrong problem when state omits information needed by the future.

CS221 Lecture 6: Search II: Priorities in UCS and A*

Lecture 6 of Stanford CS221 Autumn 2025 follows the official material on Search II: Priorities in UCS and A* and makes its assumptions and limits explicit.

CS221 Lecture 7: MDPs I: Putting Uncertainty into State Transitions

Lecture 7 of Stanford CS221 Autumn 2025 follows the official material on MDPs I: Putting Uncertainty into State Transitions and makes its assumptions and limits explicit.

CS221 Lecture 8: MDPs II: Learning Q-Values Without a Transition Model

Lecture 8 of Stanford CS221 Autumn 2025 follows the official material on MDPs II: Learning Q-Values Without a Transition Model and makes its assumptions and limits explicit.

CS221 Lecture 9: MDPs III: Differentiating Expected Return Directly

Lecture 9 moves from tabular RL to function approximation, derives REINFORCE with the log-derivative identity, and connects the derivation to the executable PyTorch implementation.

CS221 Lecture 10: Games I: From Expectimax to Minimax

Lecture 10 extends single-agent search into adversarial game trees: expectimax averages chance outcomes, minimax takes the opponent's worst case, and alpha-beta removes irrelevant branches without changing the answer.

CS221 Lecture 11: Games II: TD Learning, Simultaneous Games, and Nash Equilibria

Lecture 11 first learns game values from experience with temporal-difference updates, then moves from sequential play to simultaneous games described by mixed strategies, minimax guarantees, and Nash equilibria.

CS221 Lecture 12: Bayesian Networks I: From Joint Distributions to Factorization

Lecture 12 builds a joint distribution from random variables and factors, then uses Bayesian-network factorization to express conditional independence and make conditioning and marginalization executable.

CS221 Lecture 13: Bayesian Networks II: Gibbs Sampling and the Markov Blanket

Lecture 13 replaces costly exact inference with Gibbs sampling: resample one variable at a time from a conditional determined by its Markov blanket, then approximate query probabilities with sample frequencies.

CS221 Lecture 14: Bayesian Networks III: From Counts and Smoothing to EM

Lecture 14 moves from maximum-likelihood counts and Laplace smoothing with complete data to EM, which alternates posterior responsibilities for latent variables with parameter updates.

CS221 Lecture 15: Logic I: Models, Entailment, and SAT

Lecture 15 separates propositional syntax from semantics: model checking defines entailment through satisfying assignments, SAT finds witnesses, and inference rules must be judged for both soundness and completeness.

CS221 Lecture 16: Logic II: Quantifiers Beyond Individual Propositions

Lecture 16 compresses knowledge across objects with predicates, quantifiers, and functions, then derives conclusions through substitution, unification, and definite-clause forward inference while exposing termination and completeness limits.

CS221 Lecture 17: Language Models: From Next-Token Prediction to Generation

Lecture 17 defines a language model as a chain-rule factorization of sequence probability, compares n-gram and neural conditional models, and shows how sampling, temperature, and evaluation shape generation.

CS221 Lecture 18: AI & Society: Benefits, Misuse, Accidents, and Institutions

Lecture 18 classifies AI's social effects as benefits, misuse, accidents, and structural harms, then connects fairness audits, research ethics, copyright, and platform terms to accountable institutional choices.

CS221 Lecture 19: AI Supply Chains: Resources, Labor, and Markets Behind Models

Lecture 19 uses the Economics of AI deck to connect compute, data, distribution, and organizational complements to GDP, labor, and ideas-driven growth.

CS221 Lecture 20: Fireside Chat, Conclusion: Turning Twenty Lectures into Modeling Choices

Lecture 20 is Percy Liang's fireside chat on career and research, CS221 and Stanford, and AI's future, with every attribution tied to the official video and editorial synthesis kept separate from auto-caption uncertainty.

Stanford CS224W Lecture 1: Introduction: Why Relational Data Needs Graph Machine Learning

A slide-grounded reconstruction of Fall 2025 Lecture 1, covering Course map and tools, A common language for graph data, Hand-designed features and representation learning while documenting the classroom material unavailable to self-learners.

Stanford CS224W Lecture 2: Node Embeddings: From Random Walks to node2vec

A slide-grounded reconstruction of Fall 2025 Lecture 2, covering Encoder-decoder view, Similarity and the objective, Random walks while documenting the classroom material unavailable to self-learners.

Stanford CS224W Lecture 3: Graph Neural Networks: A First Complete Message-Passing Model

A slide-grounded reconstruction of Fall 2025 Lecture 3, covering From fixed embeddings to deep encoders, The message-passing framework, Aggregation and update while documenting the classroom material unavailable to self-learners.

Stanford CS224W Lecture 4: A General Perspective on GNNs: Turning a Model into Design Components

A slide-grounded reconstruction of Fall 2025 Lecture 4, covering The GNN design space, Message, aggregation, and update, GraphSAGE while documenting the classroom material unavailable to self-learners.

Stanford CS224W Lecture 5: GNN Augmentation and Training: Co-designing Data, Tasks, and Models

A slide-grounded reconstruction of Fall 2025 Lecture 5, covering Graph-data augmentation, Feature and structural augmentation, Supervision and loss while documenting the classroom material unavailable to self-learners.

Stanford CS224W Lecture 6: Theory of GNNs: The WL Test, GIN, and Expressive Limits

A Fall 2025 slide-grounded reconstruction of Lecture 6, covering What distinguishability means, The Weisfeiler–Lehman test, An upper bound for message passing while documenting unavailable classroom material.

Stanford CS224W Lecture 7: Designing Powerful Graph Encoders: Structural and Positional Awareness

A Fall 2025 slide-grounded reconstruction of Lecture 7, covering The perfect-GNN thought experiment, Three levels of standard-GNN failure, Identity-aware encoding while documenting unavailable classroom material.

Stanford CS224W Lecture 8: Graph Transformers: Connecting Attention to Graph Structure

A Fall 2025 slide-grounded reconstruction of Lecture 8, covering Self-attention and message passing, The scope of graph attention, Positional and structural encodings while documenting unavailable classroom material.

Stanford CS224W Lecture 9: Heterogenous Graphs: Adding Node and Relation Types to Message Passing

A Fall 2025 slide-grounded reconstruction of Lecture 9, covering Heterogeneous graph schemas, Relation-specific messages, R-GCN while documenting unavailable classroom material.

Stanford CS224W Lecture 10: Knowledge Graphs: Modeling Relations with TransE, ComplEx, and RotatE

A Fall 2025 slide-grounded reconstruction of Lecture 10, covering Knowledge graphs and completion, Triple scoring, TransE and relation patterns while documenting unavailable classroom material.

Stanford CS224W Lecture 11: GNNs for Recommender Systems: From Collaborative Filtering to LightGCN

A Fall 2025 slide-grounded reconstruction of Lecture 11, covering Graph formulation of recommendation, The matrix-factorization baseline, Message passing in NGCF while documenting the public-material boundary.

Stanford CS224W Lecture 12: Relational Deep Learning: Turning Databases Directly into Prediction Graphs

A Fall 2025 slide-grounded reconstruction of Lecture 12, covering Limits of the tabular pipeline, Mapping relational databases to graphs, Temporal entity graphs while documenting the public-material boundary.

Stanford CS224W Lecture 13: Advanced Architectures in RDL: RelGNN and the Relational Graph Transformer

A Fall 2025 slide-grounded reconstruction of Lecture 13, covering The multi-relational bottleneck, RelGNN composite message passing, Relation-specific aggregation while documenting the public-material boundary.

Stanford CS224W Lecture 14: Advanced Topics in GNNs: In-Context Learning and Uncertainty on Graphs

A Fall 2025 slide-grounded reconstruction of Lecture 14, covering The goal of relational foundation models, Zero-shot relational transfer, PRODIGY's prompt graph while documenting the public-material boundary.

Stanford CS224W Lecture 15: Foundation Models for Knowledge Graphs: New Entities, New Relations, and Double Equivariance

A Fall 2025 slide-grounded reconstruction of Lecture 15, covering Limits of transductive KG embeddings, Entity-inductive link prediction, The relation graph while documenting the public-material boundary.

Stanford CS224W Lecture 16: LLM + GNN: Letting Language Models Read Graphs and Graph Models Read Text

A Fall 2025 slide-grounded reconstruction of Lecture 16, covering Complementary gaps in LLMs and GNNs, Text-attributed graphs, The LLM as predictor or encoder while documenting the public-material boundary.

Stanford CS224W Lecture 17: Agents + Graphs: Retrieval, Planning, and Action in Structured Worlds

A Fall 2025 slide-grounded reconstruction of Lecture 17, covering From graph QA to agents, Multimodal retrieval in STaRK, Tool use and traversal while documenting the public-material boundary.

Stanford CS224W Lecture 18: Deep Generative Models for Graphs: GraphRNN and Goal-Directed Molecular Generation

A Fall 2025 slide-grounded reconstruction of Lecture 18, covering The graph-generation problem and representation, Evaluating generation quality, GraphRNN's autoregressive factorization while documenting the public-material boundary.

Stanford CS224W Lecture 19: Ranking 315K GNN Designs with Anchor Models

The Fall 2025 conclusion studies roughly 315K GNN designs across 32 tasks: run a small set of anchor models, derive task similarity from rankings, and transfer the best designs from similar tasks.

Linear Regression: From LMS to Locally Weighted Regression

Linear regression is more than a best-fit line: Chapter 1 connects squared loss to gradient descent, normal equations, maximum likelihood, and locally weighted regression.

Classification and Logistic Regression: Decision Boundaries and Newton's Method

Chapter 2 derives logistic loss from a sigmoid probability model, then contrasts it with the perceptron and extends it through softmax and Newton's method.

Generalized Linear Models: Unifying Regression and Classification

Chapter 3 uses exponential families, natural parameters, and link functions to place least squares and logistic regression inside one modeling template.

Generative Learning Algorithms: GDA, Naive Bayes, and Smoothing

Chapter 4 models p(x|y) and p(y), using GDA, Naive Bayes, and Laplace smoothing to expose both the power and price of generative classification.

Kernel Methods: Nonlinear Learning Without Explicit Features

Chapter 5 replaces high-dimensional feature inner products with kernels, letting inner-product-based linear algorithms learn nonlinear functions without constructing the features.

Support Vector Machines: Margins, Duality, and SMO

Chapter 6 formalizes classification confidence as geometric margin, then builds an implementable SVM through Lagrange duality, kernels, and SMO.

Deep Learning: Modules, Backpropagation, and Vectorization

Chapter 7 decomposes neural networks into composable modules and uses backpropagation and vectorization to explain how deep models can be trained efficiently.

Generalization: Bias–Variance, Double Descent, and Sample Complexity

Chapter 8 decomposes test MSE into irreducible noise, squared bias, and variance, then uses uniform convergence and VC dimension to explain when training performance transfers to new data. Double descent shows why parameter count is not a universal measure of complexity.

Regularization and Model Selection: Explicit, Implicit, and Cross-Validated

Chapter 9 presents three controls on generalization: explicit complexity penalties, optimizer-induced implicit regularization, and model selection on data excluded from training. MAP estimation then connects a Gaussian prior to an L2 penalty.

Clustering and k-Means: A First Alternating-Optimization Algorithm

Chapter 10 introduces unsupervised learning through k-means: alternating updates make distortion non-increasing and numerically convergent, but do not guarantee a global optimum.

EM Algorithms: From Gaussian Mixtures to VAEs

Chapter 11 starts from soft assignments in Gaussian mixtures, uses Jensen's inequality to construct the ELBO, interprets EM as alternating maximization over a variational distribution and model parameters, and extends the idea to VAEs through approximate posteriors and reparameterization.

Principal Components Analysis: Projection, Reconstruction, and Reduction

Chapter 12 formulates PCA as geometric optimization: maximize projected variance along a unit direction to obtain the leading eigenvector of the covariance matrix. The top k eigenvectors give both maximum retained variance and minimum linear reconstruction error.

Independent Components Analysis: Recovering Independent Sources

Chapter 13 models ICA as x=As: observations are unknown linear mixtures, and the goal is to estimate W=A^{-1} to recover independent, non-Gaussian sources. A Jacobian determinant enters the transformed density and leads to the Bell–Sejnowski likelihood update.

Diffusion Models: Forward Noise, Reverse Generation, and the ELBO

Chapter 14 starts with a fixed Gaussian noising Markov chain and learns to reverse each transition. The ELBO turns reverse-kernel matching into weighted noise prediction, while the continuous-time view explains reverse drift through the score ∇log p_t.

Foundation Models Overview: Linear Probes, Fine-Tuning, and LoRA

Chapter 15 compares linear probing, full fine-tuning, and LoRA—not only by trainable parameter count, but by representation movement, data needs, and memory cost.

Representation Learning: Contrastive Learning, Retrieval, and RAG

Chapter 16 connects representation learning to systems: contrastive objectives shape an embedding space, semantic retrieval finds neighbors in it, and RAG passes retrieved context to a generator.

Large Language Models: Tokenization, Transformers, MoE, and SFT

Chapter 17 runs from next-token loss through Transformers, KV caches, MoE, and SFT, connecting an LLM's objective and architecture to its inference costs.

Reasoning in LLMs: Chain of Thought and Long-Reasoning RLVR

Chapter 18 separates two levers for LLM reasoning: chain of thought adds test-time computation, while verifiable rewards and policy gradients train long-reasoning behavior.

Reinforcement Learning: MDPs, Value Iteration, and Continuous States

Chapter 19 uses Bellman equations to turn long-horizon decisions into one-step updates, moving from value iteration in known MDPs to model learning and continuous-state approximation.

LQR, DDP, and LQG: From Linear Control to Uncertainty

Chapter 20 exploits linear dynamics and quadratic objectives to solve LQR, then uses DDP for local nonlinearity and Kalman filtering with LQG for partially observed state.

Policy Gradient and Its Variants: REINFORCE and PPO

Chapter 21 derives REINFORCE with the log-derivative trick, then uses reward-to-go, baselines, and PPO clipping to control policy-gradient variance and update size.

ai deep-dive

Steel Browser: An Open-Source Browser API and the Boundary of Self-Hosting

Steel packages Chromium sessions, CDP, proxies, stealth, and debugging behind an Apache-2.0 browser API. Its public repository has about 7,400 stars and it entered the Stripe Projects developer preview in 2026. Self-hosting fits development and data-control needs; Cloud addresses concurrency, managed proxies, CAPTCHA, recordings, and SLAs.

ai deep-dive

Together AI: From Serverless Inference to Dedicated Endpoints and Fine-Tuning

Together AI puts serverless APIs for open-weight models, dedicated GPU endpoints, batch inference, and fine-tuning on one platform, letting teams validate per token before moving to reserved deployment when traffic or customization justifies it.

ai debug

How to Write a Claude Code Skill That Doesn't Eat Context: Entry Point, Thresholds, Cost, Sources

I turned Anthropic's session-cost advice into a global skill, and the first version made the very mistakes it was meant to prevent: a description stuffed with trigger keywords, hard thresholds based on file counts and minutes, and 'protect the main context' conflated with 'spend fewer tokens overall'. Three rounds later the entry point is one page, details live in references, numeric thresholds became four judgment dimensions, and every claim from a draft post was checked against official docs.

ai deep-dive

Vapi: Managed Voice-Agent Orchestration and the Safety Boundaries Before Going Live

Vapi connects phone and web audio, STT, LLMs, TTS, tool calls, and call observability in a managed voice runtime. Providers are swappable, but Vapi's realtime orchestration is not portable. In May 2026, the company reported one million developers and announced a $50 million Series B.

ai deep-dive

Vercel Sandbox Deep Dive: Putting the Agent Execution Layer Inside the Vercel Ecosystem

Vercel Sandbox isolates untrusted code in Firecracker microVMs and integrates with Fluid compute, Active CPU pricing, and Vercel OIDC. It fits agents already running on Vercel, but network defaults, memory billing, and persistence still require deliberate design.

ai deep-dive

Vertex AI Explained: From Model APIs to Gemini Enterprise Agent Platform

Vertex AI is more than the Gemini API: it puts access to 200+ models, training, evaluation, deployment, and governance under one Google Cloud control plane. Since April 2026, its products and roadmap have moved into Gemini Enterprise Agent Platform, while the Vertex AI API, documentation paths, and many resource names remain in active use.

Web Extraction Quality Benchmark: Crawl4AI, Firecrawl, Jina Reader, and Readability

Extraction tools cannot be compared by HTTP 200s. The same 20 URLs must be scored for body text, headings, tables, code, links, metadata, noise, latency, and cost. This article publishes the corpus, adapter contract, and gates, but no winner without a same-version raw run across all four paths.

ai deep-dive

Zep Complete Guide: Temporal Knowledge Graphs for Agent Memory

Zep does not merely vectorize chat history. It turns episodes into entities and facts with validity time, allowing new information to invalidate an old relationship without erasing history. Graphiti is the open-source framework; Zep adds managed scale and governance.

career guide

Beyond Upwork: Seven Remote Work Platforms Worth Bookmarking

From the free We Work Remotely to the elite 3%-acceptance Toptal, seven platforms each serve a different niche — job boards (FlexJobs, WWR, Remote OK), startup talent (Wellfound), community (Remotive), market research (Working Nomads), and premium freelancing (Toptal). All friendlier than competing on price at Upwork.

CS188 Bayes Nets and Ghostbusters: Inference When Ghosts Are Invisible

Lectures 13–18 and Project 4 move from factor operations and variable elimination to exact inference and particle filtering, letting Pacman track invisible ghosts through noisy distance sensors.

Completing CS188: Turn 28 Lectures and Projects P0–P5 into a Portfolio

Lectures 26–28 close with nuclear monitoring, AI safety, and reflection. Independent completion should preserve assumptions, test evidence, and failure analysis for Projects 1–5 instead of reporting only autograder scores.

CS188 CSPs and Multi-Agent Search: Choosing Minimax, Alpha-Beta, and Expectimax

Lectures 5–8 use CSPs to practice variables, constraints, and search order before Project 2 implements minimax, alpha-beta, and expectimax. Their key difference is the assumption made about other agents.

CS188 Decisions and Machine Learning: From VPI and Naive Bayes to Attention

Lectures 19–25 connect rational decisions and VPI to machine learning, while Project 5 uses PyTorch for regression, classification, CNNs, attention, and an optional character-GPT.

CS188 MDPs and Reinforcement Learning: From Value Iteration to Q-Learning

Lectures 9–12 and Project 3 use the same Gridworld to contrast value iteration with a known model, Q-learning from unknown dynamics, and approximate Q-learning that generalizes through features.

CS188 Search and Heuristics: Pacman from DFS and BFS to A*

Lectures 1–4 and Project 1 connect DFS, BFS, UCS, A*, state representation, and heuristic design. The goal is not memorizing algorithms but separating what the frontier, cost, and state each control.

Berkeley CS188 Spring 2026: Learn AI Through Projects P0–P5

CS188 Spring 2026 publishes 28 recordings, 27 lecture slide sets, 11 discussions, and Projects P0–P5. P0 is a Python/autograder tutorial, P1–P4 use Pacman settings, and P5 contains general machine-learning tasks.

Berkeley CS189 Spring 2025 Overview: HW1–7 with Code and Data You Can Run, Plus What Fall 2026 Looks Like

Spring 2025 at people.eecs.berkeley.edu/~jrs/189s25 is the only A3 self-study edition with notes, videos, HW1–7, code/data and past exams; Fall 2026 at eecs189.org/fa26 has a 27-lecture schedule but most materials are not yet released and the rotating site can 404 old URLs.

Berkeley CS285 L19–25: Exploration, RL Theory, Multitask Learning, and Open Problems

The final seven lectures move from exploration and theoretical limits through two review lectures to advanced exploration, multitask RL, and unresolved research problems.

Berkeley CS285 Homework and Final Projects: The CPU, GPU, and H100 Boundary

Five assignments move from CPU-friendly imitation learning to H100-based LLM RL and six-hour offline-RL runs; self-learners should use three compute tiers instead of copying the entire enrolled workflow.

Berkeley CS285 L1–4: Imitation Learning, Distribution Shift, and RL Basics

The first four lectures move from behavioral cloning to MDPs; HW1 turns distribution shift into an observable failure through MSE policies, DAgger, and flow matching.

Berkeley CS285 L11–18: From Variational Inference and LLM RL to Offline RL

L11–18 connect control as inference, LLM RL, model-based RL, and offline RL, with HW4 and HW5 providing two compute-intensive implementations.

Berkeley CS285 L5–10: Policy Gradients, Actor-Critic, DQN, and SAC

L5–10 build the deep-RL core through policy- and value-based routes; HW2 is CPU-friendly, while HW3's Atari and HalfCheetah runs can require hours of GPU time.

Berkeley CS285 Spring 2026 Guide: 25 Lectures, Five Assignments, and the Self-Study Boundary

Spring 2026 CS185/285 publishes slides for 25 lectures, nine discussion units, five assignments, and starter code; current recordings require bCourses access, while HW4 defaults to an H100, so this is not a zero-cost open course.

Berkeley CS288 Part 5: Inference-time Compute, Reasoning, and Embodied Agents

Units 15–18 place NLP models inside perception, reasoning, tool, and environment loops; the question shifts from next-token prediction to allocating inference compute and validating multi-step action.

Berkeley CS288 Part 1: From N-grams and Word Representations to Text Classification

The first four units make text countable, representable, and classifiable; A1 then moves from n-grams and perceptrons to an NBOW MLP.

Berkeley CS288 Spring 2026: 18 Slide Units, Three Assignments, and the Limits of Self-Study

CS288 moves from n-grams to RAG, reasoning, and agents through 18 public slide units and three assignments; Berkeley-only recordings make this an A3 materials route, not a public video course.

Berkeley CS288 Part 3: Pre-training, Post-training, Generation, and Evaluation

Units 08–12 turn a base model into an interactive system: pre-training establishes capability, post-training shapes behavior, and generation plus evaluation determine how outputs are used.

Berkeley CS288 Part 4: Turning Retrieval, RAG, and Advanced Architectures into a System

Units 13–14 connect models to external knowledge; A3 requires data collection, QA annotation, indexing, and ablations under CPU and latency constraints.

Berkeley CS288 Part 2: Sequence Models, Seq2Seq, and Transformers

Units 05–07 move from recurrent state to encoder-decoder models, then rewrite the information path with attention and Transformer blocks.

CMU 07-380 Fall 2026 Overview: 26 Lectures from Logic and Planning to Diffusion, HW and Project Not Yet Fully Released

07-380 Fall 2026 is the first offering of CMU's new AI II, 26 lectures from logic, planning and optimization to probabilistic graphs and generative systems; Lec01 and Prop Logic are public, HW1-7, six quizzes and the final project are still TBD — an A2→A3 transition with the 07-280 bridge.

CMU 10-301 HW1: Find ML Foundation Gaps with Mathematics and Python

HW1 is written and programming work: mathematical and CS foundations followed by a majority-vote classifier.

CMU 10-301 HW2: From Information Calculations to a Complete Decision Tree

HW2 moves from hand-calculated entropy and mutual information to an end-to-end tree learner, predictor, and evaluator.

CMU 10-301 HW3: Compare K-NN, Perceptron, and Linear Regression

HW3 is written work: a decision-tree review followed by K-NN, Perceptron, and Linear Regression through inductive bias, errors, and model selection.

CMU 10-301 HW4: Turn Logistic Regression Likelihood into a Classifier

HW4 joins probabilistic interpretation, cross-entropy gradients, and implementation into one traceable training pipeline.

CMU 10-301 HW5: Expose Neural Networks and Backpropagation with NumPy

HW5 avoids automatic differentiation so learners must track forward shapes, caches, and backward gradients themselves.

CMU 10-301 HW6: Learning Theory, MLE/MAP, and Fairness Metrics

HW6 combines generalization, MLE/MAP, probabilistic learning, fairness metrics, and social impact in one written assignment about assumptions and tradeoffs.

CMU 10-301 HW7: Move from Basic Neural Networks to Deep Learning

HW7 builds on HW5 backpropagation to address deep-model architecture and training failures, emphasizing diagnosis over merely adding layers.

CMU 10-301 HW8: From MDPs to Reinforcement-Learning Updates

HW8 connects states, actions, rewards, transitions, and value updates while separating environment dynamics, policy, and estimation error.

CMU 10-301 HW9: Close the Course with Ensembles, k-Means, PCA, and Recommenders

The final written assignment combines ensembles, clustering, representation, and recommendation to test whether you can choose a learning paradigm from problem structure.

CMU 10-301/601 Spring 2026: Learn Machine Learning Through Nine Assignments

Spring 2026 publishes material for 27 lectures and nine homework bundles; outsiders can do the core work but cannot access Panopto, Piazza, Gradescope, or official homework solutions.

learning deep-dive

CMU's AI Core Redesign: From 15-281 + 10-315 to 07-280 + 07-380

In 2026, CMU recombined its separate general-AI and SCS machine-learning introductions into the 07-280 → 07-380 sequence. This is a redistribution of content and prerequisites, not a pair of simple course renames.

Stanford CS107 Lecture 4: Bitwise Operators, Conversions, and Masks

Lecture 4 first shows that signed/unsigned conversion can preserve bits while changing meaning, that mixed comparisons may surprise, and how sign extension, zero extension, and truncation alter width. It then derives AND, OR, NOT, XOR, and bitmask idioms for testing, setting, clearing, and combining fields.

Stanford CS107 Lecture 3: Integers, Bytes, and Two's Complement

Lecture 3 starts with 32/64-bit address spaces, derives the ranges of unsigned and two's-complement signed integers, inversion-plus-one, and shared addition hardware, then separates unsigned modular arithmetic from C signed overflow and tests the model against four failure cases.

Stanford CS107 Lecture 5: Bit Shifts, Bit Tricks, and GDB

Lecture 5 extends masks to shifts, power-of-two and popcount tricks, then uses an absolute-value example to expose signed intermediate overflow at INT_MIN. Its second half establishes a GDB workflow around breakpoints, execution control, formatted printing, memory examination, and backtraces.

Stanford CS107 Lecture 2: A First C Program, Binary, and Hexadecimal

Lecture 2 puts C back into its Unix history and development environment: headers, main, printf, argc/argv, ssh, emacs, make, and executables. It then derives 8 bits = 1 byte, 256 byte patterns, and reliable conversion among decimal, binary, and hexadecimal.

Stanford CS107 Lecture 1: From the Course Map to the Unix Command Line

Winter 2026 opens by explaining why CS107 goes below programming-language abstractions: from bytes and memory through assembly and heap allocators. It then lays out the 40/10/20/30 grading structure and closes with a first tour of the Unix command line.

Harvard AI/ML Course Guide: Do CS50 AI, CS181, and CS182 Videos Match Their Assignments?

CS50 AI is Harvard's most complete public entry point, but the Summer 2026 course still uses 2020 recordings and assignment assets while the rolling OCW projects have moved to other editions. CS181 Spring 2026 exposes current homework and notes without current recordings; CS182 Fall 2026 has not yet completed an offering.

learning deep-dive

The Pacman AI Project Lineage: How Berkeley CS188 and CMU 15-281 Restructure the Same Material

CMU 15-281's Search and Games explicitly credits Berkeley's Pacman AI projects. The official course site separately lists a zero-point P0 tutorial and five programming assignments, P1–P5.

Stanford CS103 Lecture 0: From Set Language to Cantor's Diagonal

Starting with elements, subsets, and power sets, this lecture culminates in Cantor's diagonal proof that no set is as large as its own power set.

Stanford CS103 Lecture 1: Building a First Direct Proof from Even and Odd

The even-square and odd-sum examples show how arbitrary choices, assumptions, witnesses, and a want-to-show become a checkable direct proof.

Stanford CS103 Lecture 2: Negation, Contraposition, and Contradiction

This lecture identifies exactly when an implication is false, then turns quantified negation, contraposition, and contradiction into checkable proof tools.

Stanford CS103 Lecture 3: Propositional Logic, Truth Tables, and Equivalence

Propositional logic abstracts English statements into Boolean variables, then uses truth tables to check connectives, translation direction, and equivalences.

Stanford CS103 Lecture 4: Objects, Quantifiers, and Types in First-Order Logic

This lecture extends propositional logic into a language about objects: distinguish constants, predicates, functions, and propositions, then express some and every with existential and universal quantifiers.

Stanford CS103 Lecture 5: First-Order Logic II—Nested Quantifiers, Negation, and Uniqueness

Translate natural language one layer at a time: identify universal and existential forms, then handle quantifier order, negation, restricted quantifiers, and uniqueness.

Stanford CS103 Lecture 6: Functions I, from Definitions to Injection and Surjection Proofs

A function is more than a formula: domain, codomain, totality, and determinism are essential, while the quantifiers defining involutions, injections, and surjections dictate their proofs.

Stanford CS103 Lecture 7: Functions II—Surjections, Assumptions, and Composition

This lecture uses surjections and a proof about birds to separate assuming from proving, then shows that involutions are injective and surjective and carries those ideas into function composition.

Stanford CS103 Lecture 8: Cardinality by Bijections and Cantor's Diagonal Argument

Two sets have equal cardinality when a bijection pairs their elements; Cantor's diagonal set defeats every function from S to its power set by constructing a value it misses.

Stanford CS103 Lecture 9: Graphs, Part I

This lecture moves from the formal definitions of graphs and digraphs to independent sets, vertex covers, and their complement relationship.

Stanford CS103 Lecture 10: Walks, Graph Complements, and the Pigeonhole Principle

Starting with walks, paths, cycles, and components, this lecture proves that a graph or its complement is connected and develops the pigeonhole principle through degrees and monochromatic triangles.

Stanford CS103 Lecture 11: Generalized Pigeonhole, Ramsey Theory, and Average Load

Use the generalized pigeonhole principle to force a monochromatic triangle at a six-person party, then solve a movie-preference puzzle through average load and contradiction.

Stanford CS103 Lecture 12: Induction, Counterfeit Coins, and Invariants

Induction is not a list of checked examples: establish a true starting point, prove that an arbitrary true case transmits truth to the next case, and invoke the induction principle.

Stanford CS103 Lecture 13: Mathematical Induction, Part II

This lecture connects starting from ordinary induction to induction may start later, following the official examples and proof obligations.

Stanford CS103 Lecture 14: Finite Automata, Part I

This lecture connects why begin with a weak computer to from device behavior to a state machine, following the official examples and proof obligations.

Stanford CS103 Lecture 15: Finite Automata, Part II

This lecture connects the dfa definition connects the first half of cs103 to regular means that some dfa exists, following the official examples and proof obligations.

Stanford CS103 Lecture 16: Finite Automata, Part III

This lecture connects the automata ladder measures power with languages to dfa transition tables, following the official examples and proof obligations.

Stanford CS103 Lecture 17: Regular Expressions

This lecture connects from closure properties to a language syntax to regex is mathematics, not one library, following the official examples and proof obligations.

Stanford CS103 Lecture 18: Nonregular Languages

This lecture connects four equivalent descriptions of regularity to the precise finite-memory intuition, following the official examples and proof obligations.

Stanford CS103 Lecture 19: Context-Free Languages

This lecture connects from finite-state limits to recursion to the arithmetic grammar, following the official examples and proof obligations.

Stanford CS103 Lecture 20: Turing Machines, Part I

This lecture connects why the model changes after cfgs to long addition and local access, following the official examples and proof obligations.

Stanford CS103 Lecture 21: Turing Machines, Part II

This lecture connects the sample tm looks back from the end to beyond pairwise marking, following the official examples and proof obligations.

Stanford CS103 Lecture 22: Turing Machines, Part III

This lecture connects a quick quantifier audit for recognizers and deciders to why every decision problem can be represented as a language, following the official examples and proof obligations.

Stanford CS103 Lecture 23: Unsolvable Problems, Part I

This lecture connects returning from r, re, and utm to three self-reference warm-ups, following the official examples and proof obligations.

Stanford CS103 Lecture 24: Unsolvable Problems, Part II

This lecture connects defining and locating halt to why halt is recognizable, following the official examples and proof obligations.

Stanford CS103 Lecture 25: Unsolvable Problems, Part III

This lecture connects the lava diagram's two classification tasks to the deck's operational reading of rice's theorem, following the official examples and proof obligations.

Stanford CS103 Lecture 26: Complexity Theory

This lecture connects decidable does not mean feasible to efficiency requires choosing a resource, following the official examples and proof obligations.

Stanford CS103 Wrap-Up: Four Foundations and Where to Go Next

The final deck reconnects proofs, graphs, automata, and computability, then maps those foundations to Stanford courses that use them.

Stanford CS107 Lecture 15: Reading x86-64 Addressing Modes Without Confusing Addresses and Values

CS107 Lecture 15 decomposes x86-64 mov operands into immediate, register, absolute, indirect, displacement, indexed, and scaled-indexed forms, then unifies pointer dereference and array access with D + R[b] + R[i]×s.

Stanford CS107 Lecture 16: From Subregisters to x86-64 Arithmetic and Logic

CS107 Lecture 16 connects b/w/l/q data widths, subregisters, movs/movz, lea, calling conventions, arithmetic and logic, and shifts through one method: establish operand width before tracing sources, destinations, and real memory accesses.

Stanford CS107 Lecture 18: From Condition Codes to x86-64 Loops

CS107 Lecture 18 connects ZF/SF/CF/OF to cmp, test, signed and unsigned conditional jumps, then reconstructs if statements, loops, dynamic instruction counts, setcc, and cmovcc.

Stanford CS107 Lecture 17: From Multiply and Divide to x86-64 Control Flow

CS107 Lecture 17 completes full-width x86-64 multiplication and division, traces %rip through instruction bytes, and uses direct and indirect jmp to show how execution leaves its default sequential path.

Stanford CS107 Lecture 19: Understanding x86-64 Function Calls and Calling Conventions

CS107 Lecture 19 traces %rsp, push/pop, call/ret, parameters, return values, stack locals, and caller/callee register discipline to build the ABI contract that preserves data and control across functions.

Stanford CS107 Lecture 14: From C to x86-64, Reading Disassembly for the First Time

CS107 Lecture 14 dissects the ten x86-64 instructions for sum_array: addresses and machine bytes appear on the left, AT&T assembly on the right, and the reader's job is to recover C-level effects from opcodes, operands, registers, and control flow—not to write assembly from scratch.

Stanford CS107 Lecture 7: From String Search to Buffer Overflows—Input Validation Is Not Capacity Checking

CS107 Lecture 7 builds pointer-based string scanning with strchr, strstr, and strspn, then shows why valid content can still overflow a buffer: safety requires input rules, destination capacity, termination, and memory-error detection.

Stanford CS107 Lecture 25: Caching, Memory Hierarchy, and Locality

CS107 Lecture 25 builds the essential cache model from a concise deck: memory access costs are nonuniform, smaller and faster layers retain data likely to be reused, and temporal and spatial locality determine whether a program benefits.

Stanford CS107 Lecture 6: A C String Is Not a Type but a Memory Contract

CS107 Lecture 6 reduces C strings to character arrays, a terminator, and an address: every convenience in strlen, strcmp, strcpy, strncpy, and strcat depends on the caller preserving capacity and termination invariants.

Stanford CS107 Lecture 24: Profile with Callgrind, Then Read What GCC Optimized

CS107 Lecture 24 builds a measurement workflow with matrix multiplication and Callgrind, then examines GCC constant folding, common-subexpression elimination, dead-code elimination, strength reduction, code motion, and recursion-to-loop conversion. Optimization starts with bottleneck evidence.

Stanford CS107 Lecture 23: The Allocator Invariants Behind In-Place realloc

CS107 Lecture 23 advances the explicit free list to in-place realloc: split a useful remainder when shrinking, absorb free right neighbors when growing, and allocate-copy-free only as a fallback, while preserving both the physical heap and logical list.

Stanford CS107 Lecture 13: From Comparators to a Fully Generic Bubble Sort

CS107 Lecture 13 upgrades a Boolean callback to a three-way comparator, then combines void *, element width, and const void * callbacks into a fully generic bubble sort before mapping the design to qsort, bsearch, lfind, and lsearch.

Stanford CS107 Lecture 12: Function Pointers Inject Ordering into Generic C

CS107 Lecture 12 first uses char * for byte-wise generic swap and rotate, then uses a function pointer to separate bubble sort's traversal mechanism from its ordering rule: void * abstracts data types, while callbacks abstract behavior.

Stanford CS107 Lecture 11: How void * Gives C Generics Without Pretending Types Still Exist

CS107 Lecture 11 finishes the heap contracts of calloc, strdup, free, and realloc, then turns several typed swap functions into void * plus a byte count: C generics do not preserve an unknown type; they explicitly transfer responsibility for addresses, widths, and interpretation.

Stanford CS107 Lecture 21: A First Heap Allocator and the Tension Between Speed and Space

CS107 Lecture 21 starts with alignment, throughput, and utilization, then uses a bump allocator and an implicit free list to explain metadata, splitting, placement, internal and external fragmentation, and the need to coalesce freed blocks.

Stanford CS107 Lecture 22: Why an Explicit Free List Lives in Two Orders at Once

CS107 Lecture 22 replaces an implicit list with an explicit free list. Searches visit only reusable blocks, but every free block now has both physical neighbors and logical links, so unlinking, coalescing, and reinsertion must preserve both structures.

Stanford CS107 Lecture 8: A Pointer Is Not Magic, but a Copyable Address

CS107 Lecture 8 starts with address-of and dereference, explains why C pointer parameters are still passed by value, and shows how int *, char *, and char ** can modify caller-owned ints, chars, and pointers respectively.

Stanford CS107 Lecture 9: An Array Is Not a Pointer, but They Cooperate in Expressions

CS107 Lecture 9 uses seven C-string rules to separate array objects, pointer variables, and string literals: arrays often convert to first-element pointers in expressions, but storage, assignment, mutability, and sizeof remain different.

Stanford CS107 Lecture 20: After Reverse Engineering, Ask About Privacy and Trust Before Building a Heap Allocator

CS107 Lecture 20 places reverse-engineering capability in an ethical context: privacy has individual and social models, while trust combines reliance with a risk of betrayal. It then reviews process memory and shifts from heap-allocation client to allocator implementer.

Stanford CS107 Lecture 10: Stack vs. Heap Is About Lifetime and Ownership, Not Just Speed

CS107 Lecture 10 moves from sizeof and pointer arithmetic to stack-frame lifetime: returning a local array leaves a dangling pointer; malloc crosses function returns but makes NULL handling, size arithmetic, ownership, free, and leaks the programmer's responsibility.

Stanford CS107 Lecture 26: Wrap-up, Six Systems Questions, and What Comes Next

CS107 Lecture 26 closes ten weeks through six big questions: representation, text, memory, generics, execution, and allocation. It checks the learning goals through the explicit allocator and points toward CS111 and other systems courses.

Stanford CS109 Lecture 1 | What is Probability?: List outcomes first; only then assign probabilities to events.

List outcomes first; only then assign probabilities to events.

Stanford CS109 Lecture 2 | Conditional Probability: A condition restricts the sample space to outcomes still compatible with the evidence.

A condition restricts the sample space to outcomes still compatible with the evidence.

Stanford CS109 Lecture 3 | Bayes Theorem: Bayes’ theorem turns an easier generative direction into the inferential direction we need.

Bayes’ theorem turns an easier generative direction into the inferential direction we need.

Stanford CS109 Lecture 4 | Counting and Combinatorics: Decide whether order matters and repetition is allowed before choosing a formula.

Decide whether order matters and repetition is allowed before choosing a formula.

Stanford CS109 Lecture 5 | Random Variables and Expectation: A random variable maps outcomes to numbers; expectation is a weighted average, not necessarily an attainable value.

A random variable maps outcomes to numbers; expectation is a weighted average, not necessarily an attainable value.

Stanford CS109 Lecture 6 | Moments: Expectation, LOTUS, and linearity

Expectation compresses a distribution into a weighted average; LOTUS handles transformed values, while linearity makes sums tractable even without independence.

Stanford CS109 Lecture 7 | Variance and Poisson: From spread to rare-event counts

Variance describes a random variable's spread; Poisson models counts in a fixed interval and approximates a large-n, small-p binomial.

Stanford CS109 Lecture 8 | Continuous Random Variables: PDFs, CDFs, Uniform, and Exponential

A continuous variable assigns zero probability to a point and area to intervals; CDFs, Uniform, Exponential, and memorylessness build on that distinction.

Stanford CS109 Lecture 9 | Normal Distribution: Standardization, Phi, and continuity correction

Standardization maps Normal variables to Z; Phi, linear transforms, and continuity correction turn intervals and large binomials into computable probabilities.

Stanford CS109 Lecture 10 | Probabilistic Models: Joints, marginals, independence, and Bayes

A joint distribution retains the full relationship among variables; marginals, conditionals, independence, and Bayes extract different answers from it.

Stanford CS109 Lecture 11 | Inference: Prior times likelihood, then normalize

Inference multiplies each hidden-variable prior by an observation likelihood and normalizes; the same loop handles repeated evidence and discretized continuous beliefs.

Stanford CS109 Lecture 12 | General Inference: Bayesian networks, sampling, and rare evidence

A Bayesian network factorizes a huge joint through conditional independence; ancestral sampling generates joint samples, and rejection sampling filters them into a conditional.

Stanford CS109 Lecture 13 | Multinomial: Category counts, bag of words, and log probability

The Multinomial extends two-category Binomial counts to many categories; the same PMF models documents as word counts for Bayesian authorship with log-scores.

Stanford CS109 Lecture 14 | Beta: Turn an unknown probability into an updatable random variable

A Beta distribution represents full belief about an unknown success rate; success/failure data updates two parameters for posteriors, smoothing, and Thompson-sampling decisions.

Stanford CS109 Lecture 15 | Adding Random Variables and the Central Limit Theorem

A few independent sums have closed forms; general IID sums become approximately Normal under the CLT, with continuity correction for discrete sums.

Stanford CS109 Lecture 16 | Bootstrapping: Sampling statistics, error bars, and p-values

The bootstrap treats a sample histogram as a population proxy, resampling with replacement to approximate a statistic's sampling distribution, error bar, or null p-value.

Stanford CS109 Lecture 17 | Algorithmic Analysis: Conditional expectation, indicators, and recursion

Expected cost in randomized code can be conditioned on the first random choice; counting problems become indicator sums, often avoiding the full distribution entirely.

Stanford CS109 Lecture 18 | Information Theory: Surprise, entropy, information gain, and KL

Surprise turns rare events into bits; entropy is expected surprise, information gain selects uncertainty-reducing questions, and KL measures excess cost from a model distribution.

Stanford CS109 Lecture 19 | Maximum Likelihood Estimation: Hold data fixed and optimize the parameter

MLE fixes observed data and optimizes parameters; log-likelihood turns products into sums, but a maximum can also lie on a boundary.

Stanford CS109 Lecture 20 | Logistic Regression: Derive the gradient from Bernoulli likelihood

Logistic regression turns a linear score into a Bernoulli probability with sigmoid; the gradient xⱼ(y-ŷ) follows directly from the log-likelihood chain rule.

Stanford CS109 Lecture 21 | Comparing Classifiers: Beyond accuracy to calibration, error costs, and fairness

Classifier comparison requires held-out data, baselines, calibration, precision/recall, and an explicit fairness criterion—not accuracy alone.

Stanford CS109 Lecture 22 | Deep Learning: Derive backpropagation with the chain rule

A neural network stacks logistic units; a forward pass computes probabilities, while backpropagation reuses output error to obtain every gradient.

Stanford CS111 Lecture 1: Welcome to CS111!

Lecture 1 follows shared I/O cards in the 1940s, batch processing, multiprogramming, and personal computers to explain how OS responsibilities accumulated as hardware costs and user needs changed.

Stanford CS111 Lecture 2: Threads, Processes, and Dispatching

Lecture 2 defines shared and private process/thread state, then uses fork, execvp, waitpid, and thread creation to show how the kernel creates execution units.

Stanford CS111 Lecture 3: Threads, Processes, and Dispatching, Continued

Lecture 3 follows running, blocked, and ready transitions to show how PCBs, context save/restore, and the dispatcher complete one CPU-control handoff.

Stanford CS111 Lecture 4: Concurrency

Lecture 4 defeats each Too Much Milk attempt with an explicit schedule, deriving race condition, atomicity, critical section, and synchronization requirements from concrete interleavings.

Stanford CS111 Lecture 5: Mutexes, Condition Variables, and Mesa Semantics

Lecture 5 uses an eight-slot circular Pipe to prove that a mutex supplies exclusion, while a condition variable atomically releases the lock and blocks when a predicate is false; under Mesa semantics, wait must return to a while loop that rechecks the predicate.

Stanford CS111 Lecture 6: Implementing Locks

Lecture 6 evolves a one-core interrupt-masking lock through multicore version 5, tracking guard, lock, and wait-queue state to prevent races and lost wakeups.

Stanford CS111 Lecture 7: Deadlock Conditions and Global Lock Ordering

Lecture 7 extracts four necessary deadlock conditions from request/ownership graphs, then compares detection, prevention, and lock ranking; breaking circular wait is common in practice, but every module must obey one global order.

Stanford CS111 Lecture 8: FIFO, Round Robin, Priorities, and Multicore Scheduling

Lecture 8 moves from FIFO and round robin through the unimplementable SRPT ideal to adaptive priority queues and the multicore conflict among queue contention, core affinity, and work conservation.

Stanford CS111 Lecture 9: Linkers and Dynamic Linking

Lecture 9 follows source through assembly, object, executable, and process, explaining the linker's three passes and how a dynamic loader resolves shared-library addresses through a jump table at startup.

Stanford CS111 Lecture 10: Dynamic Storage Management

Lecture 10 moves from predictable LIFO stacks to heap free lists, first/best fit, and slabs, then compares reference counting with mark-and-sweep across dangling pointers, leaks, cycles, and fragmentation.

Stanford CS111 Lecture 11: Dynamic Storage Management, Continued

Lecture 11's official PDF is byte-identical to Lecture 10; this article preserves that artifact gap and focuses on reachability, dangling pointers, leaks, reference-count cycles, and mark/compact garbage collection.

Stanford CS111 Lecture 12: Trust and Operating Systems

Lecture 12 defines trust as voluntary vulnerability, separates over-trust from untrustworthiness, and applies assumption, inference, and substitution to the Linux TCB, the xz attack, and AI-code policy.

Stanford CS111 Lecture 13: Virtual Memory

Lecture 13 starts from the failures of single-tasking and load-time relocation, uses an MMU with base/bound to create isolated virtual and physical address spaces and traps, then introduces segmentation to escape one contiguous region.

Stanford CS111 Lecture 14: Virtual Memory, Continued

Lecture 14's official PDF is byte-identical to Lecture 13; this article records the gap and focuses on how multiple base/bound/protection entries enable growth, sharing, and compaction while retaining fixed-count, fragmentation, and rigid-layout limits.

Stanford CS111 Lecture 15: Paging

Lecture 15 uses fixed pages to remove inter-process external fragmentation, then connects x86-64's four-level walk, sharing and aliasing, and the TLB to trade-offs among translation speed, sparse tables, context switches, and page size.

Stanford CS111 Lecture 16: Page Faults, Demand Fetching, and Prefetch

Demand paging loads pages only when needed; present bits, precise exceptions, and restartable instructions let the kernel safely fill them from executables, zero-fill, or backing store.

Stanford CS111 Lecture 17: From Page Faults to Clock—Who Leaves When Memory Is Full?

Lecture 17 separates demand paging into fetching and replacement: MIN cannot know the future, exact LRU is too expensive, and Clock uses reference/dirty bits to find a page old enough to evict; when active working sets exceed RAM, even a 1% fault rate can cause an approximately 1,000-fold slowdown.

Stanford CS111 Lecture 18: Disk Geometry, Interrupts, and DMA

A disk hides mechanical seek and rotation behind a linear block API; modern I/O then uses memory-mapped registers, DMA queues, and interrupts so the CPU mainly issues commands and receives completions.

Stanford CS111 Lecture 19: File Abstractions, Allocation, and FAT

A file system maps durable byte collections onto disk blocks; contiguous, linked, and FAT allocation trade locality, growth, random access, and metadata cost.

Stanford CS111 Lecture 20: Multilevel Inodes, Index Walks, and Disk Scheduling

The 4.3BSD inode uses direct, single-indirect, and double-indirect tiers so lookup depth scales with file size; FIFO, SPTF, SCAN, and CSCAN then trade seek cost, fairness, and wait time.

Stanford CS111 Lecture 21: Block Cache, Free Bitmaps, and Delayed Allocation

Block cache retains hot indexes, bitmap slack preserves placement choices, and fragments plus delayed allocation trade later, better information for locality.

Stanford CS111 Lecture 22: Directory Lookup, Hard Links, and Symbolic Links

Directories map text names to file-system-local inode numbers; hard links share inode identity and reference counts, while symlinks store paths and permit cross-filesystem references with loops and dangling targets.

Stanford CS111 Lecture 23: From fsck and Ordered Writes to Write-Ahead Logging

A single file-system operation updates several blocks, but a crash can occur between any two writes; this lecture compares how fsck, ordered writes, and write-ahead logging trade recovery time, performance, durability, and consistency.

Stanford CS111 Lecture 24: Journaling, Transactions, and Checkpoints

Lecture 24 continues from the WAL entry point into transactions, idempotent replay, and checkpoints, showing why consistency is not durability and why a journal does not replace fsync or backups.

Stanford CS111 Lecture 25: Truth, Trust, and Technology—How Algorithms, Generative AI, and Deepfakes Reshape Trust

Lecture 25 separates assumption, inference, and substitution as ways to establish trust, then examines how social recommendations, generative AI, and synthetic media amplify over-trust; the response is preserved provenance, independent validation, and coordinated responsibility.

Stanford CS111 Lecture 26: Flash Translation Layers, Garbage Collection, and Wear Leveling

Flash programs pages but erases whole units; an FTL hides the asymmetry with out-of-place mapping, then manages amplification through garbage collection, temperature segregation, wear leveling, and TRIM.

Stanford CS111 Lecture 27: Trap-and-Emulate, Virtual I/O, and Nested Page Tables

A VM expands the process interface into a machine interface; the hypervisor directly executes ordinary instructions, traps privileged operations, and virtualizes interrupts, I/O, and two-stage address translation.

Stanford CS111 Lecture 28: Four Ideas Connecting Concurrency, Memory, and Storage

Lecture 28 reduces the semester to concurrency, memory, and storage, then uses four ideas—virtualization, atomicity, locality, and layering—to explain how operating systems manage shared resources.

Ably: Global Realtime Messaging with Channels, Presence, and Recovery

Ably manages global realtime connections, channels, presence, and short-window recovery; applications still own idempotency, durable business state, token capabilities, and offline resynchronization.

Delegated Authorization for AI Agents: Do Not Hand User Tokens to the Model

An agent should execute one task with a short-lived, audience- and permission-restricted credential while preserving user and agent identities, execution-time authorization, confirmation, and audit lineage.

AI Agent Sandbox Escapes and Permission Boundaries: A Container Is Not the Whole Boundary

Agent execution must constrain kernels, filesystems, processes, networks, credentials, and tool authorization; sandbox escape is only one path, and an overpowered API token is often more direct.

Linode and Akamai Cloud: Connecting Classic VPS Compute to Edge and Managed Kubernetes

Linode is now the compute foundation of Akamai Cloud Computing; VMs, LKE, storage, and databases retain regional and network boundaries, so the Akamai brand does not imply complete integration.

tech deep-dive

Algolia Site Search Deep Dive: Hosted Indexing, Ranking, and InstantSearch

Algolia packages indexing, search-as-you-type, facets, and UI components as a hosted service; it ships quickly, but data synchronization, relevance, and usage costs remain your responsibility.

Apache Kafka: A Replayable Event Log, Not Merely a Message Queue

Kafka is a distributed log ordered by partition and retained by policy. Consumer groups divide work through offsets, while exactly-once processing holds only inside boundaries covered by Kafka transactions.

Apache Pulsar: An Event Platform Separating Stateless Brokers from BookKeeper Storage

Pulsar separates serving from storage: stateless brokers handle connections while BookKeeper stores ledgers and subscription cursors. Elasticity and multi-tenancy come with more operational components.

Appwrite: A Self-Hostable BaaS for Auth, Databases, Storage, Functions, and Realtime

Appwrite combines Auth, TablesDB, Storage, Functions, Realtime, and Messaging behind consistent APIs; Cloud and self-hosted products resemble each other but have different operational ownership.

tech deep-dive

assistant-ui Explained: Runtime and Primitives for Backend-Portable Agent Chat

assistant-ui separates Agent Chat into headless React primitives, a conversation runtime, and backend adapters, so the UI does not have to bind directly to one model SDK's message state.

AWS App Runner: The Shortest AWS Path from Source or Container to a Web Service

App Runner packages build, deployment, TLS, load balancing, and autoscaling as a web service, trading orchestration control for a simpler platform with explicit VPC, instance, and health boundaries.

AWS Fargate: No EC2 Management Does Not Mean No ECS Management

Fargate removes container-host operations, but task definitions, ECS services, VPCs, IAM, scaling, deployment, and observability remain your system.

AWS Lambda: Understand Events, Retries, and Concurrency Before Choosing Functions

Lambda fits short-lived, event-driven, bursty work; its design center is invocation, retries, idempotency, and downstream capacity—not merely smaller containers.

AWS SQS and SNS: One Stores Work, the Other Fans Out Events

SQS is a pull-based durable queue and SNS is a push-based topic. Reliable fan-out commonly connects one SNS topic to multiple SQS queues rather than sharing one queue across services.

Azure App Service: The App Service Plan Matters More Than the Container

App Service manages web runtimes, TLS, deployment, and scaling, while capacity, cost, and isolation live in the shared App Service Plan rather than one app.

Azure Container Apps: Serverless Containers with Revisions, KEDA, and Environments

Azure Container Apps hides Kubernetes while exposing HTTP/TCP ingress, revisions, KEDA scaling, jobs, and Dapr; replica concurrency, event idempotency, VNet design, and identity remain yours.

Better Auth: A TypeScript Authentication Framework, Not Application Authorization

Better Auth unifies login, sessions, providers, and plugins; applications still own resource authorization, revocation latency, and policy for agent actions.

tech deep-dive

Better Auth: Should Authentication Live Inside Your TypeScript App?

Better Auth trades a managed identity platform for an in-app library and your own database; that gives you control, but migrations, security updates, and incident response become your responsibility.

tech deep-dive

Bright Data Deep Dive: From Proxies and Web Unlocker to Browser API and Datasets

Bright Data splits web data access into four layers: proxies preserve control, Web Unlocker returns unblocked content, Browser API hosts interactive browsers, and Web Scraper APIs or Datasets deliver structured data.

CapRover: Self-Hosted PaaS with Docker Swarm, Nginx, and Persistent Apps

CapRover wraps Docker Swarm, Nginx, and captain-definition in a simpler PaaS; stateless apps scale, while local persistent apps remain pinned to one node.

tech deep-dive

Clerk Authentication Platform: From UI Components and Session Tokens to Organization Authorization

Clerk's real value is an integrated identity lifecycle, not a sign-in box; resource authorization, tenant isolation, and business-data consistency remain your application's responsibility.

Cloudflare Durable Objects: The Stateful Coordination Layer for Workers and WebSockets

Durable Objects map a name or ID to a globally unique, single-threaded actor with private SQLite storage. They fit per-room, per-user, per-tenant, and per-run coordination boundaries; the real design question is where the object key belongs.

Cloudflare Queues: Move Work off the Request Path into Retryable Batches

Cloudflare Queues is the message queue beside Workers: producers enqueue slow work, consumers process it with batching, ack/retry, delays, and DLQs. It is good for single-step background jobs; durable multi-step state belongs in Workflows.

CodeQL: Extracting a Codebase into a Database and Querying Data Flow

CodeQL builds a code database with language extractors, then queries syntax, types, calls, control flow, and data flow; its depth depends on models and carries extraction and query-maintenance costs.

Convex: Building Reactive TypeScript Backends with Queries, Mutations, and Actions

Convex combines typed backend functions, a transactional document database, and reactive query subscriptions; correctness depends on separating deterministic mutations from side-effecting actions.

Coolify: Control Planes, Docker Servers, and the Self-Hosted PaaS Boundary

Coolify controls Docker, proxies, and resources on your servers over SSH; deployment gets easier, but OS, security, capacity, data backup, and recovery remain yours.

tech deep-dive

CopilotKit Explained: Bring Agent State, Tools, and Human Approval into React

CopilotKit is more than a chat box. Its React components, AG-UI events, shared state, and interrupt flows connect an agent's execution to an existing product interface.

CoreWeave: An AI Cloud Built from Kubernetes, GPU Fabric, and Storage

CoreWeave is more than rented GPUs: it combines Kubernetes, GPU networking, storage, and inference into AI infrastructure, while platform engineering and capacity governance remain yours.

Crusoe Cloud: From GPU VMs and Managed Kubernetes to Managed AI

Crusoe offers both Infrastructure Cloud and Managed AI: GPU VMs and clusters provide control, while serverless and dedicated inference provide higher abstractions with different responsibilities.

DeepSeek Harness (dsh): A Coding Agent Framework That Takes Everything-is-a-Plugin All the Way

DeepSeek Harness (dsh) is DeepSeek's official open-source coding agent framework, released as a v0.1 developer preview on 2026-08-13, accumulating 184,000+ stars in 9 days. Its core is the Cordis plugin kernel — model adapters, tools, agent loop, and UI are all swappable plugins. Four runtime modes, with the ability to use Claude Code and Codex as sub-agents. Web UI first, no native CLI.

Deno Deploy: Deno 2, Revisions, Timelines, and Global TypeScript Serverless

The new Deno Deploy runs application revisions on Deno 2; understand it through timelines, contexts, databases, and telemetry rather than Deploy Classic assumptions.

DigitalOcean App Platform: Managing PaaS Topology with Components and App Specs

DigitalOcean App Platform composes Services, Workers, Jobs, Static Sites, and Functions into an App, with an App Spec as the reviewable deployment contract.

DigitalOcean: A Simplified Cloud from Droplets to App Platform

DigitalOcean covers common product architectures with Droplets, DOKS, Managed Databases, and App Platform; simplicity comes from a smaller surface, not from eliminating OS, network, backup, or HA design.

Django: Python Web Applications with ORM, Admin, Auth, and Long-Term Evolution

Django combines data models, migrations, authentication, admin, forms, and security defaults into one system; using it well still requires understanding QuerySets, middleware, async boundaries, and production settings.

Dokku: Turning One Docker Host into a Git-Push PaaS

Dokku combines a Git receiver, buildpacks or Dockerfiles, process models, Nginx, and plugins for a single-host Heroku workflow; simplicity comes from narrow orchestration scope.

Dokploy: Self-Hosted PaaS with Applications, Compose, Remote Servers, and Swarm

Dokploy supports single-container Applications and Compose or Stack, while treating one host, independent remote servers, and a Swarm cluster as distinct topologies.

DuckDB: An OLAP Query Engine Inside Your Process

DuckDB is an in-process columnar OLAP database for Parquet, CSV, and DataFrames, not a conventional multi-user OLTP server for web applications.

tech deep-dive

Elasticsearch and OpenSearch: Choosing a Lucene-Based Site Search Engine

Elasticsearch and OpenSearch both build text analysis, BM25, aggregations, and vector search on Lucene, but licensing, governance, hybrid-search APIs, and managed ecosystems have followed separate paths since the 2021 fork.

Elysia: Bun-First APIs with Runtime Schemas and End-to-End Types

Elysia connects runtime schemas, TypeScript inference, OpenAPI, and Eden clients into one contract pipeline, but Bun-first performance, plugin scope, and cross-runtime compatibility still require separate verification.

Fastify: Plugin Encapsulation, JSON Schema, and Efficient Node.js APIs

Fastify is more than benchmarks: plugin scopes, hooks, decorators, and compiled JSON Schema build composable Node.js APIs with explicit request and response contracts.

Firebase: The BaaS Boundary of Auth, Firestore, Functions, and Security Rules

Firebase moves quickly because client SDKs directly access managed Auth, Firestore, and Storage; the real backend contract lives in data models, Security Rules, Functions, and cost limits.

Fly.io: The Real Boundaries of Machines, Fly Proxy, and Multi-Region Deployment

Fly.io places fast-starting Machines in chosen regions and connects them through Fly Proxy and private 6PN; cross-region state consistency remains your hard problem.

gitleaks: Secret Detection across Working Trees, Git History, and CI Diffs

gitleaks scans files or Git patches with rules, regexes, entropy, and allowlists; after a finding, revoke and rotate first rather than merely deleting a file or rewriting history.

Google Cloud Model Armor: Runtime Filters around Prompts, Responses, and Agent Tools

Model Armor can inspect prompt injection, jailbreaks, sensitive data, malicious URLs, and unsafe content at runtime; it is a probabilistic detector, not an authorization or sandbox boundary.

Google Cloud Run: Services, Jobs, and Worker Pools Have Different Container Lifecycles

Cloud Run is more than an HTTP container platform: choose request-serving, run-to-completion, or always-on pull work first, then design concurrency, identity, and scaling.

Google Kubernetes Engine: GKE Autopilot Reduces Node Operations, Not Kubernetes

GKE manages the Kubernetes control plane and Autopilot manages most node infrastructure, while workloads, policy, networking, upgrade compatibility, and cost governance remain yours.

GraphQL and Code Generator: Type Safety Comes from Schema Plus Operations

A GraphQL schema defines available capabilities; GraphQL Code Generator combines it with actual queries, mutations, and fragments to emit precise results, variables, and typed documents.

gRPC and Connect: One Protobuf Contract across Services and Browsers

gRPC generates cross-language stubs from Protobuf services. A Connect server can support gRPC, gRPC-Web, and the Connect protocol together, avoiding a translating proxy for browsers.

Hatchet: One Engine for Task Queues, DAGs, and Durable Tasks

Hatchet unifies regular tasks, DAGs, and durable tasks behind a Postgres-backed control plane. Durable tasks checkpoint at waits and child tasks, then replay deterministic orchestration code on recovery.

Hetzner Cloud: Cheap VMs Mean Owning the Entire Operating System

Hetzner Cloud offers lean IaaS through servers, networks, volumes, load balancers, and firewalls; its price advantage is real only after patches, HA, backups, egress, and on-call are counted.

Hono RPC: Infer Fetch Client Types from Route Implementations

Hono RPC exports a route's `typeof AppType` so `hc` can infer inputs, response bodies, and status codes. The shared artifact is a TypeScript type, not an independent wire schema.

Inngest: Turn Serverless Functions into Recoverable Workflows with Steps

Inngest makes steps the persistence boundary for ordinary TypeScript, Python, and Go functions. Recovery re-executes the function while memoized steps avoid repeating completed side effects.

Kamal: Docker Deployment Without a Resident Control Plane

Kamal deploys immutable images from an operator over SSH and switches traffic through kamal-proxy; it is not a scheduler and does not operate hosts or data.

Koyeb: A Global Serverless Model of Apps, Services, and Instances

Koyeb groups Services in Apps, runs revisions as Instances in selected regions, and integrates global routing, autoscaling, private discovery, and CPU or GPU compute.

Kubernetes: Container Orchestration from Pods and Controllers to Declarative Reconciliation

Kubernetes is fundamentally API objects, controllers, and reconciliation loops—not YAML; it manages workload lifecycles without solving application state, data consistency, or organizational governance.

Kysely: A Type-Safe TypeScript Query Builder That Keeps SQL Visible

Kysely derives query results from database types while preserving SQL and escape hatches, but teams must still keep migrations, the live schema, and generated types aligned.

Lambda Cloud: GPU Compute from One VM to Multi-Node AI Clusters

Lambda Cloud provides on-demand GPU VMs and 1-Click Clusters; it offers direct AI compute environments rather than automatically solving training, serving, and MLOps.

Liveblocks: Collaboration Backends with Presence, Storage, Comments, and Notifications

Liveblocks packages rooms, presence, conflict-free storage, comments, and notifications as managed product primitives; adoption still requires alignment with existing auth, canonical databases, and data lifecycle.

MongoDB: Document Flexibility Moves the Cost to Data Boundaries

MongoDB fits aggregate-oriented data and evolving documents; the hard decisions are embedding, transaction boundaries, indexes, and shard keys.

MySQL: A Mature Relational Database, Not Merely the Popular Default

MySQL offers a mature ecosystem, InnoDB transactions, and predictable operations, but teams must still own indexes, isolation, replication lag, and migrations.

NATS and JetStream: Choose Ephemeral Messaging or Durable Events First

Core NATS is storage-free, at-most-once pub/sub. JetStream adds streams, consumers, acknowledgments, retention, and replay. They share subjects but expose fundamentally different reliability contracts.

Nebius AI Cloud: A Full Platform for GPU Clusters, Managed Kubernetes, and Serverless AI

Nebius combines GPU VMs and clusters, Kubernetes, Slurm, storage, and Serverless AI; choose the responsibility layer before comparing hardware and price.

Neon vs Turso: Do Not Call Both Serverless Databases Managed Postgres

Neon is serverless PostgreSQL; Turso Cloud currently follows the libSQL and SQLite-compatible path. Their compatibility boundaries are fundamentally different.

NestJS: A Node.js Architecture Framework of Modules, Dependency Injection, and Request Lifecycles

NestJS is valuable not for decorators alone, but for Modules, Providers, DI, and a predictable pipeline across HTTP, GraphQL, WebSocket, and microservice architectures.

Netlify: A Web Platform for Atomic Deploys, Functions, and Edge Functions

Netlify centers on atomic deploys and previews, then adds Functions, Edge Functions, Blobs, and Database; each runtime and state layer has distinct consistency and limits.

ngrok: From Localhost Tunnels to Controlled Global Ingress

ngrok is an agent-initiated reverse proxy and ingress, not a VPN for the whole machine; public endpoints still need explicit authentication, traffic policy, and data boundaries.

Nhost: A BaaS of PostgreSQL, Hasura GraphQL, Auth, and Storage

Nhost uses PostgreSQL as the source of truth, Hasura to generate GraphQL, and connects Auth claims, role permissions, Storage, and Functions into one platform.

OMP 2 (Oh My Pi 2): From Pi Fork to Full Rust Rewrite as an Independent Coding Harness

OMP 2 is no longer a Pi fork. The entire codebase has been rewritten from scratch in Rust, with ~41 crates covering a custom bash engine, GPU-accelerated GUI, embedded CPython 3.14t, gRPC transport, and Kokoro-82M TTS. Currently in pre-release with no stable version yet.

openapi-typescript: Turn OpenAPI into Runtime-Free Types and a Fetch Client

openapi-typescript converts OpenAPI 3.0 and 3.1 into pure TypeScript types; openapi-fetch then infers methods, literal paths, parameters, and response unions from that schema.

Opencode 2: The Cost of Swapping Bun for Node, Tauri for Electron, and Rebuilding the Entire API

Opencode 2 is a major rewrite led by Anomaly (Dax Raad). Runtime migrated from Bun to Node.js (memory issues), desktop from Tauri to Electron (WebKit perf and Node integration), v1 API intentionally incompatible. New: multi-tab parallel sessions, persistent backend service, HTTP API + SDK. Currently beta, stable estimated ~September 2026. ~200K stars.

OpenStack: Not a Virtualization UI, but Cloud Services You Operate Continuously

OpenStack combines Keystone, Nova, Neutron, Glance, Placement, Cinder, and other services into multi-tenant IaaS; adopting it means staffing a cloud platform team, not finishing an installation.

Oracle Cloud Infrastructure: Start with Compartments, VCNs, and Fault Domains

OCI is a full hyperscale cloud; architecture starts with tenancy and compartment IAM, region/AD/fault domains, and VCNs before selecting Compute, OKE, databases, and storage.

oRPC: Put End-to-End Type Safety and OpenAPI on the Same Path

oRPC supports implementation-first and contract-first APIs, offers an RPC client, and can expose the same router through OpenAPI 3.1.1 HTTP endpoints.

OVHcloud: Combining Public Cloud, OpenStack, vRack, and Dedicated Servers

OVHcloud combines Public Cloud, OpenStack APIs, Managed Kubernetes, vRack, and dedicated or private cloud; that flexibility also creates more networking and responsibility boundaries.

Oxc and Oxlint: Speed Comes From Reconnecting Parsing, Types, and Rules

Oxlint has grown from a fast ESLint companion into a standalone linter with type-aware rules and JavaScript plugins. Migration depends on rule, framework-file, and plugin compatibility—not only a 50–100x benchmark.

tech deep-dive

Pagefind Explained: Full-Text Search for Astro Without a Search Backend

Pagefind scans static HTML after an Astro build and ships its index with a WebAssembly search runtime; the browser fetches only the index chunks required by a query, so no search server is needed.

PartyKit: Turning Collaboration Rooms into Stateful Edge Servers

PartyKit concentrates WebSocket coordination in room-keyed stateful servers for multiplayer and presence, while durable documents, authorization, hibernation, and platform ownership still need explicit design.

Pi v2: AgentHarness API Goes Stable, Earendil Incorporates — Minimalism Enters Its Next Chapter

Pi v0.84.0 (2026-08-06) promotes the AgentHarness v2 API to stable. Lane-based v4 Session model makes operations durable and interruptible. CBOR replaces JSON, Unix sockets replace HTTP. Earendil Inc. (Armin Ronacher's PBC) behind it has secured initial funding. 95.4K stars, still MIT, still minimal.

PocketBase: A Single-Binary Backend with SQLite, Auth, and Realtime

PocketBase packages SQLite, collections, Auth, file storage, SSE realtime, and an admin UI into a small executable; deployment is easy, but single-host and pre-v1 compatibility limits matter.

Promptfoo Red Team: Turning Prompt Injection, Tool Misuse, and Data Leaks into Regression Tests

Promptfoo plugins generate risk probes, strategies transform attacks, targets execute the system, and graders judge outcomes; useful red teams exercise the full agent application rather than only a foundation model.

Protobuf and Buf: Put Cross-Language Schema Linting, Codegen, and Compatibility in CI

Protobuf defines binary messages with stable field numbers; Buf adds modules, linting, remote plugins, generation, and breaking-change checks to govern those schemas.

Proxmox VE: Combining KVM, LXC, Clusters, Ceph, and Backups On-Premises

Proxmox VE integrates VMs, containers, clusters, HA, storage, and backup; it simplifies virtualization management while hardware, quorum, networks, capacity, and DR remain yours.

Pulumi: Infrastructure as Code in TypeScript, Python, Go, and Other Languages

Pulumi lets general-purpose programs register cloud resources for a deployment engine, providers, and stack state to preview and update; greater language power demands stronger abstraction discipline.

RabbitMQ: Express Routing with Exchanges, Preserve Work with Quorum Queues

RabbitMQ's strength is routing through exchanges, bindings, and queues. For replicated work, default to quorum queues and combine publisher confirms, manual acknowledgments, and idempotent consumers.

Railway: Deploying an Application Topology with Projects, Services, and Environments

Railway is not merely one-click deployment; it puts container services, environments, variables, and private networking into one operable application project.

Self-Hosting Inference with Ray Serve: Python Service Graphs, GPU Scheduling, and Autoscaling

Ray Serve is a distributed serving layer on Ray. Deployments and handles compose Python service graphs, while replicas, CPU/GPU scheduling, autoscaling, and model multiplexing handle orchestration; it complements rather than replaces vLLM or SGLang.

Redis Streams: An Append-Only Log and Consumer Groups for Reliable Redis Messaging

Redis Streams stores replayable entries with `XADD`; consumer groups add a Pending Entries List and `XACK`. It is more durable than Pub/Sub, but it does not automatically become Kafka.

Redpanda: Kafka API Compatibility Still Requires Broker Migration Testing

Redpanda reimplements a Kafka-compatible event log with C++/Seastar, thread-per-core execution, and a Raft group per partition. Client compatibility is broad, but operational and edge semantics still require testing.

Render: A Full PaaS Topology of Web, Private, Worker, and Cron Services

Render's value exceeds turning a repository into a URL: distinct service types model public HTTP, private listeners, queue workers, cron, data stores, and Blueprint infrastructure.

Renovate: Turning Dependency Updates into a Governed Continuous Process

Renovate is more than an update-PR bot: packageRules, grouping, schedules, minimum release age, and automerge policy determine update speed, review noise, and supply-chain exposure.

Replicate: Turn Model Versions into Prediction APIs Instead of Renting GPUs

Replicate abstracts GPUs behind versioned models, predictions, Cog, and deployments; integrators still own version pinning, async workflows, webhook verification, data persistence, and spending limits.

Restate: Put Journals, Durable State, and Service Calls in One Execution Model

Restate journals operations and results, then re-executes handlers while skipping completed work. Virtual Objects and Workflows add keyed state, single-writer semantics, and long-lived coordination.

RunPod: GPU Pods and Serverless Endpoints Are Different Products

RunPod Pods fit interactive and persistent GPU work, while Serverless fits queued or load-balanced inference; choosing incorrectly mixes persistence, cold starts, and retry semantics.

Scaleway: European Cloud Instances, Kapsule, Serverless, and Managed Data

Scaleway now spans compute, Kapsule, serverless, databases, storage, AI, and IAM rather than only low-cost VMs; maturity and integration still require per-region verification.

tech deep-dive

Scrapy Deep Dive: A Self-Hosted Crawler from Engine to Pipeline

Scrapy separates crawling into the Engine, Scheduler, Downloader, Spider, Item Pipeline, and middleware; it fits high-volume, rule-driven HTTP crawling where you need control over scheduling, throttling, retries, and storage.

tech deep-dive

Selenium Deep Dive: Browser Automation from WebDriver Sessions to Grid

Selenium drives real browsers through standardized WebDriver sessions, making it useful for cross-browser workflows, existing test assets, and remote Grid capacity; it can render JavaScript applications, but it does not guarantee bypassing CAPTCHAs or other anti-automation controls.

Semgrep: Encoding Security Policy as Readable, Tested Static Analysis

Semgrep lets teams express SAST policy with source-like patterns and taint rules; rule quality depends on positive and negative tests, framework modeling, and exception lifecycle.

Server-Sent Events: Recoverable One-Way Push over HTTP Event Streams

SSE sends text/event-stream over ordinary HTTP, with browser reconnection and Last-Event-ID; simple transport is not durable unless the server retains events behind that cursor.

Self-Hosting Inference with SGLang: RadixAttention, OpenAI APIs, and Multi-GPU Serving

SGLang is an inference engine for generative models. RadixAttention reuses KV cache across shared prefixes, while OpenAI-compatible APIs, structured output, and multi-GPU parallelism support production LLM serving; it is not a complete product backend.

Sigstore and SLSA: Verifying Who Built an Artifact with Which Process

Sigstore provides identity-bound signing, short-lived certificates, and transparency logs; SLSA describes trustworthy build provenance. They protect only when admission verifies identity, issuer, digest, and build expectations.

Snyk: Connecting SCA, SAST, Container, and IaC Findings to Development

Snyk maps code, open-source, container, and IaC findings to projects, remediation paths, and developer workflows; successful adoption depends on baselines, ownership, and executable policy.

Socket.dev: Blocking Malicious Package Behavior at Dependency-Diff Time

Socket.dev goes beyond CVEs by analyzing install scripts, obfuscation, network and shell access, and ownership changes when packages enter a dependency diff.

Socket.IO: Realtime Events with Rooms, Acknowledgements, and Recovery

Socket.IO is an event protocol and runtime over WebSocket or long-polling, not a native WebSocket compatibility layer; ordering, arrival, recovery, and horizontal scaling need separate designs.

Speakeasy: Manage Multi-Language SDKs with OpenAPI Overlays and Workflows

Speakeasy records OpenAPI, Overlays, targets, and generator versions in `.speakeasy/workflow.yaml`, enabling local or CI generation, compilation, and publishing of multi-language SDKs.

SST v3: Composing Full-Stack Cloud Apps with Components, Link, and Dev Mode

SST v3 composes cloud resources through TypeScript, high-level app components, Pulumi and Terraform providers, and resource linking; it is application-first IaC, not a hosted PaaS.

Stainless: Continuously Generate Publishable Multi-Language SDKs from OpenAPI

Stainless uses OpenAPI plus its configuration to generate multi-language SDKs, docs, CLIs, and MCP servers. Its value is a continuous preview, publishing, and upgrade pipeline rather than one-off codegen.

tech deep-dive

Stytch Deep Dive: From B2C Login and Sessions to B2B Organizations and Authorization

Stytch is an API-first managed identity platform: choose the Consumer or B2B model, converge authentication factors into sessions, then enforce organization and RBAC boundaries on the server.

Teleport: Infrastructure Access with Short-Lived Credentials, RBAC, and Session Audit

Teleport is a protocol-aware infrastructure access platform: its Auth Service signs short-lived credentials, while Proxies and Agents mediate SSH, Kubernetes, database, and app access with audit evidence.

Terraform: The IaC Workflow of Providers, Plans, Applies, and State

Terraform compares provider schemas, configuration, state, and real APIs to create a plan and apply it; controlled state and change workflow matter more than HCL itself.

NVIDIA Triton Inference Server: Multi-Framework Models, Dynamic Batching, and Pipelines

Triton Inference Server serves TensorRT, ONNX, PyTorch, and other models through consistent HTTP and gRPC APIs. Its defining tools are the model repository, dynamic batching, instance groups, and ensembles—not LLM-specific KV-cache scheduling.

tRPC: Connect Client and Server Contracts through TypeScript Inference

tRPC lets a client reference the server router type without a separate schema or code generation. Version 11 also has an official, alpha-stage OpenAPI 3.1 generator.

ts-rest: A TypeScript Contract-First API that Preserves REST Semantics

ts-rest describes methods, paths, status codes, and schemas in a shared contract, providing end-to-end server and client types without a code-generation step.

Twingate: Identity-to-Resource Access Instead of Whole Networks

Twingate uses clients, connectors, a controller, and relays to narrow user authorization to specific resources; it is managed ZTNA rather than a general peer-to-peer overlay.

TypeScript 7: A Native Go Rewrite Makes Type Checking Fast, but Migration Goes Beyond tsc

TypeScript 7 rewrites the compiler and language service in Go. The order-of-magnitude gains come from native execution and shared-memory parallelism, while old APIs and compiler options define the migration cost.

tech deep-dive

Typesense Site Search: Full-Text Search You Can Treat as a Product Feature

Typesense is a search server centered on instant, typo-tolerant keyword search. Collection schemas, field weights, facets and sorting are enough to build site search, with a choice between self-hosting and Typesense Cloud; Chinese fields need locale: zh, and domain terminology may require custom segmentation.

Vercel: Frontend Cloud Gains Its Advantage from Framework-Aware Deployments

Vercel binds framework build output, previews, CDN caching, and Functions into deployments; deep integration adds speed while defining runtime, locality, cost, and portability boundaries.

Vite 8 and Rolldown: One Pipeline Replaces the Two-Bundler Split

Vite 8 replaces the esbuild-for-development and Rollup-for-production split with Rolldown. Speed is the visible result; consistent bundler semantics across development and production are the deeper change.

Vitest: Share Vite Configuration Without Mistaking Component Tests for E2E

Vitest's advantage is not merely a familiar Jest-style API. Tests share Vite transforms, aliases, and plugins with the application; Browser Mode adds real-browser confidence but does not replace full E2E testing.

Vultr: Choosing Between Cloud Compute, VKE, and GPUs

Vultr spans VMs, bare metal, GPUs, VKE, databases, and storage; breadth and regional choice help, but product availability alone does not create an integrated architecture.

WebSocket: Duplex Connections Are the Start; Protocol and Backpressure Are the Work

WebSocket supplies a duplex message transport, not auth renewal, schemas, acknowledgements, replay, rooms, or backpressure policy; those layers determine production correctness.

WireGuard: Minimal VPN Tunnels with Cryptokey Routing

WireGuard is a small, explicit layer-3 encrypted tunnel that binds public keys, peers, and AllowedIPs, but it does not supply identity, device management, or a policy control plane.

tech deep-dive

WorkOS Enterprise Auth: Growing from AuthKit to SSO, SCIM, and Audit Logs

WorkOS lets a B2B SaaS add SAML/OIDC, SCIM, and audit exports around one organization identity model; application authorization and data governance remain your responsibility.

Yjs and CRDTs: Concurrent Editing as Commutative, Replayable Document Updates

Yjs uses shared types and commutative, associative, idempotent binary updates to converge concurrent edits; it does not prescribe transport or include authorization, persistence, or domain conflict resolution.

ZeroTier: Virtual Networks with Controllers and Flow Rules

ZeroTier places devices on a managed virtual L2/L3 network, attempts peer-to-peer transport, and uses a controller to publish membership and policy; it resembles software-defined networking more than a single tunnel.

zizmor: Finding Template Injection and Token Risks in GitHub Actions

zizmor performs domain-specific static analysis on workflow and action YAML for template injection, broad permissions, artifact credential leaks, and unpinned uses; it does not analyze called shell scripts.

Zodios: Build an Axios Type-Safe Client from Zod Endpoint Definitions

Zodios uses a central Zod endpoint definition for Axios client types, runtime validation, and aliases, with optional Express and OpenAPI packages.

tech deep-dive

Zyte Deep Dive: From Scrapy Development and Anti-Bot Fetching to Scrapy Cloud

Scrapy owns crawl flow and data models, Zyte API handles fetching, browsers, and anti-bot infrastructure, and Scrapy Cloud adds deployment, scheduling, and output; each layer can be adopted independently.

Agent Plugins 1.0: OpenAI, Google, and AWS Unite to Standardize AI Agent Extensions

Agent Plugins 1.0 is a packaging format that bundles Agent Skills (markdown instructions) and MCP server configs into a single directory, loadable by ChatGPT, Cursor, GitHub Copilot, Kiro, and VS Code. It's not a new protocol — it's the wrapper above protocols. Vercel initiated it, OpenAI/AWS/Microsoft/Cursor co-authored it, and Google joined on launch day. Anthropic isn't on the governance board, but MCP is a core primitive of the spec.

ai guide

AgentQL Complete Guide: Semantic Web Extraction and Playwright Automation

AgentQL replaces brittle CSS and XPath selectors with queries shaped like the data you want: `query_data` returns structured values, while `query_elements` returns interactive Playwright locators. The public Starter plan lists 50 free API calls per month, but its payment, hard-stop, and remote-browser reset rules still need to be verified in Billing.

ai guide

Apify Complete Guide: How Actors, Tasks, Schedules, and Datasets Form a Scraping Platform

Apify is not a single crawler. It packages scraping programs as Actors, saves reusable configurations as Tasks, triggers them with Schedules, and delivers results through Datasets. It fits teams that do not want to operate queues, schedulers, and workers, but Actor fees, compute, proxies, storage, and transfer all draw from the same platform budget.

ai guide

Browser Use Complete Guide: The Agent Loop Behind Browser Automation

Browser Use combines browser state, model decisions, and actions such as click, type, and extract into a repeatable loop. The open-source package favors custom tools and execution control; Cloud manages browsers, profiles, proxies, and concurrent work.

ai guide

changedetection.io Complete Guide: Selectors, Notifications, and Browser Steps

changedetection.io is a web-change signal layer: narrow the monitored content, suppress noise, and notify downstream systems only when a meaningful change occurs. It is neither a search API nor a crawler replacement.

Composio: Who Holds Every User's Token When Your Agent Connects a Hundred SaaS Apps

This site covers MCP thoroughly but has never written about the layer underneath it: when your agent acts for ten thousand end users reading their own Gmail, whose database holds those refresh tokens, who rotates them, who revokes them. Composio is currently the most complete answer — MIT-licensed SDKs, a commercial hosted execution and OAuth layer. It claims 1,000+ toolkits; the managed-auth page actually lists 121 with a Composio OAuth app and 96 that require your own credentials. New pricing effective 2026-08-15: 100K free tool calls, $29/mo Pro. This post takes the authorization model down to an operational level and draws the line between wiring up MCP servers yourself and buying an integration platform.

ai deep-dive

Seven Answers to a Full Context Window, and No Consensus

Chroma's controlled study shows that even when it fits, a full context degrades performance. Coding agent vendors have landed on seven different responses: compact, hand off, prune, defer loading, isolate, train it into the model, or change the unit of work. Amp removed /compact outright, Atlassian argues summarization should be a last resort, and Cursor's A/B test measured a 46.9% token reduction. The three real disagreements come down to what each team is measuring.

ai guide

Crawl4AI Complete Guide: From Markdown Crawling to Structured Extraction

Crawl4AI handles retrieval after a URL is known: use JsonCssExtractionStrategy for stable DOMs, and switch to LLMExtractionStrategy only when extraction needs semantic judgment or must tolerate irregular layouts.

CrewAI: Organizing Multi-Agent Collaboration Through Role-Playing

CrewAI (GitHub 57.4k stars, MIT, PyPI 11.6M weekly downloads) defines agents by role, goal, and backstory, then groups them into crews for collaboration. Unlike LangGraph's graph-first and MAF's workflow-first approach, CrewAI is team-first — you don't draw nodes and edges, you describe who's on the team and what each person does. It fully removed its LangChain dependency in late 2024 and is now a standalone framework. The commercial side splits into the open-source package and AMP, a managed platform adding visual building, deployment, tracing, and compliance.

Exa: Neural Search Built for Agents, Not People

Exa turns every indexed web page into an embedding and retrieves by vector similarity instead of keyword matching. Official pricing as checked on 2026-08-21: $7 / 1k requests for /search (first 10 results included), $1 / 1k pages for /contents, $12–15 / 1k for the deep tiers, with $20 in free credits for new accounts. This blog's CLAUDE.md puts Exa first among cloud fetch tools, 16 of its 38 skills reference it directly, and only four existing posts mention it in passing — with zero dedicated posts. This is that post.

ai guide

Firecrawl Complete Guide: Choosing Scrape, Crawl, Map, and Structured Extraction

Firecrawl puts single-page scraping, site discovery, whole-site crawling, and JSON extraction behind one API. Cloud removes browser, proxy, and worker operations; self-hosting gives infrastructure control, but not the complete Cloud feature set.

ai deep-dive

Choosing Free Search, Scraping, and Browser APIs: Recurring Quotas, Trials, and Self-Hosting

Free access is not one model: recurring allowances, balance top-ups, rate-limited access, one-time credits, and self-hosting have different steady-state costs.

Linkup Search API Guide: From standard and deep to Structured Output

Linkup separates search depth from response shape: start most agent queries with standard + searchResults, move to deep only for multi-step browsing, and treat the monthly $20 as a balance refill rather than a new $20 grant.

LlamaIndex Is Not a RAG Framework Anymore, and Old Tutorials Won't Tell You

LlamaIndex (51,775 GitHub stars, MIT, verified 2026-08-21) has moved its center of gravity from indexing to Workflows: the standalone llama-index-workflows package pulls 2.81M weekly PyPI downloads, more than the 1.97M of the llama-index umbrella package itself. This post covers the core abstractions, the trade-off against hand-rolling a pipeline, and a hands-on test of its defaults on Traditional Chinese text — at the same chunk_size=1024, English fits 4,645 characters and Traditional Chinese only 1,332. Plus one fact you need before choosing: the TypeScript port is archived and unmaintained.

ai deep-dive

Meilisearch Complete Guide: Indexing, Chinese Search, and Tenant Security

Meilisearch turns application data into fast, typo-tolerant full-text search; the hard parts are index settings, asynchronous tasks, Chinese tokenization, access filters, and tested recovery.

Microsoft Agent Framework: After the Merge, Who Does the Name AutoGen Point To?

Microsoft merged Semantic Kernel and its own AutoGen into Microsoft Agent Framework, which hit 1.0 GA on 2026-04-02 for .NET and Python (Go is still public preview). The absorbed autogen-agentchat has not shipped since 2025-09-30. But AG2, the fork on the original authors' side, never merged — it shipped 1.0.2 six days ago, and `pip install autogen` gets you AG2, not Microsoft. This post covers MAF's abstractions, the migration clock, and how to read the tangle of names.

ai guide Reading MIT 6.S191

MIT 6.S191 Guide: Nine Lectures and Three Labs Are Public, but the Full Path Still Uses Three External Services

MIT 6.S191's 2026 edition publishes nine lecture videos, slides, three software labs, and solutions, making it an A3 self-study course. The supplied path still depends on Google/Colab, Comet, and OpenRouter for Lab 3, while unaffiliated learners do not receive MIT credit, project feedback, or API credits.

Modal: The Layer Your Inference Engine Runs On — and When the Premium Isn't Worth It

Modal is a per-second-billed serverless GPU platform that also treats agent sandboxes as a first-class primitive (company-reported: over 1 billion sandboxes launched, more than a third of revenue). The selection question isn't how convenient it is — it's your GPU utilization. Verified 2026-08-21: Modal's A100 80GB works out to $2.50/hr against RunPod's $1.59/hr for the same card, so above 64% utilization renting your own is cheaper. But on the same day, H100 SXM is $3.95/hr on Modal against $3.99 on Lambda — on that card the premium is gone.

ai deep-dive

Qdrant Complete Guide: Collections, Hybrid Search, and Self-Hosted Operations

Qdrant is not just a place to store embeddings: define the vector schema, index frequently filtered payload fields, then add dense+sparse queries, tenant boundaries, snapshots, and monitoring to build an operable retrieval service.

ai guide

Scrapling Complete Guide: From Adaptive Selectors to Concurrent Spiders

Scrapling puts HTTP, Playwright browsers, CSS/XPath extraction, and a Spider API behind one Python interface. Adaptive selectors save element properties and relocate a target by similarity after a layout change, but the output still needs validation.

ai guide

SearXNG Complete Guide: Engine Tuning, JSON API, and Self-Hosted Operations

SearXNG is a metasearch engine, not a crawler, and it does not own a web-wide index. Based on the official 2026.8.20 documentation, this guide covers Compose installation, settings.yml, engine selection, the JSON API, and empty-result diagnosis.

Build Your Own Search Backend: SearXNG + Crawl4AI, From Zero to Claude Code

Three components, one job each: SearXNG finds, Crawl4AI reads, and a thin layer of your own code glues them together. Three defaults will stop you cold — `formats` only emits HTML, `secret_key` ships as the literal string `ultrasecretkey`, and turning on the limiter blocks your own code. Also, the official install path changed in March 2026, so most tutorials online point at a repository that is now archived.

Tavily and Exa Can't Be Self-Hosted: How to Build Your Own

Tavily and Exa are cloud-only APIs and can't be self-hosted. What you can assemble instead is SearXNG (269 upstream engines, 82 on by default) plus Crawl4AI (78.8k stars, Apache-2.0), and the ready-made Tavily-compatible wrappers are all still double-digit-star solo projects you should not depend on. But SearXNG has no index of its own, and running it from a datacenter IP gets you empty results — those two facts decide whether self-hosting is worth it.

Stanford CS124: Numbered 100, Four Prerequisites Written Into the Catalog, and Not Offered at All Next Year

CS124 is the first course in Stanford's NLP branch. Its textbook is Jurafsky's own Speech and Language Processing, free online, and all nine assignment repos are public. But a banner sits on the course homepage: it will not be taught at all in AY 2026–27. And the chapter numbers the syllabus points at no longer match the August 2026 textbook.

Stanford CS221: The AI Intro Course Whose Prerequisites Field Reads CS103, CS106B, CS109, CS161

CS221 lays AI out along one axis, and reflex models — deep learning — sit in the lowest slot, with states, variables and logic above them. When Percy Liang took over in Autumn 2025 he replaced the slides with runnable Python and wrote 'Cut constraint satisfaction problems :(' into the source of the first lecture — yet ExploreCourses and Stanford Online both still advertise constraint satisfaction as a course topic. The project has gone from 20% of the grade in 2019 to extra credit only.

Stanford CS224N: Open the 2019 Syllabus and Transformers Are Still Lecture 14

CS224N has kept every course website since 2000 online. In Winter 2019, Transformers were lecture 14, taught by a guest. In Winter 2026 they are lecture 5, and every lecture after that assumes you already know them. The machine translation assignment is gone; assignment 3 now has you code a decoder-only Transformer from scratch, with pytest suites that run on your laptop.

Stanford CS224U: The Course Site Stopped in Spring 2023, but You Can Clone the Whole Thing

CS224U's teaching material isn't a slide deck — it's an Apache-2.0 GitHub repo holding the lecture notebooks, all three assignments, and the grading document for the final project. But the on-campus course has skipped three straight academic years since Spring 2023, and ExploreCourses has it back on the books for Spring 2026-27. The official description still lists relation extraction and semantic parsing; the 2023 syllabus covers neither. And the data-loading cell in the first assignment breaks in a fresh environment today, on a Hugging Face compatibility change.

Stanford CS224V: Renamed to Agentic AI in 2026, but What It Teaches Is Formal Methods Against Hallucination

CS224V only became Agentic AI in the 2026–2027 catalog, and the rename changed nothing underneath: the course still translates natural language into formal semantics and constrains agents with SMT solvers and knowledge graphs instead of wiring frameworks together. Seven of the eleven mandatory readings come out of the instructor's own lab. Every slide deck is public, and the course site says outright that they are deliberately incomplete.

Stanford CS224W: Every Assignment Runs in Colab, but the Biggest Slice of the Grade Is Closed to Self-Learners

All six CS224W Colabs download and run today, and the first one needs only NetworkX — no PyG install at all. But the exam is 35% of the grade, the largest single piece, and it's an in-person closed-book sitting. The public recordings stop at 2021 and cover none of the current syllabus's second half: graph transformers, relational deep learning, LLM+GNN.

Stanford CS228: The Prerequisites Are One Sentence About Probability and Algorithms — But the Course Hasn't Run in Two Years

CS228's official prerequisite is a single line — 'basic probability theory and algorithm design and analysis' — with no named course. But ExploreCourses shows it was last offered in Winter 2024, and the next slot, Winter 2027, still has a blank instructor field. What a self-learner can actually get is cs228-notes: 16 chapters, complete, last touched in June 2025.

Stanford CS229: Notes Rewritten Every Year, Public Problem Sets Frozen at 2020, and an Official Self-Test From 2008

The three things you need to self-study CS229 run on three different clocks. The lecture notes are 278 pages and were recompiled in August 2026. The newest problem sets you can download are from summer 2020. The self-assessment Stanford Online tells you to attempt before enrolling is a PDF created in 2008. Seventeen lectures from spring 2026 are public, and the last three are mislabeled.

ai deep-dive

Stanford CS25 V6: A Course Called Transformers United Whose First Two Talks Weren't About Transformers

CS25 is Stanford's 1-unit seminar where attendance is the only homework and anyone can audit. Of the nine talks in the Spring 2026 season, the three worth your time are Albert Gu on the inductive biases of SSMs vs Transformers, Charles Frye on serving inference across thousands of GPUs, and Victoria Lin on what native multimodality still hasn't solved.

Stanford CS329Z: Hand-Build the Agent with litellm First, Then Let DSPy Take It Away

CS329Z is a new three-unit agent engineering course debuting at Stanford in Autumn 2026. Its first homework asks you to build RAG, tool calling and a ReAct loop from scratch with litellm, then rewrite the same components in DSPy and hand in the comparison. The course site lives in a public GitHub repo, and the commit log shows the assignment count dropping from three to two in mid-August — the one that got cut was 'Data for Agents'.

Stanford CS336: The Lectures Are Runnable Python, and From Assignment 2 On You Pay for the GPUs

Of the seventeen regular CS336 lectures, only nine are executable Python programs; the other eight are PDF slide decks — and the split falls exactly along the two instructors. Assignment 1's handout carries eight 'Low-Resource Tips' for finishing it on a laptop. Assignments 2 through 5 carry none. The course page lists the hourly price of a B200; the handouts list how many B200 hours each problem needs.

ai guide

Tavily Search API Complete Guide: Search, Extract, Map, and Crawl

Tavily exposes Search, Extract, Map, and Crawl through one web API for agents. The free plan includes 1,000 credits per month; basic, fast, and ultra-fast Search cost 1 credit each, while advanced costs 2.

vLLM: The Default Choice for Self-Hosted Inference — and When It's Over-Engineering

vLLM is the de facto standard for self-hosted LLM inference (89,470 GitHub stars, verified 2026-08-21), built on managing the KV cache the way an OS manages paged memory. But the selection question isn't how fast it is — it's your GPU utilization. Using Red Hat's measured 793 output tokens/second, a fully saturated A100 costs roughly $0.70 per million output tokens; at 10% utilization that becomes $7, more than most cloud APIs.

How to Evaluate Agent Search Quality: Building a Web Retrieval Benchmark

A web retrieval benchmark must evaluate complete tasks, not HTTP 200s: 30 fixed cases across five failure strata and three live channels, measuring answers, citations, freshness, latency, cost, and unnecessary escalation. This article delivers the harness and gates, but no fabricated ranking while the three live channels remain unconfigured.

A Complete Web Retrieval Route for AI Agents: When to Use Search, Fetch, Crawlers, and Browsers

An agent should not open a browser for every web task: route first to Search or Fetch, then escalate on explicit signals such as status codes, weak content, JavaScript shells, authentication, or challenge pages, with retry, budget, cache, deduplication, and provenance constraints at every step.

Berkeley AI/ML Course Guide: From CS61A to CS288, What Can You Actually Study Online?

Berkeley has no standalone undergraduate AI degree. A workable path builds on the CS BA or EECS BS foundation, enters through either CS188's broad AI curriculum or CS189's mathematical machine learning curriculum, then branches into deep learning, NLP, vision, or reinforcement learning. Many 2025–2026 courses are A3, but the newest class, the newest stable URL, and the best self-study edition are not always the same.

learning deep-dive

CMU's AI Degrees: The First U.S. AI Bachelor's Turned 'What Should AI Students Learn?' into Graduation Requirements

Stanford has no AI degree; AI is a track inside CS. CMU launched the first U.S. B.S. in Artificial Intelligence in 2018, divided AI into four clusters, required one course from each, and made ethics a graduation requirement. At the master's level, MSAII sits not in CS but in the Language Technologies Institute; 84 of its 195 units cover an innovation process ending in a fundable capstone. Two official-page conflicts emerged during verification: whether the AI Core has two or three courses, and whether MSAII totals 192 or 195 units.

CMU AI/ML Course Guide: The New 07-280 Core and a Public Self-Study Route

CMU's current BSAI now runs through 07-280 and 07-380 before branching into an NLP/vision core and four AI clusters, but 07-380 does not debut until Fall 2026. The residual Spring 2026 materials for 07-280 and the complete 10-301/601 site already support self-study; retired 15-281 remains a useful legacy route.

learning deep-dive

The Conference as a Content Factory: AI Engineer's Structural Advantage

AI Engineer reached 600,000 YouTube subscribers in under three years not because it mastered video production, but because it barely needs to produce videos at all: recordings from eight conferences a year create an inexhaustible supply of YouTube material. The real constraint on content creation is structure, not skill.

A Global Map of AI and CS Courses: Which Ones Can You Actually Study in Public?

This map audits AI and CS courses at Stanford, CMU, MIT, and UC Berkeley in 2025–2026 using four access labels: A0 for a visible catalog entry, A1 for a public syllabus, A2 for partial materials, and A3 for a self-study-ready package. A course site or YouTube playlist can exist without giving outsiders access to the current videos, assignments, or starter code.

MIT AI/ML Course Guide: Course 6-4 Is a Real AI Degree, but Its Public Materials Span Three Eras

MIT has offered Course 6-4, a formal BS in Artificial Intelligence and Decision Making, since 2022. For an outside learner, however, the current degree requirements, the 2025–2026 course sites, and the best OCW editions rarely line up. A workable route follows 6-4's programming, algorithms, linear algebra, and probability foundation, then selects among 6.S191, 6.3900, 6.4110, 6.7960, vision, and robotics according to what is actually public.

Stanford CS103: A Math Course Whose First Assignment Is Installing a C++ Compiler

CS103 teaches you how to write proofs, then teaches you what can't be proven — but the part nobody mentions is that it ships C++ programming assignments, starting with PS0: install Qt Creator. Its real asset is a shelf of homegrown 'Guide to X' handouts and a Proofwriting Checklist that graders actually deduct points against, all public. Solutions and practice exams sit behind Stanford login, and the Honor Code page explains why.

Stanford CS107: The Same Course Weights Assignments at 40% One Quarter and 20% the Next

CS107 runs from Unix and C all the way to x86-64 and writing your own malloc, across seven assignments. But line up four archived syllabi and the course stops looking like one course: assignments are worth 40% in three quarters and 20% in Summer 2026, where in-class quizzes take 40%. The resubmission policy exists only in the quarters Cain taught; Troccoli's quarter has none. The one assignment that accepts no late days is the final heap allocator. And what blocks a self-learner isn't the autograder — it's that every starter repo lives on AFS.

Stanford CS109: A Probability Course That Turned "How to Read This Lecture With an LLM" Into Official Coursework

Every lecture in CS109's Summer 2026 offering ships with an official LLM Learning Guide — six concepts, a Learn prompt and a Test me prompt for each, written week by week across the quarter for a total of 23 PDFs. The same course's honor code Rule 4 forbids asking an LLM to solve your homework, and 65% of the grade sits in proctored exam rooms. Those two facts are halves of one design.

Stanford CS111: Nine Assignments Build an Operating System, and the Exams Don't Test Them

CS111's nine assignments run from lambdas to crash recovery in a journaling file system. Reading the site page by page turns up three things the syllabus blurb never mentions: assignment 3 is the point of no return, because assignment 4 compiles your assignment 3 code; a whole block of the final exam asks for definitions of ethics terms, and the public practice sheet ships with answers; and pasting your own code into an AI tool to ask about it is written down, in plain words, as an Honor Code violation.

Stanford CS161: The Algorithms Course That Lists Writing Clearly as Its Third Learning Goal

The first slide of CS161 names three goals: design, analysis, communication. The third one is why handwritten homework scores zero and why solutions have to read like a memo to a colleague. Of the eight problem sets, HW2 is the wall. The lecture notebooks exist to show that timing runs can't tell you which algorithm is faster. And the summer offering is a completely different course wearing the same number.

Stanford CS161 Lecture 1: Why Algorithm Analysis Starts with Karatsuba Multiplication

Splitting two n-digit integers in half still creates four recursive products and leaves the runtime at n². Karatsuba reconstructs the cross term with (a+b)(c+d)-ac-bd, cuts the branching factor to three, and reaches roughly n^1.585.

Stanford CS161 Lecture 2: From an InsertionSort Proof to MergeSort's n log n

Lecture 2 turns 'fast' into a worst-case bound that can be proved. A loop invariant establishes InsertionSort's correctness while its worst case is n²; a recursion invariant and O(n) work per level give MergeSort O(n log n).

Stanford CS161 Lecture 3: Reading a Recursion Tree Through the Master Theorem

For T(n)=aT(n/b)+O(n^d), the central comparison is branching growth a versus per-problem shrinkage b^d. Equality makes every level equally heavy, a<b^d makes the root dominate, and a>b^d makes the leaves dominate; outside the template, use substitution.

Stanford CS161 Lecture 4: How Median of Medians Guarantees Linear-Time Selection

Selection does not require sorting. Median of medians groups elements by five, selects the median of the group medians as a pivot, and guarantees that the larger recursive side has at most 7n/10+5 elements; substitution proves O(n) worst-case time.

Stanford CS161 Lecture 5: Proving Randomized QuickSort's Expected Time

Randomized QuickSort has O(n log n) expected time on every fixed input but Θ(n²) worst-case time. The valid proof does not substitute expected subproblem sizes into a recurrence; it computes the probability that each pair is compared.

Stanford CS161 Lecture 6: Sorting Lower Bounds and Linear-Time Radix Sort

The Ω(n log n) lower bound applies to comparison sorting. When integer keys can index buckets directly, stable Counting Sort can power Radix Sort and achieve O(n) under conditions such as M≤n^c.

Stanford CS161 Lecture 7: Binary Search Trees, Red-Black Trees, and the Source of Worst-Case O(log n)

Ordinary BST operations cost O(h) and can degrade to O(n); five red-black invariants cap the height at 2 log₂(n+1), giving search, insertion, and deletion worst-case O(log n) bounds.

Stanford CS161 Lecture 8: Hashing, Collisions, and What Expected O(1) Actually Guarantees

A universal hash family only needs to keep the collision probability of every distinct key pair at most 1/n; that makes the expected bucket size below 2, yielding expected O(1), not per-operation worst-case O(1).

Stanford CS161 Lecture 9: Graph Representations, DFS, BFS, and Proofs About Search Order

DFS and BFS both scan an adjacency-list graph in O(n+m); DFS finish times produce a topological order for a DAG, while BFS layers equal exact unweighted shortest-path distances.

Stanford CS161 Lecture 10: Why Two DFS Passes Find Strongly Connected Components

Contracting each SCC always produces a DAG; first-pass DFS finish times order those components, and a second pass on the transposed orientation discovers exactly one SCC per DFS tree in O(n+m).

Stanford CS161 Lecture 11: Dijkstra, Bellman-Ford, and Two Orders of Relaxation

Dijkstra finalizes the minimum estimate and relies on nonnegative weights; Bellman-Ford repeatedly relaxes every edge, spending O(nm) to support negative edges and detect a negative cycle reachable from the source.

Stanford CS161 Lecture 12: Dynamic Programming with Bellman–Ford and Floyd–Warshall

Dynamic programming starts by defining subproblems, derives a recurrence from optimal substructure, and evaluates states in dependency order; Bellman–Ford layers by edge count, while Floyd–Warshall layers by allowed intermediate vertices.

Stanford CS161 Lecture 13: Designing Dynamic Programs for LCS, Knapsack, and Independent Set

Lecture 13 turns dynamic programming into five steps: choose a state, derive transitions, fill the table, reconstruct a solution, and then improve the implementation. LCS takes O(mn), both knapsack variants take O(nW) pseudo-polynomial time, and maximum-weight independent set on a tree takes O(|V|).

Stanford CS161 Lecture 14: When a Greedy Algorithm Turns Local Choices into a Global Optimum

A greedy algorithm is not merely 'pick what looks best.' It keeps one choice at each step and needs an exchange argument proving that the choice preserves an optimum. Lecture 14 develops that proof pattern through activity selection, weighted completion time, and Huffman coding.

Stanford CS161 Lecture 15: Proving Prim and Kruskal with the Cut Property

The heart of MST algorithms is an invariant: the selected edges remain contained in some MST. The cut property proves that every step of Prim and Kruskal is safe.

Stanford CS161 Lecture 16: Ford–Fulkerson, Residual Networks, and Max-Flow Min-Cut

Ford–Fulkerson augments through a residual network. When no path remains, residual reachability yields a cut equal to the flow, certifying max flow, min cut, and their equality.

Stanford CS161 Lecture 17: Gale–Shapley and Revocable Greedy Choices

Deferred Acceptance permits tentative choices to be revoked. Monotone proposals prove O(n²) termination and stability, with an outcome favoring the proposing side.

Stanford CS161 Lecture 18: From the Algorithmic Toolbox to LP, Coding, and ML

The finale recaps the CS161 toolbox and points toward LP duality, Reed–Solomon coding, and ML-assisted algorithms. Officially, this lecture has slides but no notes.

Reading Guide: Pick What Most People Use — the Other Five Criteria Are Tie-Breakers

The primary criterion has not changed: it is still adoption — and AI makes it matter more, not less, because more users means more training data means higher agent accuracy. The five criteria this series collects (machine-readable docs, types, whether the source is in your repo, data shape, machine-callability) are for breaking ties when adoption is comparable, or for costing out what picking the less popular option will charge you.

AI SDK Message Parts: The Data Skeleton of a Conversation UI

The AI SDK splits an AI message into a parts array — text, reasoning, source-url, tool-* — each an independent typed fragment (introduced in v5, unchanged since). That data structure dictates how modern AI conversation UIs are written: render by switching on part.type, handing each fragment to its component. This post unpacks the design logic of the parts model, useChat's streaming behavior, and how it became the foundation for component libraries like AI Elements.

Drizzle ORM: A SQL-First Database Access Layer for TypeScript

Drizzle ORM is a SQL-first TypeScript ORM — its query builder reads like SQL, so queries written by agents are auditable in diffs. Zero dependencies, ~7.4 KB gzipped, native support for edge databases like Cloudflare D1, Neon, and Turso. Still at version 0.45.2 with no 1.0, yet weekly downloads have reached 16.9 million — surpassing Prisma's 13.8 million.

llms.txt: The Copy of Your Docs Written for Machines

llms.txt is a convention proposed by Jeremy Howard on 2024-09-03 (the spec is now at v2): a Markdown index at your site root written for LLMs. Hand-tested across six frontend docs sites: TanStack, shadcn, Zustand, AI SDK, and Next.js all ship it; React Router is the lone 404. The companion llms-full.txt (full-text version) is live at Anthropic, Cloudflare, and others. This post covers the spec, who uses it, and why it has started to influence library selection.

shadcn Registries and MCP: The Third Way to Distribute Components

Component distribution used to offer two roads: npm packages (black-box dependencies) or manual copy-paste. The shadcn registry standardizes a third — components described as JSON with embedded source and dependencies, installed by CLI straight into your repo as your own code. Anyone can host a registry (AI Elements is one), and the official MCP server lets AI agents browse and install components directly.

Supabase: A Platform Built Entirely on PostgreSQL

Supabase isn't just an open-source Firebase alternative — its core design builds Auth, Storage, and Realtime entirely on PostgreSQL schemas and WAL. The result: everything is queryable with SQL, pgvector works out of the box, and AI agents can operate the entire platform by writing SQL. 108k GitHub stars, Apache 2.0, free tier with 500 MB database.

Tailscale: Your Agent Lives at Home, You're Not Dialing Home

Self-hosting an agent that runs 24/7 means opening something on your own network that must be reachable from outside and must never sit on the public internet. This post takes apart what each Tailscale mechanism actually solves: the tailnet for reachability, subnet routers for private resources, tags plus ACLs for the permission boundary, and seconds-fast policy propagation plus Tailnet Lock for revocation. Pricing checked 2026-08: Personal is free, up to 6 users, unlimited user devices, 50 tagged resources included.

TanStack Router: Making Routes Compile-Time Verifiable

TanStack Router (1.0 in December 2023, ~20M weekly downloads) makes paths, params, and search params compile-time inferred: navigating to a nonexistent route is a type error, not a runtime 404. This post unpacks its three core designs — type safety, first-class search params, and Query-integrated loaders — and why AI agents writing code amplifies their value.

Temporal: Write the Process as Code, and It Finishes Even After a Crash

Temporal is a durable execution platform (Server 1.31.2, Python SDK temporalio 1.31.0, MIT, verified 2026-08). What separates it from BullMQ / Celery isn't scale but the guarantee: a queue guarantees a message gets consumed, Temporal guarantees a multi-call process runs to completion. The price is that Workflow code must be deterministic — and LLM calls are inherently non-deterministic. This post covers how to resolve that tension and when the constraint isn't worth it.

Trigger.dev: Durable Tasks via Process Snapshots, No Determinism Required

Trigger.dev is an Apache 2.0 durable task platform (v4.5.12, checked 2026-08) that uses CRIU to snapshot entire Node.js processes for pause and resume. Unlike Temporal's replay model, it never re-executes your orchestration code and imposes no determinism constraint — LLM calls go directly in the task. The tradeoff: snapshots can't preserve TCP connections (you reconnect manually), and checkpointing is cloud-only — self-hosted deployments don't get it.

WebMCP: Letting a Web Page Hand Its Own Functions to an Agent

WebMCP lets a page register its own functions as agent-callable tools via document.modelContext.registerTool(), replacing the agent's guess-the-button DOM scraping. Chrome opened an origin trial in 149 and estimates stable in 157; Edge followed in 150. But WebKit has formally opposed it ('an agent acting on a user's behalf is, in effect, assistive technology... the site should not single it out for different treatment') and Mozilla filed neutral. This post covers both APIs, where the security gates sit, and whether to invest now with one and a half engines behind it.

tech deep-dive

Testing Five zh-TW Terminology Linters: One Is Usable, One's --fix Turns 只是 Into 隻是

On 109 posts that genuinely contain Mainland vocabulary, zhtw-mcp scored 29.4% precision at 85.2% recall; twlint scored 21.2% / 82.5% — but 195 of twlint's error-level findings are legitimate Traditional characters misread as Simplified (干→幹 60 times, 只→隻 15), so --fix corrupts the text. The most useful result was not a winner: zhtw-mcp found five Mainland terms my own list had missed, and correctly declined to flag one it wrongly included (審計). These tools calibrate your wordlist; they do not replace it.

Zod: From Form Validation to TypeScript's Universal Contract

Zod's 224M weekly downloads (checked August 2026) put it far beyond 'form validation library': API boundaries, environment variables, route search params, LLM tool schemas and structured output all run on the same schemas. The core mechanism is one definition, two payoffs — runtime validation and static types derived from a single source. Zod 4 (on npm July 2025) is faster, slimmer, and easier on tsc.

Behavioral & Ethics Interview Guide: AI Ethics, Teamwork, and Impact Narratives

Behavioral interviews aren't about improvisation — they're about a pre-prepared story library. AI Engineer behavioral interviews have unique focus areas: AI ethics (bias, fairness, privacy), technical decision impact narratives (why you chose this model/architecture), and experience driving ML projects across teams. Strategy: build 8-10 STAR stories, practice each until you can deliver it in under 2 minutes.

Coding Interview Guide: Strategies for ML-Flavored Programming Problems

AI Engineer coding interviews aren't identical to SWE — beyond LeetCode medium, you'll face ML-flavored problems (implementing a tokenizer, writing a batch inference pipeline, handling sparse matrices). Strategy: practice LeetCode medium to 70% pass rate, then spend remaining time on numpy/pandas operations, data processing pipelines, and ML-related programming problems.

Deep Learning Interview Guide: Core Intuitions from CNN to Transformer

Deep learning interviews don't ask you to derive backpropagation — they test whether you can explain the design intuition behind architectures. High-frequency topics: CNN's locality and translation invariance, why the evolution from RNN to Transformer was necessary, self-attention computation and complexity, BatchNorm vs LayerNorm use cases, and common training tricks (learning rate scheduling, gradient clipping, mixed precision).

LLM Application Design Interview Guide: From RAG to Agent Architecture

LLM Application Design is the hottest new interview topic in 2025-2026. Key focus areas: RAG pipeline chunking/retrieval/reranking design, agent tool-use and planning loops, context window management strategies, guardrails and safety design, and LLM application evaluation methods. Interviewers especially value whether you've hit real-world pitfalls.

ML Fundamentals Interview Guide: From Bias-Variance to Evaluation Metrics

ML fundamentals interviews don't test formula memorization — they test whether you can explain concepts intuitively and hold up under follow-up questions. High-frequency topics: the practical meaning of bias-variance tradeoff, the selection logic for L1/L2 regularization, why cross-entropy beats MSE for classification, SGD vs. Adam tradeoffs, and how precision/recall priorities differ by scenario.

ML System Design Interview Guide: From Requirements to Production Architecture

The core of ML System Design interviews isn't choosing the model — it's how to turn a business objective into a system that's deployable, monitorable, and iterable. Interviewers want to see if you can: translate business goals into ML objectives, design data pipelines and feature stores, choose reasonable serving strategies, and plan monitoring and A/B testing.

MLOps & Deployment Interview Guide: From CI/CD to Model Monitoring

MLOps interviews test whether you have experience pushing models to production. Key topics: ML pipeline CI/CD (how it differs from software CI/CD), model registry and version management, A/B testing design and pitfalls, inference scaling strategies (horizontal scaling, model compression, caching), and production monitoring and alerting design.

NLP & LLM Interview Guide: From Tokenization to RLHF

The dividing line in LLM interviews is whether you've actually used these things. High-frequency topics: BPE tokenization logic and multilingual challenges, pretraining objectives (CLM vs MLM), three levels of fine-tuning (full/LoRA/prompt tuning), RLHF workflow and failure modes, prompting as engineering practice, and the difficulty of LLM evaluation with current methods.

AI Engineer Interview Overview: From Company Types to Preparation Strategy

AI Engineer interviews go beyond ML — big tech emphasizes system design and coding, startups look for end-to-end delivery, and AI-native companies test LLM engineering depth. Strategy: identify your target company types first, then allocate prep time across six dimensions (ML fundamentals, system design, LLM applications, coding, paper reading, and behavioral).

Paper Reading Interview Guide: How to Read, Discuss, and a Must-Read List

Paper reading interviews don't test whether you've read that specific paper — they test whether you can quickly understand a new method and identify its limitations. AI-native companies (Anthropic, OpenAI) particularly favor this format. Strategy: practice reading a paper in 30 minutes and verbally stating contribution + limitation, build your own must-read list, and practice summarizing each paper in three sentences.

Stanford CS329A: A Course on Self-Improvement That Says Out Loud What It Can't Improve

CS329A is built around the generation–verification gap: models can produce the right answer but can't tell which one it is. The conclusion the course draws about itself matters more — today's methods make models more consistent, not smarter. Nine lectures are public, out of twenty.

A Reading Guide to Stanford's CS Courses: Ordered by Prerequisites, from CS106A to CS336

Stanford CS rests on CS103, CS107, CS109, CS111, and CS161; CS221 names three of those plus CS106B as preparation. This guide combines official prerequisites with an explicitly editorial reading order and marks public-material and offering risks.

In the AI Era, Taste Is an Amplifier

AI pushes execution cost toward zero. People with good taste create more value; people with poor taste create more garbage. The difference is not whether you can use AI, but whether your mind contains something worth amplifying before you use it. This series documents my attempt to sharpen judgment systematically.

AI Product Design Interview Guide: From Human-in-the-Loop to Trust Building

AI Product Design is the hottest new interview topic in 2025-2026. Core areas: when to use AI (not every problem needs it), human-in-the-loop design patterns (when to let humans intervene), trust building (how to make users believe AI output), AI product challenges (hallucination, latency, cost), and AI product evaluation metrics.

Behavioral & Leadership Interview Guide: Influence, Conflict Resolution, and Vision

Product Builder behavioral interviews differ from SWE — they don't just test teamwork, they specifically test how you drive things without formal authority. Core skills: influence narratives (how to convince engineers to build your feature), conflict resolution (disagreements with designers/engineers/stakeholders), vision expression (how to make someone understand your product direction in 30 seconds), and failure stories (learning from failure without deflecting blame).

Execution Interview Guide: From Roadmap to Cross-Team Collaboration

Execution interviews test whether you can turn ideas into deliverables. Core skills: roadmap planning (how to prioritize with limited resources), priority defense (why A before B), cross-team collaboration (how to drive engineering and design), stakeholder management (how to handle conflicts), and the ability to track progress with data.

Growth & Experimentation Interview Guide: From Growth Loops to Experiment Design

Growth interviews don't test whether you can growth hack — they test whether you have systematic growth thinking. Core skills: growth loop design (the acquisition → activation → retention → referral flywheel), experiment design (the full hypothesis → metric → experiment → analysis process), retention strategy (finding the aha moment, designing habit loops), and using data to decide what's worth continued investment.

Metrics & Analytics Interview Guide: From North Star to Experiment Design

Metrics interviews test whether you can make decisions with numbers, not how much statistics you know. Core skills: north star metric selection logic (why this one and not that one), metric tree decomposition (finding actionable levers), funnel analysis (which step's drop-off is most worth fixing), A/B testing design and pitfalls, and judgment when facing counterintuitive data.

Product Builder Interview Overview: From PM to Builder Mindset

A Product Builder isn't a traditional PM — you need to build from 0 to 1, not just write PRDs. Interviews test the intersection of product intuition, metrics thinking, technical understanding, and execution ability. Prep strategy: first figure out whether your target company wants a PM or a Builder, then allocate time across nine dimensions.

Product Design Interview Guide: From Problem to Solution

Product Design interviews don't test whether you can draw wireframes — they test how you go from problem to solution. Core skills: MVP scope judgment (what to build and what not to), trade-off analysis (speed vs completeness, generic vs custom), communication of design decisions (why A instead of B), and iterative thinking.

Product Sense Interview Guide: From User Insight to Feature Prioritization

Product Sense interviews don't test how many features you can think of — they test whether you can find the problem truly worth solving within a vague requirement. Core skills: user segmentation thinking, problem reframing (turning 'add a feature' into 'what problem are we solving'), structured reasoning for feature prioritization, and the ability to hold or revise your judgment under follow-up questions.

Strategy Interview Guide: From Market Positioning to Competitive Moats

Strategy interviews don't test whether you can recite frameworks — they test whether you can make judgments with incomplete information. Core skills: market sizing (the practical use of TAM/SAM/SOM, not rote numbers), competitive moat analysis (network effects, switching costs, brand), go/no-go decisions for new markets, and using elimination rather than addition for strategic trade-offs.

Technical PM Interview Guide: From API Design to Architecture Understanding

Technical PM interviews don't require you to write production code, but you need to be able to read trade-offs. Core skills: API design fundamentals (RESTful, versioning, error handling), high-level system architecture understanding (microservices, database selection, caching), collaboration patterns with engineers (RFC process, technical spec review), and making product decisions under technical constraints.

Choosing Among the Three AWS AI Certifications: MLA-C01 Has 40 Days Left, and Only in English

AIF-C01, MLA-C01, and AIP-C01 are not a difficulty ladder — they are three different job surfaces: AIF tests whether you can talk about it, MLA tests putting ML into production, AIP tests integrating someone else's foundation models into a system. But in August 2026 the choice is gated by time: the official certification page announces that the last day to take MLA-C01 in English is September 28, 2026 — 40 days from today — while MLA-C02 registration does not open until September 1 and its exam guide is unpublished. Non-English candidates (Japanese, Korean, Simplified Chinese) have a materially longer window. The same page also carries two codes, MLA-C02 and ME1-C02, with no stated relationship. And the renewal graph works backwards on ordering: passing AIP-C01 renews AIF-C01, MLA-C01, and Data Engineer – Associate for three years each.

Choosing Among the Four Claude Certifications: First Check Whether You Can Even Register

Before comparing exam objectives there is one fact that outranks all of them: registration for the Claude certifications is open only to organizations in the Claude Partner Network — individuals cannot sign up. For those who clear that gate, four things actually decide the answer. CCAO-F ($99) does not count toward partner tier eligibility while the other three do. Claude Code is 20% of CCAR-F but only 3.1% of CCDV-F — the architect exam tests the tool far more heavily than the developer exam. CCAR-P is 28% non-technical (governance 14% plus stakeholder communication 14%), which nothing else in this series is. And CCAO-F is the cheapest but only 14% Prompting; Output Evaluation at 21% is its real spine. All four are valid 12 months, with retakes at 14 / 30 / 90 days and 4 attempts per rolling 12 months.

Which Microsoft AI Certification: The Forks Between AI-103, AI-500, AB-620, and AB-100

Across Microsoft's four AI/agent certifications, only AI-103 → AI-500 is an official ladder; everything else is positioning. Three forks decide it: whether you write Python (AI-103/AI-500 vs AB-620), whether you build or judge (AB-100 vs the rest), and whether you can actually start today — AI-500's four official learning paths currently 404, AB-620 has no practice test, AB-100 has a free one. For readers who prefer Chinese there is a fourth fork: AI-103 and AB-620 offer Traditional Chinese, AI-500 and AB-100 are English only. All four cost $165, expire after one year, and renew free but only inside a six-month window.

Choosing Among NVIDIA's Four: Two Can't Be Registered For, Training Is All Paid, and the Docs Contradict Themselves

NVIDIA's generative AI line has four exams: NCA-GENL and NCA-GENM ($125 each, associate), NCP-GENL and NCP-AAI ($200 each, professional). Three decision inputs no other vendor forces on you. One: both professional exams still show 'Coming soon' next to Register, so any near-term plan is down to the two associates. Two: NVIDIA is the only vendor in this series whose official prep courses are all paid — real cost is exam fee plus courses, and the self-paced totals are $390 (NCA-GENL), $210 for only three of five courses (NCA-GENM), and $1,620 list price across NCP-GENL's five. Three: the official documents disagree with themselves — NCP-AAI's weights total 98% on the web page and 92% in the PDF, and two cells of NCP-GENL's web table carry misplaced text, one of it about OpenUSD. Lock-in also varies sharply: NCP-AAI is 7% NVIDIA-specific, NCP-GENL is 31% GPU and model-compression work.

Aider: The Oldest Terminal AI Pair Programmer, and Where Its Maintenance Stands

Aider is a terminal AI pair programmer dating back to 2023 (Python, Apache-2.0, ~48.3k stars), designed against the grain of today's autonomous agents: you control context by hand with /add, every edit becomes its own atomic git commit, and architect/editor mode splits planning from editing across two models. But note the maintenance cadence: the latest PyPI release is 0.86.2 from 2026-02, the last commit was 2026-05, and the site still recommends Claude 3.7 Sonnet and o1.

Amp: The Coding Agent That Defines Itself by What It Deletes

Amp spun out of Sourcegraph in December 2025 as Amp Frontier Corporation, and its npm package moved from @sourcegraph/amp to @ampcode/cli. Its defining trait is deletion: the editor extension, Amp Tab, TODO lists, Fork, custom commands, and public threads have all been removed. Monthly subscriptions only arrived on 2026-07-18 (Megawatt $20, Gigawatt $200); before that it was pay-as-you-go only. The current focus is orbs — remote machines that keep working after you close your laptop.

GitHub Copilot CLI: An Agent That Runs on GitHub the Platform

Copilot CLI went GA on 2026-02-25 and is included in every Copilot plan, Free included. Its differentiator isn't the agent — it's the GitHub integration: a built-in GitHub MCP server that works on issues and PRs, org policies inherited automatically, and an `&` prefix that hands work to the cloud coding agent. Billing runs on GitHub AI Credits (1 credit = $0.01): Pro $10/mo includes $15, Pro+ $39 includes $70, Max $100 includes $200.

omp (Oh My Pi): The Fork That Inverts Pi's Minimalism

omp is a fork of Pi, but it is not just a plugin layer stacked on top: it adds roughly 80,000 lines of Rust, pulling grep, shell, AST, and PTY in-process. Built-in tools go from Pi's 7 to 31, plus 14 LSP ops, 28 DAP ops, and 60+ providers. One codebase, two opposite bets.

Choosing a React Stack in the AI Era: From the TanStack Trio to the Full Map

TanStack Router (19.7M weekly downloads) + Query (55.8M) + Zustand (44.5M) as the core, with Vite, react-hook-form + Zod (224M), Tailwind + shadcn, and Vitest + Playwright — the current default stack for serious SPAs. The AI era adds three new selection criteria: does the docs site ship llms.txt (all of TanStack does; React Router doesn't), can type safety act as an agent guardrail, and does the source code live in your repo where an agent can read it.

AI Elements: Vercel's ChatGPT-Style Interface as Copy-In shadcn Blocks

AI Elements is Vercel's React component library for the AI SDK ecosystem — the registry currently holds 48 components covering Conversation, Reasoning, Sources, Tool, and the rest of the AI-interface vocabulary. It follows the shadcn model: npx ai-elements@latest copies the source into your project, fully editable, mapped one-to-one onto useChat's message parts.

ai deep-dive

AI Agents Generating Slides: Letting the Model See Its Own Layout

The 2026 consensus for agent-built slide decks: outline-first, separate content from construction, then render to images and let a fresh-eyes subagent do visual QA. Anthropic's and OpenAI's official slides skills both converged on PptxGenJS plus a visual verification loop, and the research line (PPTAgent → PreGenie → DeepPresenter) points the same way. But two later corrections matter: PresentBench shows the widely cited PPTEval scores too generously, and SeaSlides argues the model should not write free-form HTML/SVG at all.

AI Governance Frameworks vs. Exam Objectives: EU AI Act, NIST AI RMF, ISO/IEC 42001 — and Why No Certification Names Them

Governance carries more weight on these exams than most engineers expect — CCAR-P is 14% governance plus 14% stakeholder work (28% non-technical), CCAO-F is 15%, AB-100's deploy-and-govern block is 40–45%, AIF-C01 is 14% responsible AI plus 14% security/compliance/governance. But across all fifteen official exam guides in this series, not one names the EU AI Act, the NIST AI RMF, or ISO/IEC 42001; the only regulations any of them names are CCAR-P's GDPR, HIPAA, and FedRAMP. So the use of this post isn't memorizing frameworks for an exam — it's using the three frameworks as a skeleton to file six certifications' scattered governance objectives. The three split cleanly: the EU AI Act is law (fully applicable 2026-08-02, high-risk duties pushed to 2027-12-02 by the AI Omnibus), the NIST AI RMF is voluntary (GOVERN/MAP/MEASURE/MANAGE, and 1.0 is currently being revised), and ISO/IEC 42001 is a certifiable management system standard whose clauses sit behind a paywall — so this post uses only what ISO's own public page states.

AWS AI Practitioner (AIF-C01): v1.1 Turned It Into an Agentic AI Exam

The AIF-C01 exam guide moved to v1.1 on April 30, 2026, adding seven objectives at once — MCP, multi-agent patterns, context engineering, token-based pricing, and hallucination detection all became testable, and Bedrock AgentCore, Kiro, and Strands Agents joined the in-scope services. Almost every summary in circulation describes the older version. This guide builds a four-week path on the official five-domain weighting (20/24/28/14/14). Official specs: $100, 90 minutes, 65 questions (50 scored), pass at 700, valid 3 years — and it is the only exam in this series offered in Traditional Chinese.

AWS GenAI Developer Professional (AIP-C01): The Whole Exam Is About Integrating Someone Else's Model

AIP-C01 is AWS's only professional-level certification dedicated to GenAI application development, and it explicitly excludes model development, advanced ML, and feature engineering — it tests integrating someone else's foundation model into a production system. Domain weights are 31/26/20/12/11, and nearly half the skills in the heaviest domain sit in vector stores and RAG. A March 2026 refresh added Bedrock AgentCore and the beta ended March 31, so anything older is stale. Official specs: $300, 180 minutes, 75 questions (65 scored), pass at 750, valid 3 years — and passing it renews AIF-C01, MLA-C01, and Data Engineer – Associate at the same time.

Claude Certified Architect Foundations Exam Complete Guide

A complete study guide for Claude's official architect certification (CCAR-F): five domains weighted 27/18/20/20/15, four scenarios drawn from six, common anti-patterns, and hands-on preparation. Official specs are 60 items / 120 minutes / $125 / 12-month validity / 720 to pass; registration is limited to Claude Partner Network members, and on-time renewal is free and non-proctored.

Claude Certified Architect Professional (CCAR-P): 28% of It Isn't Technical

CCAR-P is the most expensive and most senior of Anthropic's four exams ($175, 63 items, 120 minutes). Integration is the heaviest domain at 19%, but what really separates it from everything else in this series is the other two: Governance, Safety & Risk Management at 14% and Stakeholder Communication & Lifecycle Management at 14% — 28% combined on compliance, risk, discovery interviews, and delivery lifecycle rather than code. The guide names GDPR, HIPAA, and FedRAMP, and its Intended Audience explicitly excludes entry-level developers and anyone doing 'prompt writing without broader system design responsibility.'

Claude Certified Associate (CCAO-F): The Heaviest Domain Is Knowing When Claude Is Wrong

CCAO-F is the cheapest of Anthropic's four exams ($99, 60 items, 120 minutes), aimed at people who work with Claude rather than build against it. The heaviest of its seven domains is Output Evaluation and Validation at 21% — spotting hallucinations, deciding when human review is required, and adapting outputs — with Governance, Risk, and Responsible Use at another 15%. Anthropic states plainly that it is not for developers building against APIs or designing agentic systems. One easily missed limitation: this credential does not count toward Claude Partner Network tier eligibility, while the other three do.

Claude Certified Developer (CCDV-F): A Third of It Is Ordinary Software Engineering

CCDV-F is the engineer's exam among Anthropic's four certifications. The official blueprint has eight domains, and the heaviest — Applications and Integration at 33.1% — is led by Claude Application Design (8.6%) and Software Engineering Foundations (7.4%), meaning a third of the exam is API mechanics and ordinary software engineering. The counterintuitive part: Claude Code is only 3.1% and Eval only 2.6%, while the sibling Architect exam gives Claude Code 20%. Official specs: $125, 53 items, 120 minutes, pass at 720, valid 12 months, registration limited to Claude Partner Network organizations.

Cost, Latency, and Availability Across Six Exams: One Topic Tested From Three Altitudes

Google PMLE, AWS AIF-C01 and AIP-C01, Microsoft AI-103 and AI-500, and NVIDIA NCP-GENL all test how to make a GenAI application fast, cheap, and reliable — and they form a three-rung ladder: AIF-C01 asks whether you know cost scales with tokens, AIP-C01 and the two Microsoft exams ask whether you can instrument and control it, NCP-GENL asks whether you can change the model and the hardware. Three different altitudes. NVIDIA works at the kernel and quantization layer (Model Optimization 17% + GPU Acceleration 14% = 31%, the heaviest single cost/latency block in the whole series), AWS and Microsoft at the application layer (three caching tiers, token caps, chargeback), and Google at the MLOps layer (CPU/GPU/TPU evaluation, data vs model parallelism, scaling serving backends by throughput). The shared core is eight levers, but each lever becomes a different question at each altitude. This post deliberately carries no prices and no hardware specs — that is the part of this topic that rots fastest.

Preparing for Google PMLE After the Exam Guide Rewrite

Google's Professional ML Engineer exam guide was rewritten in 2026: Vertex AI is renamed Gemini Enterprise Agent Platform throughout, so older study material no longer matches the product names in the questions. This guide uses the official six-section weighting as its skeleton, listing what each section tests, which official materials cover it, and what to build — plus a study schedule whose reasoning is spelled out. Official specs: $200, two hours, 50–60 multiple-choice and multiple-select questions, two-year validity, 3+ years of industry experience recommended including 1+ year on Google Cloud.

The Research Side of Hermes Agent: Batch-Running Thousands of Prompts Into Training Data

`batch_runner.py` runs thousands of prompts in parallel into ShareGPT-format tool-calling trajectories, lets each prompt name its own container image, and resumes by matching prompt content rather than index. Two quality filters run before you see the data: samples with zero reasoning are discarded, and entries calling hallucinated tool names are dropped at merge time. This is why a research lab builds a personal agent — the agent is the data pipeline.

The Hermes Agent Gateway and Scheduler: An Unattended Agent's Biggest Risk Is Spending Your Money

One gateway process fronts 30-plus chat platforms and denies every user not on an allowlist or paired by DM. The scheduler adds two unusual guards: pre-dispatch validation marks a misconfigured job `blocked_config` without making a single LLM call, and the model drift guard makes unpinned jobs fail closed when the global model changes — protecting you from an hourly job quietly following you onto a paid model.

Installing and Upgrading Hermes Agent: Check the Support Tier Before You Pick a Path

Hermes install paths come in three support tiers: macOS (Apple Silicon), Windows 10/11, Linux/WSL2, and Docker are Tier 1; Termux and Nix are Tier 2; pip, brew, AUR, and Intel Macs are explicitly unsupported — fixes for those won't be merged. On the upgrade side, `hermes update` snapshots state first, then compiles nine critical files after the pull and hard-resets the checkout if any fail to parse.

Hermes Agent: Nous Research's Self-Improving Agent, and Its Real Relationship With OpenClaw

Hermes Agent is Nous Research's MIT-licensed agent framework, built around a learning loop: it writes its own skills, curates its memory, and searches past sessions with FTS5. It ships `hermes claw migrate` to move you off OpenClaw — but OpenClaw was not replaced, and both projects are still moving. This is the series opener: what it is, how it differs, and when not to pick it.

Memory and Skills in Hermes Agent: A System That Rewrites Itself, and Where You Can Intervene

Hermes memory has hard caps: 2,200 characters for MEMORY.md and 1,375 for USER.md, and an over-limit write returns an error instead of auto-compacting, forcing the agent to make room itself. Skills are maintained by a curator that runs every 7 days after 2 hours of idle, marks skills stale at 30 days and archives at 90 — but never deletes. The switches actually worth flipping are `memory.write_approval` and `skills.write_approval`, which stage the background self-improvement writes for review.

Migrating From OpenClaw to Hermes Agent: What Moves, What Doesn't, and the Archive Directory

`hermes claw migrate` imports persona, memory, skills from four locations, model and provider config, platform tokens, and the approval allowlist — but secrets are never imported silently, and even `--preset full` requires an explicit `--migrate-secrets`. What can't move (cron jobs, plugins, hooks, the multi-agent list, deep channel config) isn't discarded but parked in `~/.hermes/migration/openclaw/<timestamp>/archive/` for manual work. Coming from Claude Code or Codex is a different command: `hermes import-agent`.

Hermes Agent Model Providers: The Subscription Billing Trap, and Why Fallback Fires Only Once

Hermes supports 40+ providers, and the consumer-subscription OAuth paths are where billing surprises live: Anthropic OAuth only spends Claude Max extra-usage credits, and Claude Pro can't use it at all. Auxiliary tasks default to `provider: auto`, meaning your expensive main model does compression and vision grunt work. The fallback chain is a one-shot switch per session, not continuous retry.

The Hermes Agent Security Model: There's a Floor Below --yolo That You Can't Remove

Approvals default to smart mode: an auxiliary model waves through low-risk commands, auto-denies genuinely dangerous ones, and escalates the uncertain cases to you. Neither `--yolo` nor `approvals.mode: off` can disable the hardline blocklist (`rm -rf /`, fork bombs, `dd` to a physical disk), and `approvals.deny` is its user-editable counterpart, evaluated before yolo. Upstream is explicit that the threat model is an honest-but-wrong agent, not an adversarial process.

Hermes Agent's Seven Terminal Backends: Moving to a Sandbox Turns Off Dangerous-Command Approval

Hermes can run commands on seven backends: local, ssh, docker, singularity, modal, daytona, and vercel_sandbox. The decisive trade-off isn't performance, it's approval — local and ssh run dangerous-command checks, the other five skip them entirely because the container is treated as the boundary. Also, Docker defaults to one long-lived container shared across sessions, not a fresh environment per conversation.

The Nous Tool Gateway: One Subscription Instead of Four Accounts, at the Cost of Concentrating Your Tool Supply Chain

The Tool Gateway routes four tool categories — web search (Firecrawl), image generation (nine FAL models), TTS (OpenAI), and cloud browser (Browser Use) — through Nous infrastructure, replacing four signups with one OAuth. It's per-tool rather than all-or-nothing, and `use_gateway: true` overrides any direct key in your `.env` — the precedence rule people most often get wrong.

The Hermes Agent Tool Layer: What Happens When 3,300 MCP Tools Won't Fit in Context

Attach enough MCP servers and the tool schemas alone eat your context — upstream's extreme example is Cloudflare's ~3,300 tools, whose names alone run about 32K tokens. Hermes answers with Tool Search: MCP and non-core plugin tools collapse into three bridge tools and schemas load on demand, while core tools never defer. Separately, plugins are disabled by default and only run when named in `plugins.enabled`.

Microsoft AB-100: The Architect Exam — Don't Prepare From the Blurb on Its Own Page

AB-100 is the architect tier of Microsoft's agent line, weighted 25-30 / 25-30 / 40-45 with deployment and governance heaviest. Its outline runs on verbs like design, recommend, and propose — it tests judgment, not configuration. Three things to know first: the scope blurb on the official exam page is wrong (it is information-protection and DLP boilerplate, which I verified verbatim), so prepare from the study guide instead; it has a free practice assessment, the only one of Microsoft's three agent credentials that does; and the 15 associate certifications it lists are described as usable, not required. Official specs: $165, English only, pass at 700, one-year validity.

Microsoft AB-620: The Low-Code Agent Track on Copilot Studio

AB-620 is the low-code branch of Microsoft's agent certification line — it tests agent flows, adaptive cards, computer use, MCP tools, A2A, and Fabric data agents in Copilot Studio, not Python. The three skill areas weigh 30-35 / 40-45 / 20-25, with integration the heaviest. Official specs: $165, 120 minutes, pass at 700, one-year validity, and 13 languages including Traditional Chinese — the only localized exam of Microsoft's three agent credentials. It is generally available, but the practice assessment is not out yet.

Microsoft AI-103: After the Foundry Rename, Every Older Azure AI Study Guide Is Void

AI-103 replaces AI-102, retired June 30, 2026, and the objectives were rewritten around Microsoft Foundry — prompt flow, Azure AI Studio, Azure OpenAI Service, and Azure AI Agent Service appear nowhere in them. The five skill areas weigh 25-30 / 30-35 / 10-15 / 10-15 / 10-15, with generative AI and agents the largest. Official specs: $165, 120 minutes, pass at 700, offered in Traditional Chinese, may include interactive components — and it is valid for only one year, though renewal is a free, open-book, unproctored online assessment.

Microsoft AI-500: The Objectives Are Published, the Training Isn't

AI-500 is a rare thing among the major clouds — an expert-level certification dedicated to multi-agent systems, weighted 15-20 / 30-35 / 20-25 / 20-25, naming Agent Framework, LangGraph, Hugging Face Transformers, MCP servers on Azure Functions / Logic Apps / API Management, A2A, Key Vault, and the AI Red Teaming Agent. Three constraints come first, though: it is still in beta (scores wait for rescoring), it requires AI-103 before you can take it, and the official training is not live — the four learning paths listed on the exam page all return 404 today, and the instructor-led course opens 2026-09-30.

Multi-Agent Architecture Across Five Exams: The Shared Core and What Doesn't Transfer

Microsoft AI-500, AB-620, AB-100, NVIDIA NCP-AAI, and Claude CCAR-F all test multi-agent architecture, and they overlap on seven things: orchestration topologies, A2A and MCP, per-agent identity boundaries, three-layer memory, observability and agent replay, human-in-the-loop, and guardrails at four intervention points. But four vendors use four vocabularies for the same ideas, and each exam has objectives that don't transfer — Microsoft names four context-window failure modes nobody else names, 7% of NVIDIA's is locked to NeMo and NIM, and Claude tests SDK-level details like stop_reason. One correction along the way: Google PMLE's wall-to-wall 'Agent Platform' is a Vertex AI rename, not a multi-agent domain.

ai deep-dive

Multimodal Models, First Half of 2026: Native Fusion vs. Bolted-On Vision, and Why Leaderboards Contradict Each Other

Pure image understanding has flattened out — four frontier models all clear 80% on MMMU-Pro within 3 points of each other. The real differentiation is video, long-document OCR, and realtime speech, each with a different leader. But the most useful lesson from assembling these rankings is that two credible sources named different Video-MME leaders more than 10 points apart — and that July and August each turned the field over again.

NVIDIA NCA-GENL: The Name Says LLM, Half the Blueprint Is Classical ML

NCA-GENL is usually what a job posting means by 'NVIDIA Generative AI / LLM certification.' But the official blueprint diverges sharply from the name — Core Machine Learning and AI Knowledge 30%, Software Development 24%, Experimentation 22%, Data Analysis 14%, Trustworthy AI 10% — with LLM and RAG content scattered at bullet level rather than forming a domain, alongside spaCy, NumPy, Keras, and cross validation. The other thing to know first: NVIDIA's official preparation courses all cost money ($30–$500), making it the only vendor in this series without a free official learning path. Official specs: $125, 1 hour, 50–60 items, two-year validity, English only, pass/fail with no score reported.

NVIDIA NCA-GENM: The Multimodal One, With Two Required Courses Only Sold as $500 Workshops

NCA-GENM matches NCA-GENL on price, length, and level but not on emphasis: Experimentation rises to 25% (the heaviest), Core ML drops from 30% to 20%, and two new areas appear — Multimodal Data 15% and Performance Optimization 10%. The content covers U-Net, CLIP, diffusion models, multimodal loss functions, attention maps, and NVIDIA's Riva / NeMo / Triton / ACE SDKs. Watch the cost structure: two of the five recommended courses exist only as $500 workshops with no self-paced option, so a self-study path cannot cover the official set. Official specs: $125, 1 hour, 50–60 items, two-year validity, English only.

NVIDIA NCP-AAI: Registration Isn't Open, and the Official Weights Contradict Each Other

NCP-AAI is NVIDIA's professional-level agentic AI credential — $200, 120 minutes, 60–70 items, two-year validity. Two things come first: registration is not open (the Register button carries a 'Coming soon' label), and NVIDIA's own web page and PDF study guide disagree on the weights — Deployment and Scaling is 13% on the page and 5% in the PDF, Run/Monitor/Maintain is 5% on the page and 7% in the PDF, and the two versions total 98% and 92% respectively. Both are nvidia.com. This guide treats that as a range and an uncertainty rather than picking one.

NVIDIA NCP-GENL: 31% Is GPU and Model Optimization, and Two Cells of the Official Table Are Broken

NCP-GENL is NVIDIA's professional-level LLM credential — $200, 120 minutes, 60–70 items. What separates it from every other GenAI exam is where the weight sits: Model Optimization 17% plus GPU Acceleration 14% is 31% on quantization, distillation, pruning, distributed parallelism, and CUDA profiling — not on calling APIs. Two things first: the Register button says Coming soon, so you cannot sit it yet; and two description cells in the official weight table are corrupted — Fine-Tuning is described with OpenUSD data-interchange text and Model Optimization with deployment text. I verified both verbatim; the correct descriptions are in the official PDF.

How Ten Certifications Actually Test Prompting: Exam Framing vs. Practice

Most people assume GenAI certifications are built around prompt writing. CCAO-F gives Prompting 14% while Output Evaluation gets 21%; CCDV-F gives Prompt and Context Engineering 11.0%. What actually gets tested is structured output, injection-resistant prompting, dynamic context injection, context compression and caching, prompt lifecycle governance, and proving a prompt change helped — closer to context engineering and software engineering than to writing craft. None of the ten asks you to write a prompt on the spot; they are all multiple choice, so explaining why beats having a feel for it. The single most useful line comes from CCAR-F: when business logic must be guaranteed, 'change the prompt first' is usually the wrong answer.

RAG and Retrieval Evaluation Across Four Exams — and One Everyone Assumes Tests It, Which Doesn't

Four certifications genuinely test RAG and retrieval evaluation: AWS AIF-C01 (chapters 2 and 3 total 52%, covering RAG, vector stores, and FM evaluation metrics), AWS AIP-C01 (11 of the 27 skill points in its 31% Domain 1 sit in vector storage and RAG), NVIDIA NCP-AAI (Knowledge Integration 10% plus Evaluation and Tuning 13%), and Microsoft AI-500 ('multi-agent RAG architecture' inside its 30–35% Develop area). Google PMLE contributes exactly one LLM-as-a-judge objective, and Claude CCDV-F — the developer certification people most readily assume covers RAG — has no retrieval objective across its eight domains, with Eval at just 2.6%. Includes a same-vendor foundational-vs-professional comparison, a four-vendor terminology map, non-transferable objectives, and a practice project.

ai deep-dive

Nine Self-Hosted Personal Agents, One Security Question: Where Does the Execution Boundary Go?

OpenClaw has 386k stars to Hermes Agent's 232k, yet Hermes passed it on OpenRouter daily tokens back on 2026-05-10 (224B vs 186B). The nine self-hosted agents that appeared this year aren't nine competitors — they're nine incompatible answers to one question. CVE-2026-44112 broke OpenClaw's own sandbox, and in the Meta alignment director's inbox incident there was no attacker at all: context compaction ate the safety instruction.

Cloudflare Workers AI Model Picking Guide: By Use Case, Price, and Context

The Workers AI catalog currently holds 84 models. For general chat pick glm-4.7-flash ($0.06 / $0.40 per M, 131K context), for vision pick gemma-4-26b-a4b-it ($0.10 / $0.30, 256K), for cheap high-volume steps pick granite-4.0-h-micro ($0.017 / $0.112), and for embeddings pick qwen3-embedding-0.6b or bge-m3 (both $0.012 per M). This post is updated on a schedule.

travel guide

Do Frequent Japan/Korea Travellers Need a Foreign Currency Account? Breaking Down the 1.5% Card Fee and Real Exchange Costs

The 1.5% overseas card fee is 1% network + 0.5% issuer, and dual-currency cards pay it too. Bank of Taiwan quotes KRW only as cash with a 15.7% round-trip spread, versus 2.45% for JPY spot — so a JPY account is worth it, a KRW one isn't (and mostly can't be opened).

ai deep-dive

The 45 Rules of microsoft/AI-Engineering-Coach: An Opinion About Agentic Engineering, Written as Executable Thresholds

A VS Code extension open-sourced by Microsoft employees that reads your local Claude Code / Codex / OpenCode session logs. The real payload is 45 Markdown rules: prompts under 30 characters, sending the next message within 15 seconds of receiving 20 lines of AI code, instruction files over 4,000 bytes — turning 'context engineering' into numbers you can argue with.

CS146S Week 4: What Goes in CLAUDE.md, What Hooks Should Block, Where Subagents Cut

The course lists four techniques for directing agents: instruction files, hooks, commands, subagents. The instruction file is the only one loaded in full every startup, making it config rather than memory; hooks cover what instructions can't, because a rule can be ignored and a hook cannot; commands are the only one a human triggers. The course also marks just one and a half of seven task steps as human work.

CS146S Week 1: A Coding Agent Is, Underneath, a While Loop

Week 1 of CS146S is 'build Claude Code in 200 lines' plus a dissection of production system prompts. The agent loop really is that small. The course slides close with four things Claude does underneath, one of them being `<system-reminder>` tags scattered everywhere to stop the model drifting — which appears in no official documentation.

CS146S Week 5: Express Scores 28, CockroachDB Scores 74 — Agent Readiness Is Measurable

Factory breaks 'can an agent work in this repo' into eight pillars and five levels, and published real scores: CockroachDB L4 (74%), FastAPI L3 (53%), Express L2 (28%). The thesis is that agent readiness approximates the density of deterministic validation loops — linters, type checkers, tests are reward signals for agents.

CS146S Week 7: o3 Found a Linux Kernel Zero-Day at a 1:50 Signal-to-Noise Ratio

The course measured AI SAST false positive rates at 50–100%, against 50%+ for traditional SAST — the genuinely new problem is nondeterminism: run the same prompt twice, get different results, and you can never answer "am I done scanning?" The course lists five agent attack vectors, one of which, intent breaking, attacks the agent's plan itself.

CS146S Week 3: An Agent Skill Is a Folder — the Hard Part Is Two Lines of Description

The Agent Skills spec fits in a sentence: a directory containing a SKILL.md. The real design is three levels of progressive disclosure — only name and description load at startup, the body loads on a match, bundled files load on demand. This site's own repo carries 35 skills and 7,893 lines of SKILL.md, and startup still costs only those 35 metadata pairs.

CS146S Week 6: To Make AI Review Useful, Google Deleted 17 Rules First

Google deployed AutoCommenter to tens of thousands of engineers and published the whole tuning process: suppressing 17 'technically correct but low-value' rules raised the useful ratio from 54% to 66%, with 80% set as the bar for the next rollout stage. Final comment-resolution rate landed around 40%. The bottleneck in AI code review was never detection — it's volume.

CS146S Week 9: One Person Wiring Up MCP Is Fine; Three Hundred Need a Gate

How an individual connects tools is a preference; how an organization does it is governance — who can touch what data, where keys live, whose budget it lands on. Anthropic's published record of ten internal teams contains a good indicator: security engineering accounts for 50% of all custom slash commands in the entire monorepo. Adoption doesn't spread evenly; it takes off first in teams that already build their own tools.

CS146S Week 8: Once Agents Run in the Cloud, the Bottleneck Moves from Waiting to Reviewing

Background agents replace 'you watch it run' with 'it finishes and opens a PR.' Every vendor's design converges on the same parts: an isolated environment, external triggers (issues, Slack, Linear), and a PR as the output. The genuinely new problem is that you become the bottleneck — five agents finish at once, five diffs queue for you, and none of them know the others exist.

CS146S Week 2: Context Engineering, RePPIT, and MCP's 98.7% Cut

Fall 2026 compresses a full week of prompting into one bullet here and adds RePPIT (Research, Propose, Plan, Implement, Test) and MCP. Two RePPIT rules are worth stealing outright: always ask for exactly two proposals, and never let the instance that wrote the code review it. On the MCP side, Anthropic measured turning tools into code calls dropping 150,000 tokens to 2,000.

Stanford CS146S, Two Syllabi Side by Side: What Changed in a Year

Stanford CS146S's Fall 2026 syllabus compresses prompting from a full week into a single bullet, drops the terminal and UI-generation weeks, and adds Agent Skills, Agent-Ready Codebases, Background Agents, and AI-Native Team. Grading moved too: the final project fell from 80% to 50%, with 30% now on open source contributions. This series reads all ten weeks.

CS146S Week 10: The Software Factory Isn't Automation — It's Handing Over the Feedback Loop

The final session is 'self-running, self-improving software systems.' The parts all appeared in the previous nine weeks: deterministic validation loops, skills that can be written back, background agents, centralized governance. One easily missed proportion from the slides — coding is 30% of engineering time, and running it in production is the other 70%.

Adversarial Robustness and Generative Models: It's Not Nonlinearity, It's Linearity

Researchers initially assumed neural networks are easy to fool because they're nonlinear. That was wrong — Goodfellow's 2014 paper argues the primary cause is their linear nature, and high dimensionality lets every tiny perturbation compound. The second half covers generative models: GANs' three pathologies, and why diffusion sidesteps two of them by adding noise and learning to remove it.

Agents, Prompts, and RAG: What's Left After the Lecture Is the Hard Part

A BCG experiment found a jagged frontier: inside it, AI substantially improved consultants' work; outside it, AI made results worse — and people fell asleep at the wheel. The lecture also takes a strong position: avoid fine-tuning wherever possible, because by the time you're done tuning, the next model already beats your fine-tuned version.

AI Project Strategy: Three Hours in a Spreadsheet Buys Back Weeks

Andrew Ng demonstrates error analysis on a deep researcher: columns are the pipeline stages, rows are 10 to 100 queries, you only look at the ones that went badly, and you mark each cell where something broke. The percentages don't have to sum to 100%. He says it takes three or four hours and saves weeks of going the wrong direction — and the fraction of people who actually do it is far below 100%.

Deep Reinforcement Learning: Putting RLHF Back Inside the RL Frame

The third reason Go can't be learned with supervision is the interesting one: the ground truth itself is ill-defined — the strongest human doesn't play their best moves every day, and even their best move isn't optimal. The last 20 minutes map RLHF fully back onto RL: the agent is the model being fine-tuned, the action is the next token, an episode is one full generation, and the reward is extremely sparse.

Full Cycle of a DL Project: You Get Two Days to Collect Data

Andrew Ng walks a face-recognition door system through the entire project lifecycle, and the whole lecture has one thesis: speed. He gives teams a two-day deadline, on the reasoning that 'time spent preparing data should be commensurate with the time it takes to train the model once.' It closes on a line: my job is to build something that actually works, and that is not the same as building something that works on the test set.

Supervised, Self-Supervised & Weakly Supervised Learning: From Comparing Pixels to Comparing Meaning

CS230's second lecture derives embeddings through three case studies: day/night classification teaches you to use humans as a proxy for choosing resolution, trigger-word detection teaches you to manufacture a million training examples in three hours, and face verification walks you through designing your first loss function. The final step — from supervised triplets to self-supervised pairs — is why modern models can consume billions of unlabeled images.

What's Going On Inside My Model? Where You Look First When It Regresses

Ask a model what a goose looks like to it and it draws a whole flock — because the labeled data tagged a flock as 'goose,' so it thinks the flock is the label. This lecture gives seven ways to open a CNN up, then says honestly: applied to transformers, even the frontier of this research only explains two layers.

Introduction to Deep Learning: The Two Moments Prompting Stops Being Enough

CS230's first lecture is a course overview, but Andrew Ng spends most of it on three things: why scaling works, when prompting stops being enough, and why he thinks 'don't learn to code' is one of the worst pieces of career advice ever given.

Scanned PDF Benchmark: How Did 10 Parsers Handle Graduate Entrance Exams?

I tested 10 open-source PDF parsing tools on four scanned NTU graduate entrance exams. VLM-based tools—Firecrawl, MinerU 3.4, and Marker v2—overwhelmingly beat conventional OCR on formulas and code, but installation was the real barrier: MinerU's old package name creates dependency hell, Marker's first model download takes 10 minutes, and PaddleOCR needs a separate engine. In practice, use RapidOCR for screening and MinerU or Firecrawl for close inspection.

Career Advice in AI: The 10x Engineer Who Failed 300 Applications

An elite-level engineer applied to over 300 jobs and failed every one, because the interview guides told him to stand his ground and have a backbone, and he read that as being hard-nosed. The interviewers' read: this is the 10x engineer, and I don't want him anywhere near my team. This lecture also gives the best criterion anyone has offered for vibe coding — technical debt.

ai deep-dive

X Open-Sources Its Algorithm: Publishing the Weights Made It Clearer Where the Black Box Is

The 2026-08-13 release added 363,246 lines and published the For You ranking weights for the first time: favorite 0.5, reply 5.0, report −234.0. But the weights are constants — the P(action) that actually decides order comes from a 2560-dim, 8-layer transformer.

Context and Memory: Where Agents Actually Fail

Chroma tested 18 frontier models and all of them degrade as input grows — as a cliff, not a slope. Memory failures are usually retrieval failures in disguise. And the real cost of KV cache is bandwidth, not storage: every generated token reads the whole cache.

Security: Prompt Injection Can Only Be Contained in the Harness

In November 2025 three frontier labs jointly broke all 12 previously proposed prompt-injection defenses. EchoLeak's payload passed Microsoft's own dedicated classifier. So the goal is not blocking every attack — it is surviving the ones that land, and that is harness work.

Drawing the Lines: Agent, Workflow, RAG, and MCP

The line between workflow and agent is who decides the steps — the developer at design time, or the model at run time. By that definition most LLM systems in production today are workflows. Plus a usable test for choosing between RAG and an agent.

Launch Is Where the Work Starts: Enterprise Agent Cases Read Sideways

Salesforce's number from 20,000 deployments: 90% of the work on an agent happens after launch, the reverse of traditional software. Stripe merges 1,300 PRs a week with no human-written code, and credits the environment rather than the model.

The Protocol Layer: MCP, A2A, ACP, Skills

MCP governs agent-to-tool, A2A governs agent-to-agent, Skills govern reusable knowledge. The test is whether the data changes: if it changes between calls you need MCP; if it's stable enough to write down, a skill file is simpler and has no runtime that can fail on its own.

The Model Is a Component, the Harness Is the System

Microsoft, OpenAI, Salesforce, Stripe and three others independently say the same thing: reliability comes from the engineering around the model. And 'give the deterministic parts back to code' has been shipped as a product four separate times — Agent Script, Procedures, runtime, blueprints.

Three Shapes of RAG and the Evaluator Paradox

Standard RAG gives a wrong answer when it retrieves the wrong chunk, and nothing in the system will notice. Agentic RAG adds a self-check, at the cost of the evaluator paradox: the ceiling on self-correction is whatever the evaluating LLM can judge about relevance.

Drones Have Closed Taiwanese Airports 23 Times, and Airports May Only 'Enforce Against'

Three Legislative Yuan budget evaluations log 23 drone-caused airport closures in Taiwan between FY2019 and August 2023, and every row's response column repeats one sentence: on notification, go to the scene with the Aviation Police and investigate. Article 99-13(6) of the Civil Aviation Act gives an airport the verb 'enforce against', not 'stop or remove' — yet the FY2024 report says every airport has already procured handheld jammers.

Power Plants and Fabs Have No Row in the Civil Aviation Act's Authority Table

Paragraphs 3 and 4 of Article 99-13 of the Civil Aviation Act say 'government agencies, schools or legal persons', but paragraph 7's proviso — the only sentence letting a site operator 'take appropriate measures to stop or remove' a drone — says only 'government agencies', and Taipower, CPC and the fabs are legal persons. Their remaining path is to have a municipality announce a restricted zone and 'enforce against' violators, which drops the penalty from an NT$300,000 minimum to an NT$300,000 maximum, while importing a jammer requires 'critical infrastructure provider' status that you learn you hold by receiving a letter.

Police May Lawfully Fly a Drone, but That Is Not Authority to Film

Eight municipal police departments now run drone units, and the authority to fly is clear — Article 99-16(2) of the Civil Aviation Act exempts agencies performing statutory duties. But that is a flight-safety exemption: a Taipei City legal opinion says it 'may be unable to serve as the legal basis for conducting administrative investigation by drone', and the special-compulsory-measures chapter passed in July 2024 authorises GPS, IMSI-catchers and private-space imaging with no aerial-investigation provision at all.

Who May Bring a Drone Down? Taiwan's Law Authorises the Outcome and No Means of Achieving It

'Drone' appears in 19 of Taiwan's central statutes and regulations; 'counter-drone' appears in exactly one, and that one is a procurement act. Six settings are authorised to 'stop or remove' an intruding drone and none of them says by what means — Article 99-13 of the Civil Aviation Act says only 'take appropriate measures', while the 31-item police instrument list includes mortars but no jammer, and Article 67(1) of the Telecommunications Management Act has no public-duty exception.

A Flight Controller's Autonomy Has 56 Entries, One About Opportunity

ArduPilot's ModeReason enum is the exhaustive list of on-board autonomous decisions: of 56 values, 43 are the aircraft deciding for itself, 21 of those because it detected something dangerous, and exactly one — SOARING_THERMAL_DETECTED — because it found an opportunity. As for end-to-end, PX4 mainline's mc_nn_control has a 10 KB tensor arena and a Kconfig default of n; ArduPilot mainline has none.

Why Drone Thermal Camera Prices Jump: The Cost Steps Export Control Draws

Thermal camera prices jump in steps because ECCN 6A003.b.4.b splits control on 111,000 focal-plane elements: 384×288 clears it by 408 pixels, 640×512 is three times over, nothing common sits between, and FLIR prices the Boson 640 at 2.04–2.31× the 320. The Note 3.b carve-out then limits not resolution but focal length — you may have a thermal camera, just not a telephoto one.

Production Ramp Evidence Isn't in the Factory, It's in Failed-to-Award Notices

The public evidence of a production ramp isn't in the factory but in the government e-procurement system: the fire service's 88 thermal drone sets were priced centrally, six fire departments' first tenders drew no qualifying bidder, all seven awards I pulled equal the budget to the dollar, and the bidder pool is about six firms. A control group added afterwards — aerial ladder trucks, at 0.54 failed-to-awarded versus the drones' 0.33 — refutes this post's original first conclusion: failed tenders are the norm in Taiwanese fire-service procurement.

There Is No Swarm in a Drone Light Show: 200 Aircraft Share One Integer

Grepping all 9,199 lines of Skybrush's open-source light-show firmware for neighbour, collision and swarm turns up nothing: no aircraft knows another exists, and the entire content two hundred of them synchronise is one GPS time-of-week integer plus a millisecond offset, with collision avoidance settled on the ground in the choreography software. Chapter 7 of Taiwan's drone cybersecurity spec — the swarm chapter — leaves the 'drone' column a dash down all twelve items and tests switches and routers instead.

Taiwan's Drone Security Spec: The Five Items That Test Resilience Are Optional

Reading the whole specification turns up three counter-intuitive things: of the seven mandatory items only three sit on the aircraft and the rest on the ground control station; an unencrypted link still passes, as long as the manual or the box states the reason or the risk — the rule asks for disclosure, not encryption; and the five items that actually test resilience (spoofing, jamming, link loss, known firmware vulnerabilities, app certification) are all in Chapter 8, which is optional.

From 40 Minutes to 6 Hours: It's the Airframe, Not the Battery

The same 5 kg aircraft on the same 2 kg pack needs 512 W to hover for 41 minutes as a multirotor, or 128 W to cruise for 164 minutes and 197 km as a fixed wing — and that range is independent of cruise speed. A VTOL's cost is not the energy burned in transition (2.4%) but that it carries both a rotor system and a wing, worth about a quarter of the range; Taiwan builds something at every rung from 40-minute electric multirotors to a 6-hour piston helicopter.

Why Drones Only Fly 30 to 45 Minutes: One Equation Gives the Answer

Hover power scales with takeoff weight to the 1.5 power while battery energy scales linearly, and that one exponent gap makes added battery hit diminishing returns fast — the optimum battery fraction solves to two-thirds of takeoff weight, independent of rotor diameter, rotor count, efficiency and energy density. What actually caps endurance is payload, not the regulatory weight thresholds: 60 minutes needs a 314 Wh/kg cell, and the best high-rate cell today is 242 Wh/kg.

Frequency Hopping Is Not Encryption, and Taiwan Caps Power by Channel Count

ExpressLRS derives its hop sequence by running the bind phrase through MD5 into a UID and feeding that to a linear congruential generator — fully reproducible, and I ported it to Python and matched the original C bit for bit — while the link itself has no encryption at all, only a 14-bit CRC. Taiwan's LP0002 then turns channel count into a power ceiling: 75 or more hopping channels at 2.4 GHz allows 1 W, fewer allows 0.125 W.

Analogue FPV Video Links Have No Clause to Walk Through in Taiwan

Analogue FPV video in Taiwan has two separate problems: LP0002's 5.8 GHz window is only 125 MHz wide while the market's standard channel table spans 300 MHz, leaving eighteen of the forty classic channels inside the window once transmit width is counted. The harder one is that §4.10.1.1 recognises only frequency-hopping transmitters at 5.8 GHz, and analogue video is fixed-frequency — it doesn't violate a clause so much as fail to match any clause in the document.

The Seven Seconds After GPS Jamming: How It Notices, and Why Detection Is Off

Flying a drone to 28 m in SITL and switching on the simulator's built-in GPS jamming, the flight controller declared EKF failure about seven seconds later, switched to LAND, landed and disarmed — it neither flew away nor crashed. Reading PX4's source then shows that of the 12 gates in EKF2_GPS_CHECK, bit 11, jamming detection, is off by default: the estimator catches jamming on its own, while spoofing can only be reported by the receiver.

PX4 or ArduPilot: The Real Fork Is the Licence

After cloning and building both, three things differ from the stereotype: ArduPilot's EKF3 header credits the derivation to PX4/ecl, so the hardest layer is shared; PX4's last year of commits comes from company domains (380 from Auterion alone) while ArduPilot's comes from personal addresses with one contributor at 37%; and what really decides the choice is not performance but BSD-3 versus GPLv3, and which layer you need to modify.

The Filings Answered What I Assumed Needed an Interview

I had filed "what's the real margin on military-grade-commercial tenders, and how long is the cash cycle" under questions requiring interviews. The published filings answer both, more precisely: gross margin runs 35–39%, normal for hardware; but operating expenses consume it, and operating income has been negative for three straight quarters while reported net income came from non-operating items. The real constraint is inventory — roughly 385 days of it, producing a cash conversion cycle near 377 days. The money in this business isn't stuck in margin, it's stuck in inventory.

The CAA Published the Entire Question Bank: What Four Exam Subjects Reveal About the Regulator

In 2022 Taiwan's Ministry of Transportation issued a press release titled 'Drone Licensing Overhaul! Obscure Questions Removed, Question Banks Fully Published,' putting all 1,420 questions online — 388 general, 588 professional, 324 renewal, 120 simplified renewal. The stated reason was candid: candidates said it was too hard. And the published bank became a policy document in its own right — the meteorology subject imports the manned-aviation syllabus wholesale, and flight principles lean heavily on fixed-wing aerodynamics, while most candidates fly multirotors.

The Drone Chapter Has No Privacy Provision, Only the Criminal Code

Taiwan's Civil Aviation Act drone chapter regulates flight safety, not the people being flown over — and that isn't my commentary, it's the Legislative Yuan's own 2020 research report: the chapter 'contains no specific provision on the privacy management that matters most to the public.' So privacy falls back on Article 315-1 of the Criminal Code and the Personal Data Protection Act, and case law shows that catches part of it: pointing a drone at a hot spring room drew an indictment, raising a phone to a window drew four months. What it doesn't catch is evidence — the report's own words: 'by the time the victim notices it, it may already have vanished.'

After "NT$2M a Year Flying Drones": Agricultural Spraying Has Run a Full Cycle

Agricultural spraying is the one drone application whose unit economics are fully transparent: the billing unit is the fen (about 970 m²), rates run NT$150–300, spraying one fen takes about two minutes, equipment costs NT$300–500k, and even the substitute's price is public (manual spraying, roughly NT$200 per fen). And precisely because anyone can run the arithmetic, everyone did — operators who entered in 2018 sprayed over a thousand hectares a year and cleared over a million; later entrants broke even after two or three years and quit. Licensed operators charge NT$300 per fen; unlicensed ones undercut to NT$150. This is the industry cycle in miniature, compressed into about six years.

Inspection Is Taiwan's Furthest-Along Drone Application — Because It Routed Around BVLOS

I assumed inspection was blocked by beyond-visual-line-of-sight rules the way logistics is. It isn't. Bridge, transmission tower, and high-speed rail viaduct inspection are all running with hard numbers: one bridge went from 8 inspectors, 4 vehicles and 2 days to 5 people, 1 vehicle and half a day with no traffic control at all, at 60% of conventional cost; high-speed rail crews covered at most 700 metres a day on foot and now save 3–5x; a private power plant cut headcount by three quarters and cost by half with no outage required. The reason is that all of these are segmented, fixed-point tasks completable within visual line of sight. What BVLOS actually blocks is continuous long-range routes, not 'inspection' as a category.

Taiwan Already Has 24 Drone Logistics Corridors — It Didn't Take the Wait-for-Regulation Route

I assumed Taiwan had no real drone logistics because it has no BVLOS framework. Wrong: the CAA has approved 10 cases across 24 flight corridors, and the Institute of Transportation has run a six-year PoC → PoS → PoB path since 2020, entering commercial validation in 2025. The way it routes around regulation is the mirror image of inspection — inspection cuts work down to within visual line of sight, logistics gets corridors approved one at a time. And its value isn't being cheaper than a boat: during the typhoon sailing suspensions at Liuqiu, a drone made the crossing in a bit over ten minutes, which is what you have when the boats don't run.

Search and Rescue Drones: The One Application Whose ROI Isn't Money — and the Easiest Budget to Cut

Agricultural spraying runs NT$150–300 per fen, computable to the decimal. Search and rescue has no such number, because a life recovered has no price. That difference determines two things: first, the specification is written by terrain rather than performance — the defining feature of Taiwan's fire agency drones is that they do NOT depend on GPS, because mountain signal drops and they must thread through trees; second, when the legislature moved to cut the budget, the ministry could only point to one man pulled from a flooded river in June. Defending a budget with an anecdote is fragile, and this application has no better weapon.

Why Countering Drones Is Hard: Jamming Is Failing, and Taiwan's Problem Isn't Only Technical

Electronic warfare has one structural limitation: it needs a signal to attack. Fiber-optic control emits no radio, satellite links bypass ground jammers, and AI terminal guidance needs no link at all for the final leg — three methods systematically defeating jamming, which is why defense is shifting toward interception. Taiwan carries an extra layer: a Control Yuan investigation documents the Ministry of National Defense revising its drone response SOP three times in two months, from "may shoot down" to "flare warning only, first shot requires the Minister's authorization" and back to "shoot down with 7.62mm or smaller." That is not a technology problem.

Taking Apart Two TTSB Crash Reports: Neither Was the Operator's Fault

Only four drone occurrences have entered Taiwan's official aviation safety statistics — because the threshold is 'substantial damage to a drone over 25 kg.' A hobbyist crash never enters the count. The two published investigation reports are the same model, the same manufacturer, and the same agency, and both probable causes were hardware failures: main rotor servo electrical failure in one, a fractured tail rotor pitch link in the other. In both, the flight control computer and the operator were explicitly cleared.

How to Read a Drone Spec Sheet: Which Lines Regulation Turned Into Boundaries

The three most important lines on a drone spec sheet are exactly the three lines spec sheets don't print. Taiwan's Drone Cybersecurity Testing Specification defines a product 'series' as units whose flight control, communications, and satellite positioning chip modules are all identical — regulation decides whether two drones are the same drone by those three modules, not by looks or endurance. And the weight field isn't a marketing number either: 250 g, 1 kg, 2 kg, 15 kg, and 25 kg are five separate legal thresholds.

ai deep-dive

What AI Certifications Engineers Can Take in 2026

Every AI certification an engineer can register for in 2026, listed one by one: AWS AIP-C01 / MLA-C01 / AIF-C01, Google PMLE, Microsoft AI-103 and AI-500, all twelve NVIDIA exams, Databricks, Snowflake, Oracle's Agentic AI track, IBM watsonx, Salesforce Agentforce, GitHub GH-300, Anthropic's four Claude exams, Taiwan's iPAS AI Application Planner, plus the governance and audit line (IAPP AIGP, ISACA AAISM / AAIA, CertNexus CAIP) — prices, validity, and registration gates all checked against vendor pages. Two things that hit your wallet: Google's PMLE exam guide has renamed Vertex AI to Gemini Enterprise Agent Platform throughout, making pre-mid-2026 study material worthless, and the iPAS intermediate certificate is valid 5 years, not permanently.

anydoc: 14 Office Formats to Markdown, Firecrawl's Rust Answer

Firecrawl's open-source Rust conversion library turns 14 office formats (including legacy .doc / .ppt / .xls) into GFM at a 4.7ms median — 109× faster than Docling under the same timing basis. The trade-off: it does no OCR at all.

The Parsing Layer: When Structure Must Be Inferred — and Licensing Is the Real Selection Axis

Scans and complex layouts leave you no choice but to infer structure with a model. But the technical gap between MinerU, Marker, and Docling is far smaller than the licensing gap — MinerU needs a separate license past $20M monthly revenue, Marker's model weights need payment past a funding threshold, and only Docling is cleanly MIT. Read the LICENSE before the benchmark.

The Three-Layer Ladder of Document Parsing: Pick the Layer Before the Tool

The most common mistake in feeding documents to an LLM isn't picking the wrong tool — it's picking the wrong layer. Structure already in the file goes to the conversion layer (milliseconds); text without structure goes to extraction; only inferred structure needs parsing. anydoc's 4.7ms against Docling's 513.6ms is a 109× gap, and most people jump straight to the most expensive layer.

The Deterministic Extraction Layer: Solve 80% of Your PDFs With No Model At All

Digital-native PDFs already contain readable text — what's missing is structure, and heuristics can recover it. PyMuPDF, pdfplumber, pypdf, and Tika do this with zero GPU and zero inference cost. The biggest selection trap isn't accuracy; it's PyMuPDF's AGPL-3.0 license.

The Drone Industry Job Map: Eleven Roles, and Which Ones a Software Person Can Actually Enter

Put drone jobs back into the five-layer supply chain and the picture clarifies: Layer 2 belongs to mechanical and electrical engineers, while Layer 3 — flight control firmware, sensor fusion, RF, edge AI — is the real entry point for software people, and also exactly the layer Taiwan is short on. But check the scale first: 267 companies and NT$12.9B of 2025 output means far fewer openings than the news volume suggests.

Four Gates into Taiwan's Drone Industry: The Entry Mechanics Public Records Can Tell You

TEDIBOA has grown from 50 founding members to over 260, but since 1 July 2025 you must first be a full member of the Defense Industry Development Association before you can even apply. R&D grants under the Industrial Innovation Platform cap out at 50% of total project cost, three concurrent cases per company, three years maximum — and explicitly ban red-supply-chain components while requiring joint cybersecurity lab testing. This piece covers only what public documents can actually establish.

From Software into Drones: Use the PX4 Architecture Diagram as a Job Map

PX4's own architecture docs say the companion computer runs Linux because 'Linux is a much better platform for general software development than NuttX; there are many more Linux developers.' That sentence is the entry point. The three transition paths differ sharply in friction — CV and MAVLink work on the companion computer transfers almost directly, flight controller firmware means learning RTOS and work-queue constraints, and estimation and control means quaternions and Kalman filters.

Four Ways to Learn Drones in Taiwan: Universities, Licences, Competitions, and Vocational Training

Taiwan has no 'drone department.' The system's core is National Formosa University's Department of Aeronautical Engineering — a NT$90M Ministry of Education training base sited inside the Chiayi drone cluster, the only one of 18 regional bases located in an industrial park, plus a NT$50M lab with a large low-speed wind tunnel. And the 2026 Presidential Cup's three challenge topics (edge AI recognition, high-precision positioning, GNSS-denied autonomous navigation) are almost a mirror of the industry's Layer 3 gap.

Following Taiwan's Drone Defense Money: Three Budgets and a Bill Stuck for Two Months

Taiwan's public drone money splits three ways: an approved NT$44.2B coordination program (R&D grants), a proposed NT$210B defense procurement special statute (stuck in cross-party negotiation), and annual agency budgets (NT$7.2B+ for 2026). The largest was written to run from 1 August 2026 — that date has passed with the bill still unresolved. And the Executive Yuan version buys only three specific items.

The Drone Supply Chain Against a Four-Criteria Framework: Only One of Four Holds

Measuring the drone sector against supply chain chokepoint / structural demand / high switching cost / long-term institutional holding: Taiwan sits at the most substitutable layer, 80% of demand comes from public budgets rather than end-user behavior, and only the certification-driven switching cost genuinely holds. The Army's NT$988M counter-drone contract — failed three times, terminated in full, NT$98.78M performance bond forfeited — is the most expensive lesson in why winning a bid is not revenue.

learning deep-dive

What Paper Still Earns in a Digital-First Study Stack: Three Places It Works, and One It Doesn't

Once studying went fully digital, paper still holds three places with real evidence: reading (paper over screens at g ≈ −0.21 across 171,055 participants, widening to 0.35–0.48 when scrolling is required), writing while you answer (on screen, harder questions draw less scratch work, not more), and drawing (45% recall against 20% for writing). The one claim most people lean on, that handwritten notes stick better, spans −0.008 to +0.248 across four meta-analyses with no consensus.

Getting a Taiwanese Drone Licence: Tiers, the No-Skipping Rule, Fees, and Timeline

The general licence is written-test only and costs NT$450 total (NT$200 test + NT$250 certificate). The professional basic tier adds a practical test and medical exam for NT$1,900. Professional tiers must be climbed in order — reaching the top bracket realistically takes 18 months to two years. Each advanced group (G1/G2/G3) costs another NT$1,200. Fee figures come from Annex 17 of the regulations; the widely circulated 'NT$500 practical test' is wrong.

Taiwan's Drone Rules in Plain Language: Registration, Licence, Penalties

Register anything 250g or heavier; the registration number expires after 2 years. Individuals only need a licence at 2kg–15kg with navigation equipment. The licence term is now 3 years, not 2, and the student licence age dropped to 14, not 16. This piece covers only the currently effective text of the Remotely Piloted Drone Management Regulations, and flags the three most common outdated claims circulating online.

Four Drone Business Models, and Why Selling Airframes Is the Worst One

Hardware runs 35–55% gross margin under permanent DJI price pressure; autonomy software and DaaS subscriptions run 60–80% and recur. Skydio's software subscriptions were already ~30% of revenue in 2023 at a 38% blended margin; India's Garuda had DaaS at 62% of FY24 revenue with a 351-day cash conversion cycle against a defense-heavy peer's 597. Taiwan is almost entirely concentrated in the lowest-margin, most substitutable cell.

BVLOS in Three Jurisdictions: Taiwan Has No Framework At All

The EU has had a workable path since the end of 2020 — the Specific category grants an operational authorisation based on risk assessment, with U-space Regulation (EU) 2021/664 rolling out on top. The US Part 108 rule was still at OIRA and unpublished as of July 2026. Taiwan doesn't have the framework at all: its regulations offer 'extended visual line of sight' (900m, 400ft, observer required), while true BVLOS runs on per-activity permits valid for three months.

Drone Industry Cycles: How the 2016 Bubble Burst, and What's Different This Time

The 2016 consumer drone bubble left specific wreckage: 3D Robotics stopped making hardware, GoPro recalled all 2,500 Karma units six weeks after launch and cut 15% of staff, Parrot cut 35% of its drone workforce, and Lily Robotics collapsed after taking $34M in pre-orders. In 2023 even Skydio — $570M raised — exited consumer, and three years later it is valued at $4.4 billion. This wave runs on a completely different engine, but three things are exactly the same.

The Drone Industry Map: Components, Regulatory Ceilings, and the Non-Chinese Supply Chain Rebuild

The global drone market is roughly US$69B in 2026 (IDTechEx). China holds about 80% of it (CSIS) and DJI over 70% of multi-rotor. The FCC put every foreign-made drone on its Covered List in December 2025; Taiwan's drone output jumped from NT$5.0B to NT$12.9B in one year, and Q1 2026 exports already beat all of 2025. This piece breaks down the five-layer supply chain, the four demand blocks, and the two ceilings holding back scale.

tech deep-dive

Your Phone Isn't Listening: What FTC Filings and Meta's Own Docs Say About Why the Ads Are So Accurate

Northeastern tested 17,260 Android apps and found zero activating the microphone. In May 2026 the FTC ruled that Cox Media Group — the company that claimed to be listening — collected no voice data at all and was reselling data-broker email lists, settling for $930,000. The real pipelines are off-site event feedback, lookalike spillover, contact-graph uploads, and location brokers.

Taiwan's Drone Supply Chain: Where the 267 Companies Are, and Which Layer They're Stuck On

Of the 267 companies the Ministry of Economic Affairs counted, 164 are in northern Taiwan. But geography is not division of labor — Thunder Tiger's published bill of materials shows motors, batteries, frames, and propellers sourced locally, while flight control, comms/GPS, and camera modules go to US, European, and Japanese partners. Exports were only 23% of 2025's NT$12.9B output, and 88.1% of export value sits in the 2–7 kg weight band per Ministry of Finance statistics.

tech deep-dive

Three Routes to Hand-Drawn SVG Icons: A ~88k Free Library, a Generator That Bends Lucide, and the License Page Nobody Reads

Koboyo claims close to 90,000 free hand-drawn SVG icons (the count oscillates: 92,967 → 87,954 → 90,150), but its sitemap only lists about 17,930 icon pages, and its license page explicitly forbids building an icon library or canvas app with them. There are actually three routes to a hand-drawn look: collect a library, bend existing geometry programmatically (sketchyicons turns every straight run in Lucide into a quadratic Bézier, seeded by icon name for byte-for-byte reproducibility), or generate with AI. This piece compares seven libraries on scale and license, unpacks the algorithms behind sketchyicons and tldraw, and surveys the icon search tools now shipping MCP servers.

AI Makes Things Smooth Exactly Where They Should Be Hard: What Generative AI Does to Learning

The single most-cited meta-analysis on ChatGPT in education (g = 0.867, ~500k views) was retracted by Nature in April 2026. But the positive finding was not overturned — the issue is that it measures performance while the AI is available. Bastani's PNAS RCT measured something else: +48% accuracy during practice with GPT-4, then 17% below never-users once access was removed.

Learning How to Learn: Auditing the Course 4.17 Million People Took — What Holds Up, What's Just a Metaphor

Dunlosky's 2013 review rated 10 study techniques; only self-testing and distributed practice earned 'high utility'. But a 2026 systematic review puts the effect at 0.22–0.46, and Pan & Rickard's transfer meta-analysis finds 'no positive transfer' once publication bias is corrected — making the premise in the framework's own name the piece that tests worst.

ai deep-dive

Digital Employees: Reliability Comes From the Harness, Not the Model

"Digital employee" isn't a technology — it's a pricing and accountability unit. Anthropic's Project Vend had Claude actually run three shops, and found the most effective intervention wasn't a smarter model but forcing it to follow procedures. Their words: "we rediscovered that bureaucracy matters." Gartner estimates only ~130 of the thousands of vendors claiming to be agentic actually are.

ai deep-dive

The Image-to-Video Landscape: Architecture, Models, and Real Prices in 2026

Every serious image-to-video model in 2026 runs latent diffusion on a DiT backbone, so visual quality is no longer a useful axis for choosing one. The real axes are native audio, self-hostability, and dollars per second. Three widely-repeated errors worth correcting: Sora's app shut down on April 26 and its API goes on September 24; Wan 2.7 is described everywhere as Apache 2.0 open weights but no first-party source has them; Veo 3.1 officially costs $0.40/s, not the $0.75/s that circulates on review sites.

ai deep-dive

The 2026 Map of 3D Modeling Tools: AI Generation, Scanning, CAD, or Manual

There are four paths to a 3D model in 2026: AI generation (Meshy-6 / Tripo / Rodin Gen-2.5 / Hunyuan 3D), phone scanning, Text-to-CAD, and manual modeling. Picking wrong has concrete costs — AI-generated meshes can't be dimensionally edited, Rodin's STL exports usually need repair, and Meshy's free-tier assets are public. This guide selects by what the model is actually for, with current pricing from each vendor's own page.

AI Web Scraping Tools Landscape: A Selection Guide for 34 Open-Source Projects

From MarkItDown (175k stars, MIT) to curl_cffi (6k stars), a survey of 34 open-source tools for feeding data to AI. Categorized along five axes: whole-site crawling, AI browser agents, document conversion, smart extraction, and anti-detection infrastructure. The key to selection isn't which tool is best — it's scenario matching.

ai deep-dive

Uncle Bob Doesn't Read His Agents' Code: What He Runs Instead of Code Review

Uncle Bob's 4.18M-view post of 2026/7/23 isn't a manifesto — it's a reply to an engineer who started in 1983 asking whether needing to understand code psychologically makes him old-fashioned. And he doesn't skip the code entirely: his 6/1 four-stage pipeline post says 'I spot check the code,' with thresholds of crap ≤ 6 (convention is 30) and mutation runs that kill all survivors. Plus a breakdown of his open-sourced Acceptance-Pipeline-Specification and the three metric blind spots Grady Booch names.

product deep-dive

Value Validation for Digital Products: From Assumption Maps to the M3 Retention Baseline for AI

The unit of validation is an assumption, not an idea. Kohavi's data shows the industry median experiment success rate is ~10%, which means roughly 22% of 'winning' experiments at p<0.05 are false positives. Sean Ellis's 40% threshold has no publicly available dataset. AI product retention should be baselined at M3 rather than M0, and GRR splits from 23% below $50/mo to 70% above $250/mo.

product deep-dive

Product Builder vs PM: What the Role Is and How to Get There

A Product Builder runs the full loop from problem discovery to design to build, alone. The core difference from a PM: PMs influence execution through authority, Product Builders influence it by directly shipping working products. LinkedIn replaced its APM program with an Associate Product Builder track, and PayFit defined the role back in 2019.

ai deep-dive

Where 3D Generative Models Stand: Reading the 2026 Technical Map Through Lyra 2.0

The dominant paradigm in 3D generation in 2026 is video diffusion feeding feed-forward 3D reconstruction, and Lyra 2.0 is the flagship of that line. But three Best Papers at CVPR 2026 point at what comes next: SAM 3D brings foundation-model-scale object reconstruction, D4RT rebuilds dynamic 4D scenes in seconds from a unified transformer, and O-Voxel replaces Gaussians with structured latents. 3DGS still rules, but surface primitives are challenging it, and pixel-space diffusion is pushing back against latent space.

tech deep-dive

How Content Platforms Rank Their Feeds: From Reddit's Formula to TikTok's Interest Graph

Take apart the feed ranking of ten platforms and they're all solving the same problem: how to trade off between newest, best, and letting new content be seen. What actually decides your answer isn't how clever the algorithm is, but whether your content is oversupplied or scarce — big platforms rank to filter content out, small communities want every post to be seen.

ai guide

Which AI Courses to Take in 2026: From AI-Curious to Vibe Coding to Production

Every official course platform from OpenAI, Anthropic, and Google, plus Stanford CS146S/CS336, Elements of AI, Hugging Face, MIT 6.S191 and more — scraped page by page, then re-sorted into four tiers: AI-curious, vibe coding, shipping to production, and how models actually work. Also covers self-study repos still being updated in 2026 and browser-based platforms that need no local setup, filtered by last-commit date rather than star count. The conclusion: nearly all of it is free. What is scarce is not courses, it is the judgment to pick one. And tier four will not fix your tier three problem.

learning deep-dive

Evergreen Books Still Trending in 2026: A Reading List Built from Threads, Dcard, and Vocus Signals

24 evergreen books across productivity, life design, brain science, psychology, and money — selected using actual discussion evidence from Threads, Dcard, PTT, and Vocus between late 2025 and mid-2026. Strongest signal: Rewire by Nicole Vignola hit top 3 on both Eslite and Books.com.tw H1 2026 bestseller charts.

tech deep-dive

Is PostgreSQL Really Enough? Don't Rush to Adopt Specialized Databases

Most teams don't need five databases. PostgreSQL's extension ecosystem covers caching, queues, full-text search, and vector search — but the real decision isn't 'can it do it' but 'where does ops cost cross performance needs.'

ai deep-dive

HyperFrames Deep Dive: HTML as Video, a Paradigm Shift for the Agent Era

HeyGen's open-source HyperFrames defines video timelines with HTML data attributes, uses headless Chrome for frame-accurate seek-and-capture, then encodes via FFmpeg to MP4. 33k stars in 3 months, Apache 2.0, 21 agent skills — AI agents write HTML to produce video, no React needed.

investing deep-dive

What Is the Mini Yuanta Taiwan 50 ETF Futures (SRF): Reading a Screenshot of 0050's Futures Version and Its Leverage Design

SRF is the Mini Yuanta Taiwan 50 ETF Futures, tracking the 0050 ETF itself. A NT$7,900 initial margin controls a contract worth roughly NT$110,000 — about 14x leverage. Dividends are handled through an equity adjustment, not the backwardation mechanism used by index futures.

investing deep-dive

Understanding a Trading Post from Scratch: Warrants, Stock Futures, and Maintenance Ratio Explained

Saw a trading post about going from NT$150k to NT$2.4M in half a year. Didn't understand a word of it — warrants, stock futures, maintenance ratio — so I looked them all up.

ai deep-dive

Loop Engineering: When AI No Longer Needs You to Write Prompts

Loop Engineering is the practice of designing systems that automatically prompt AI agents, rather than prompting them manually. Boris Cherny runs hundreds of agents, Addy Osmani coined the term, and Blake Crosley identified verification cost as the real bottleneck — this article covers primary sources, the five building blocks, applicability boundaries, and criticisms.

climbing deep-dive

A Knowledge Map of Climbing Books: 60+ Books from Training Science to Mental Philosophy

Climbing books don't exist in isolation — they represent competing schools of thought, philosophical differences, and knowledge gaps. This post maps the relationships between 60+ books to help you pick the right one to read next.

Choosing a Browser MCP: CDP, Playwright MCP, or Puppeteer MCP?

It's really a two-way choice now: @playwright/mcp (cross-browser, accessibility tree, token-cheap) versus chrome-devtools-mcp (Chrome's official server, performance and memory diagnostics). @modelcontextprotocol/server-puppeteer has been archived and is no longer a candidate. The dividing line is no longer abstraction level — it's 'drive the page' versus 'diagnose Chrome'.

Chrome DevTools MCP: The MCP Server Wired Directly to CDP

chrome-devtools-mcp, maintained by the Chrome team, packages DevTools capability as an MCP server: performance traces and insights, Lighthouse audits, heap snapshots, extension management — none of which @playwright/mcp exposes. It runs on Puppeteer, so interactions auto-wait; the costs are Chrome-only support and usage statistics reported to Google by default.

@playwright/mcp: Microsoft's Official Browser Automation MCP Server

@playwright/mcp defaults to an accessibility tree (browser_snapshot) instead of screenshots, cutting token consumption sharply. Combined with Playwright's native auto-wait it's a sensible starting point for AI agents doing web automation — but note it now runs headed by default, keeps a persistent profile by default, and gates advanced tool groups behind --caps.

@modelcontextprotocol/server-puppeteer: The Official Puppeteer MCP Server

server-puppeteer is the Puppeteer wrapper in the official MCP servers monorepo — seven lean tools built around screenshots and evaluate. It has since been archived (moved to servers-archived, no longer published), so it is not a choice for new projects; if you want Puppeteer lineage in an MCP server today, look at the Chrome team's chrome-devtools-mcp.

investing deep-dive

I Saw This 2x ETF System on Threads — It Comes From 3 Books

A 2x leveraged ETF system traces its philosophy to three books: A Random Walk Down Wall Street answers 'what to hold' (index funds), Lifecycle Investing answers 'how to accelerate' (leverage to diversify time risk), and The Four Pillars of Investing answers 'how to survive' (rebalancing discipline). Combined: 60% 2x ETF + 40% cash, Beta=1.2, ±10% rebalancing trigger.

ai deep-dive

Text / Image to Lottie: A Landscape Overview of AI Animation Generation Tools

From the CLI tool kin3o to the CVPR 2026 paper OmniLottie — a survey of open-source approaches for converting text and images into Lottie animations, with performance benchmarks and selection guidance.

tech deep-dive

AI-Powered E2E Testing: How canary, Stagehand, Magnitude, and Shortest Each Solve the Problem

AI agents running tests are non-reproducible; hand-written Playwright is hard to maintain. Four tools that emerged in 2024-2025 each tackle this dilemma with very different design philosophies.

ai deep-dive

The Skill Management Revolution for LLM Agents: A Complete Landscape of Skill Lifecycle from Voyager to MUSE-Autoskill

MUSE-Autoskill (2026) introduces a five-stage skill lifecycle framework. Self-created skills achieve 60.35% (+7.16%) on SkillsBench overall, and an impressive 87.94% on tasks where skill generation succeeds — surpassing the human-authored skill ceiling. This post synthesizes six arXiv papers to map the full landscape of skill evolution research.

design deep-dive

A Guide to Design-System Color Palettes: From Tailwind to Material 3

A comparison of seven major design systems—Tailwind, Radix, Material 3, Carbon, Ant Design, Primer, and Apple HIG—covering scale structure, neutral colors, dark-mode strategies, and why new projects should prefer OKLCH over HSL.

ai deep-dive

How to Rigorously Compare Before and After Agent Changes: From Golden Sets to Statistical Testing

Even with temperature=0, LLM outputs can still fluctuate by up to 15% in practice. To rigorously compare agent changes, you need a frozen golden set, at least 3 runs per query averaged out, LLM-as-judge blind evaluation (pairwise preference flip rate reaches 35%), and paired statistical tests -- not just running each version once and going by feel.

ai deep-dive

Agent Observability: From OTel Traces to Catching Hallucinations, Tool Misuse, and Infinite Loops

The industry has converged on using OpenTelemetry GenAI semantic conventions to turn every LLM call and tool call into a span. Detecting the three major failure modes then splits into three tracks: faithfulness + semantic entropy for hallucinations, framework-level symbolic guardrails for tool misuse, and max steps + action hash deduplication for infinite loops — all wired into a Final / Trajectory / Single-step three-layer evaluation framework.

ai deep-dive

Resource Rationality for Agents: Optimal Decisions Across Tokens, Tool Calls, and Latency

Agent decision-making under resource constraints is bounded rationality reborn: Rational Metareasoning uses VOC rewards to save 20-37% of tokens, BATS proves that adding budget without budget awareness is futile, FrugalGPT cascades cut costs by up to 98%, and Speculative Actions reduce latency by 20%. The three constraints ultimately converge into a single Pareto curve, and the overarching trend is moving from humans tuning knobs to models making resource-rational decisions on their own.

ai deep-dive

The Single Crack in Agent Security: From Prompt Injection to Trust Boundaries to Multi-Agent Worms

Three seemingly distinct agent security problems — tool output injection, trust boundaries, malicious agents — share the same root cause: LLMs flatten instructions and data into a single token stream, making them architecturally unable to distinguish between the two. Understand this through-line and you can trace every attack from EchoLeak (CVE-2025-32711, zero-click) to the Morris II AI worm, and see why 'making the model behave' doesn't work — only architectural constraints (six design patterns, CaMeL) do.

ai deep-dive

How Agents Decide Whether to Retrieve, What to Retrieve, and How to Merge: Three Decision Layers of Agentic RAG

Traditional RAG is a fixed pipeline of 'retrieve then answer.' Agentic RAG splits retrieval into three decision layers: when to retrieve (FLARE uses token probabilities; Adaptive-RAG uses a complexity classifier), what to retrieve (HyDE / RAG-Fusion / decomposition / Step-back), and how to fuse (RRF k=60 then cross-encoder rerank then compression -- Anthropic measured a -67% failure rate reduction). Key counter-intuitive insight: unnecessary retrieval hurts quality -- 'deciding not to retrieve' is a first-class capability.

ai deep-dive

Stop Hand-Tuning Prompts: From GEPA to Tool Descriptions, Automating Agent Behavior Optimization

Automatic prompt optimization (APO) has evolved from APE/OPRO to GEPA: replacing sparse rewards with linguistic reflection, winning over GRPO by ~6pp with 4-35x fewer rollouts. Meanwhile, tool descriptions are the overlooked prompt -- small wording changes can shift tool selection rates by 10x, and Anthropic's experiments show Claude self-rewriting tool descriptions outperforms human experts. These two lines are converging: eval-driven automatic optimization is eating hand-tuned prompts.

ai deep-dive

How to Build a Deep Research Agent: Multi-Turn Search Planning, Conflict Resolution, and Verifiable Conclusions

An autonomous research agent = four controllable stages: planning (decompose into sub-questions), retrieval loop (search -> read -> reflect on gaps -> search again), evidence arbitration (>=2 independent sources, typed conflict handling), and verifiable output (sentence-level citations + independent verification pass). Two approaches: training-based uses RL to learn end-to-end when to search (Search-R1 +41%); orchestration-based uses orchestrator-worker division of labor (Anthropic internal eval +90.2%, at ~15x token cost).

ai deep-dive

Machine Theory of Mind: How Agents Infer Other Agents' Intentions, Knowledge, and Goals

Inferring another's beliefs/goals/intentions from observed behavior is called Machine Theory of Mind. Three lineages: symbolic BDI, Bayesian inverse planning, and deep learning ToMnet. The biggest controversy in the LLM era is that GPT-4 still trails humans by >10 points on ToMBench — are high scores genuine reasoning or statistical shortcuts?

ai deep-dive

Multi-Agent Error Propagation and Recovery: Borrowing Thirty Years of Weapons from Distributed Systems

At 99% accuracy per step over 100 steps, the error-free completion rate drops to just 36% -- error compounding is a structural problem, not something prompt tuning can fix. Distributed systems' supervisor trees, bulkheads, circuit breakers, sagas, and durable execution can be mapped almost one-to-one into agent orchestration. But LLMs introduce a failure class that traditional systems never had -- semantic errors that don't crash -- which require Inspector agents (recovering 96.4%) and redundancy voting (MAKER: one million steps with zero errors) to address.

ai deep-dive

Semantic Similarity ≠ Retrieval Relevance: Scenarios, Detection, and Remedies for Systematic Embedding Retrieval Failures

Cosine similarity and relevance systematically diverge across an entire class of scenarios: negation (most IR models score at or below random on NevIR), exact identifiers, numeric thresholds, and logical combinations (SoTA models achieve recall@100 < 20 on LIMIT) -- some of these hit the theoretical ceiling of the single-vector paradigm, and switching to a larger model will not help. Recommended remedy order: hybrid BM25 -> reranker (Anthropic measured -67%) -> upstream metadata routing -> domain fine-tuning / multi-vector.

ai deep-dive

How to Pick the Right Tool from Hundreds: The Collapse Curve of Tool Selection and Engineering Solutions

As tools scale up, selection accuracy doesn't degrade gracefully — it collapses: 4 to 51 tools drops from 43% to 2%, 10 to 100+ drops from 78% to 13.62%. The root fix is to stop stuffing everything in at once — Anthropic's Tool Search Tool uses defer loading plus retrieval to cut 85% of tokens, pushing Opus 4.5 accuracy from 79.5% to 88.1%. Description quality has conditional payoff: negligible in simple scenarios, but correctness jumps from 44% to 50% in multi-tool chaining.

ai deep-dive

A More Expensive Embedding Won't Save Your Traditional Chinese RAG: Three Layers of Failure and the Fix Order

Traditional Chinese RAG retrieval failures are a three-layer stack: embedding granularity defects (BGE/GTE from 0.1B to 7B all mis-rank on simple queries like 'fried chicken'), Simplified Chinese / English corpus dominance causing local vocabulary drift ('premium', 'exclusion clause' alignment is unreliable), and MTEB Chinese benchmarks being Simplified Chinese making model selection signals misleading. The fix is architectural: OpenCC normalization -> hybrid + jieba segmentation -> reranker -> local fine-tuning last -- and the prerequisite for all of it is building a Traditional Chinese eval set first.

ai guide

arXiv Paper Quality Assessment Guide: From Endorsement Mechanisms to a Practical Checklist

arXiv does not perform peer review, and roughly 2% of submissions are rejected. Quality judgment relies on external signals: top venue acceptance > institution + open-source reproduction > citation quality. Includes a 20-item practical checklist and a 2026 toolbox (PWC has shut down).

tech deep-dive

Bumblebee: A Design Teardown of Perplexity's Read-Only Supply Chain Endpoint Scanner

A Go read-only scanner open-sourced by Perplexity in May 2026 (v0.1.1, zero non-stdlib dependencies). It inventories npm/PyPI/Go/RubyGems/Composer/MCP/editor and browser extensions into NDJSON, matches against a custom exposure catalog, and answers the question 'which machines in my fleet are currently affected' the moment a supply chain incident hits. It deliberately never invokes any package manager and is not an EDR.

ai deep-dive

Auto-Embedding on File Upload Is a Bad Default: A Survey of Adaptive / Agentic RAG and Agentic Parsing

Making 'chunk and embed every uploaded file automatically' the default behavior means making a decision for the LLM that it could have made itself. From Self-RAG (2310.11511) and Adaptive-RAG (2403.14403) to AgenticOCR (2602.24134), the academic trajectory is pushing three layers of decision-making -- whether to retrieve, whether to parse, and how to chunk -- from the ingestion pipeline back to the agent at conversation time.

ai deep-dive

Assembling LLM Agent Skills / Tools / Code Interpreter for Real: A Paper Reading Map

The hard part of LLM agents is not building function calling, skills, code interpreter, and document tools individually -- it is assembling them into a system that selects the right tool, writes code when needed, decomposes tasks, verifies results, and resists prompt injection. This post organizes the key papers into six engineering decisions: function calling reliability, tool/skill selection, code-as-action, multi-step planning, skill systems, and safety plus document generation.

tech debug

Mobile Chrome Redirects Back to Login After Sign-In: Debugging an HTTP-to-HTTPS Entry Point Issue

When mobile Chrome keeps redirecting back to the login page after sign-in, the culprit isn't always OAuth or broken frontend state. In this case, the root cause was that the HTTP entry point for app-dev.daodao.so wasn't issuing a 301 redirect to HTTPS, so /auth/me requests sent with an http origin didn't include the auth_token cookie.

ai deep-dive

A2UI (Agent-to-User Interface): Google's Open Protocol for Agents to Ship UI as Data

A2UI is an agent generative UI protocol open-sourced by Google on 2025-12-15: agents send declarative JSON describing UI intent, and clients render it natively using their own component catalog whitelist, layered on top of A2A. It launched at format v0.8 and iterated to v0.9 within three months.

browse.sh: Turning What Browser Agents Learn into a Skill Catalog

browse.sh, launched by Browserbase in May 2026, is two things: a browser skill catalog and the Browse CLI. The core thesis: the bottleneck for browser agents isn't reasoning — it's amnesia. By storing learned site-specific workflows as plain-text SKILL.md files, Autobrowse cut Craigslist task costs from ~$0.22 to ~$0.12 by their own metrics. Note: this has nothing to do with the 2018 Browsh text-mode browser.

ai deep-dive

CodeGraph: Local Code Knowledge Graph, and the Truth About 'Walking the Graph to Save Money'

CodeGraph uses tree-sitter to extract a codebase into a local SQLite/FTS5 knowledge graph, letting AI coding agents query the graph instead of scanning files. The official end-to-end benchmark (7 repos, median of 4 runs) averages 35% cost savings and 70% fewer tool calls -- but only if the agent actually walks the graph. Delegating exploration to a file-reading subagent that ignores CodeGraph turns it into pure overhead.

ai deep-dive

How Do People Read arXiv Papers? A Complete Guide to Methods and Tools

Reading papers is two problems stacked together: methodology (Keshav's three-pass method, 5-10 min / 1 hour / 4-5 hours) determines how to read, and tools (arXiv HTML, alphaXiv, NotebookLM, Connected Papers, Zotero) shorten the time for each pass. AI lowers the barrier to understanding; judging correctness always stays with the human.

Midscene.js: Betting on Pure Vision for Cross-Platform UI Automation

An MIT-licensed open-source UI automation framework from ByteDance. UI actions rely solely on feeding screenshots to a vision-language model, with no DOM parsing. A single JS API works across Web / Android / iOS / desktop. The trade-offs: each step is slower and more token-expensive, and everything hinges on the model's grounding ability. Note that Midscene retired MCP after 1.9.8 in favour of Skills + CLI.

Antigravity CLI: How Google Folded Gemini CLI Into a Unified Terminal Agent Harness

Antigravity CLI is a terminal agent Google announced at I/O on May 19, 2026. Written in Go (versus Gemini CLI's Node.js), its binary is called agy, and it shares the same agent harness as the desktop Antigravity 2.0. It is also Gemini CLI's successor — the personal-tier Gemini CLI service ends on June 18, 2026.

ai deep-dive

How Claude Reads and Writes PDF / DOCX / PPTX: Deconstructing the Three-Layer Architecture of Skills + Sandbox

Claude has no docx_tool or pdf_tool -- it relies on bash + file tools, plus SKILL.md instructions and pre-installed libraries like pdfplumber / python-pptx inside the container, assembling file handling capabilities from three layers.

ai deep-dive

Open Design: The Open-Source Claude Design Alternative Forked in 11 Days

Anthropic shipped Claude Design on 2026-04-17. On 4-28, nexu-io/open-design went public -- same artifact-first loop, Apache-2.0, runs on the 16 coding-agent CLIs you already have. Two weeks from 0.1 to 0.7, 40k+ stars. A paradigm shift that flattens AI design tools from vertical SaaS into a skill bundle.

ai deep-dive

system_prompts_leaks Deep Dive: What Problem Does a 40k-Star AI System Prompt Archive Solve

asgeirtj/system_prompts_leaks collects the raw system prompts of 40+ AI assistants, from GPT-5.5 and Claude Opus 4.7 to Gemini 3.1 Pro, with 40.3k stars, 461 commits, and an MIT license. The value isn't in obtaining secrets -- it's in turning vendors' implicit policies into comparable engineering material. What you should study is the design decisions, not the text itself.

ai deep-dive

Dissecting Anthropic's Founder's Playbook: Four Stages, Three Moats, and One Cowork Compliance Pitfall

Anthropic's 35-page startup handbook released 2026-05-14 reorganizes Idea/MVP/Launch/Scale around agentic AI. The most valuable takeaways are 'the easier it is to build, the more important validation becomes' and treating CLAUDE.md as the first MVP artifact. The part to discount: the Launch chapter puts compliance workstreams on Cowork -- but Anthropic's own docs say Cowork doesn't write audit logs.

tech debug

LLM Agent Tool Descriptions Determine Tool Selection: Three Bug Fixes

Rewriting tool descriptions from soft suggestions to hard rules (whitelist + consequence explanation) eliminated the LLM's incorrect tool selection; adding skip_signal=True fixed vector store double-indexing.

ai deep-dive

Using AI Agents to Operate Video Generation Tools: A HyperFrames, HeyGen, and Runway Integration Guide

AI agents can operate video generation tools through three approaches — Skills, MCP Connectors, and direct APIs. Choosing the right integration method matters more than choosing the right tool.

ai deep-dive

Code Mode: Moving Tool Definitions from Context into Code

Stop stuffing all your tool descriptions into context at session start. Let the model write code, have the runtime execute it, and let tool definitions enter context only at the import line — Anthropic's GDrive→Salesforce example dropped from ~150K tokens to 2K, and Cloudflare's 2,500-endpoint schema shrank from 1.17M to 1K.

ai deep-dive

The FDE War: Why OpenAI and Anthropic Are Both Copying Palantir's Playbook

MIT research says 95% of enterprise AI pilots yield zero return. OpenAI and Anthropic announced multi-billion-dollar joint ventures in the same week, wholesale adopting the Forward Deployed Engineer model that Palantir has used for over a decade to bring AI into the enterprise battlefield.

ai deep-dive

How Others Use LLMs to Write: Trade-off Notes from Karpathy's LLM-wiki to Multi-Agent Pipelines

A survey of 11 public LLM writing pipelines, distilled into three dominant patterns: multi-agent (researcher -> writer -> critic), Karpathy LLM-wiki (raw + wiki + LLM writes, humans don't), and quality guardrails (technical verifier + never fabricate + brief gate). The Princeton GEO paper (KDD 2024) quantifies the impact: inline citations +28%, adding statistics +33%, quoting source text +41%, keyword stuffing -9%.

ai deep-dive

OpenAI's Codex Secure Deployment Strategy: Sandboxing, Auto-review, and Enterprise Governance

In May 2026, OpenAI published its internal Codex deployment practices: sandboxes define technical boundaries, approval policies determine when to pause, Auto-review delegates approval decisions to a sub-agent instead of a human, and Managed configuration lets enterprise admins enforce policies top-down. The core philosophy: zero friction for low-risk actions, mandatory review for high-risk ones.

ai guide

9Router: A Local 3-Tier Fallback Router That Routes Claude Code / Cursor / Cline to 40+ Providers

Spin up a local OpenAI-compatible endpoint at localhost:20128 that automatically routes requests from Claude Code / Cursor / Cline / Codex / Copilot through a Subscription → Cheap → Free 3-tier fallback to 40+ providers. Built-in RTK compresses tool_result (saving 20–40% input tokens), Caveman mode compresses output, OAuth auto-refresh, multi-account round-robin — install with npm install -g 9router and two commands.

Claude, Codex, and Gemini Are All in the Browser Now: Comparing Three AI Agent Approaches in Chrome

Three vendors originally took three routes: Anthropic built an extension, OpenAI built its own browser, Google welded AI into Chrome. By August 2026 there are only two — OpenAI's Atlas stopped working on 9 August, with its capabilities folded back into the ChatGPT desktop app and Codex. The remaining split is 'live alongside Chrome' versus 'be Chrome'.

ai deep-dive

15 Walls for Building Your Own Auto-Dev Agent: Concrete Lessons from Stripe Minions

Stripe Minions says 'The walls matter more than the model,' but the case studies from four Silicon Valley companies never explained how to actually build those walls. This post breaks down the 15 walls we implemented in the daodao auto-dev agent: what each wall prevents, where the files live, and what the tradeoffs are. Tier 1 is mandatory, Tier 2 strengthens governance, Tier 3 is serious governance.

ai guide

What Is an Auto-Dev Agent? An Intro to daodao's Automated Development System

A PM checks a task card in Notion → the system syncs it to a GitHub issue → writes a plan → writes code → opens a PR for human review. This post explains what the system does, what it doesn't do, and why it's feasible now — written for people who don't write code.

ai guide

Step-by-Step: Build a Notion → PR Auto-Dev Agent — A Reproducible Version of the daodao Pipeline

Build a Notion task → GitHub issue → spec PR → code PR auto-dev agent from scratch. Using the daodao case as a template, this guide walks through every step — what to do, what to verify, and how to handle problems. Notion DB schema → bin/ scaffold → two Claude Code routines → cloud env vars → staging tests.

ai deep-dive

Claude for Financial Services: Dissecting Anthropic's Multi-Agent Reference Implementation

Anthropic open-sourced 12 financial-industry Agents and 11 MCP connectors. The real takeaway isn't the Agents themselves but the layered design of 'one prompt, two runtimes' and 'pure-file extensibility.'

ai guide

From Plan to PR: Building daodao's Auto-Dev Agent in Practice

5 rounds of consensus to write the plan, then team mode with 5 workers running 12 tasks in parallel — with plenty of pitfalls along the way. Writing it down for my future self and anyone else trying the same thing.

ai deep-dive

DeepSeek-OCR: The 10x Compression Experiment That Turns Long Context into Images

DeepSeek-OCR's paper is titled Contexts Optical Compression -- OCR is just the means; what it actually validates is that 'rendering text as images and feeding them to a VLM' achieves 10x compression at 97% accuracy. This is a qualitative shift for long-context LLM and RAG token costs.

ai deep-dive

2026 LLM Inference Provider Free Tiers & Pricing: 40+ Services Ranked by Tier

For side projects, toy demos, and RAG prototypes, nobody wants to swipe a credit card on day one. This is a verified roundup of 40+ LLM inference providers still operating as of 2026/05, tiered by whether free resources auto-replenish or are one-time grants. Each entry notes credit-card requirements, supported models, paid starting prices, and catches. Chinese-origin providers including Zhipu GLM (permanently free), Doubao (2M tokens/day), Kimi, DashScope, and the Ollama local option are all included.

Claude Code /loop: Turning AI into a Background Worker with Native Scheduling (v2.1.72+)

/loop is Claude Code's native cron feature — set schedules in plain English and let Claude monitor, auto-fix PRs, and run recurring tasks in the background. Session-scoped and expires after 7 days; for cross-session scheduling, use Routines or Desktop scheduled tasks.

Claude Code Routines: Complete Guide to Cloud Automation — Setup, Triggers, and Real-World Examples

Routines is Claude Code's cloud automation system (formerly Cloud Scheduled Tasks). Beyond cron scheduling, you can trigger runs via API endpoint or GitHub events — scan issues, review PRs, run checks, open PRs — all while your computer is off.

ai deep-dive

Claude Skills: Package Domain Knowledge into a Folder, Teach Once and It Remembers

A Skill is a folder with a SKILL.md. Three-layer progressive disclosure lets Claude load details only when needed, eliminating the need to re-explain preferences every conversation.

Local Deep Research Walkthrough: A Privacy-First Deep Research Agent

Local Deep Research is a privacy-first deep research agent built on LangChain + LangGraph, integrating 20+ search engines and 30+ research strategies. Its flagship langgraph_agent_strategy takes the LLM-autonomous tool-calling approach, offering a fundamentally different paradigm from fixed-pipeline RAG graphs.

PageIndex: RAG Without Vectors — Turning Long Documents Into a Book With a Table of Contents

PageIndex skips chunking, embedding, and vector storage entirely. Instead it relies on LLM reasoning over a tree-structured table of contents the LLM itself wrote, reporting 98.7% on FinanceBench in its own vendor-run evaluation. It solves a different problem than vector RAG — finding the right section in a well-structured long document.

tech deep-dive

Accessing Your Home Mac From Anywhere: Cloudflare Tunnel and the Alternatives in 2026

Two answers stand out for remotely accessing your home Mac in 2026: Cloudflare Tunnel if you need browser-based access with no client install, and Tailscale if you just want something simple for personal use. This post compares both, covers ZeroTier, Pangolin, NetBird, and other alternatives, and explains why Cloudflare's remotely-managed tunnel makes setup significantly easier in 2026.

Search MCP Tools for AI Agents: What to Do When WebFetch / WebSearch Gets Blocked

When using AI agents like Claude Code or Cursor, built-in WebFetch / WebSearch often gets blocked by Cloudflare, geo-restrictions, or rate limits. Connecting a search MCP server is the most direct fix. This post compares the options actually available in 2026.

ai guide

Groq Console: The Developer Platform for Running Open-Source Models on LPU Inference

Groq Console is the developer portal for Groq's in-house LPU chip, offering an OpenAI-compatible API, Playground, and free tier credits. Its selling point is running open-source models like Llama, Qwen, and DeepSeek at the fastest tokens/second on the market.

Warp: From Modern Terminal to Agentic Development Environment

Warp evolved from a Rust-powered modern terminal into an AI Agent-integrated development environment (ADE), open-sourced under AGPL in April 2026, with over 700,000 developer users.

goose: Open-Source, Cross-Platform, LLM-Agnostic Local AI Agent

goose is an open-source AI Agent maintained by the Linux Foundation's AAIF, supporting 15+ LLM providers and 70+ MCP extensions, built with Rust as a Desktop App + CLI + API. It positions itself as a vendor-neutral, self-hostable alternative to Claude Code.

Gemma on Cloudflare Workers AI: A Pragmatic Choice for Traditional Chinese Applications

For running Traditional Chinese LLM workloads on Cloudflare Workers AI, the Gemma family follows instructions more reliably than same-tier Llama models. gemma-3-12b-it was marked deprecated on 2026-05-30; the current equivalent is gemma-4-26b-a4b-it: 256K context, Vision, Function calling, at $0.10 / $0.30 per M tokens.

ai deep-dive

Knowledge Management with LLMs: From Karpathy's llm-wiki to the Open-Source Ecosystem

Karpathy proposed the llm-wiki pattern in 2026, having LLMs proactively maintain a markdown wiki instead of running RAG from scratch every time. Over 100 open-source implementations now exist, ranging from local CLI tools to serverless Telegram bots.

ai deep-dive

OpenAI Workspace Agents: From Custom GPTs to a Team Automation Platform

On 2026/4/22 OpenAI launched Workspace Agents — powered by Codex, capable of long-running cloud execution, and integrating with Slack/Salesforce/Google Drive. They are the enterprise successor to Custom GPTs.

Building a Legal Contract RAG in 36 Hours: Weaviate Query Agent + ColQwen Architecture Breakdown

Using Weaviate Query Agent + ColQwen multi-vector model, a single prompt built a production-grade legal contract search system in 36 hours -- this post breaks down its architecture logic, technology choices, and what you actually need to watch out for.

marketing guide

AKIRAXCLAW's Content Model: 5 Posts a Day, a Three-Tier Funnel, and Agent-Assisted Publishing

Akira runs a Threads → Blog → Docs three-tier funnel with agent-assisted publishing, building a sustainable knowledge monetization model in the Chinese-language AI content market.

ai guide

Where AI Code Review Stands Now: Lessons from Cloudflare's Multi-Agent System

Cloudflare ran a Multi-Agent Code Review system internally for 30 days — 131K reviews, median 3 minutes. This post breaks down their architecture and compares it with solutions from Anthropic, GitHub, CodeRabbit, Greptile, and others.

ai guide

Inside the Codex Agent Loop: How OpenAI Keeps AI Agents Iterating

A detailed look at OpenAI's Codex agent loop design: how prompts are constructed, how multi-turn conversations are managed, how prompt caching prevents cost explosions, and how context window auto-compaction works.

ai guide

Codex App Server: How OpenAI Turned an Agent Harness into a Universal Protocol

OpenAI wrapped the Codex harness as a JSON-RPC over stdio App Server, enabling VS Code, JetBrains, Web, and desktop apps to share a single agent loop. Three core primitives: Item, Turn, and Thread.

OpenAI Wrote 1 Million Lines of Code with Codex: Harness Engineering in Practice

An OpenAI internal team spent 5 months with 3 people and 0 lines of hand-written code, delivering a complete product using Codex. This article distills their core lessons on AGENTS.md design, repo-local knowledge bases, architecture enforcement, and entropy management.

AEO / GEO Tool Landscape: Input, Traffic, and Output Layers — From isitagentready to aeo-radar to Profound

AEO/GEO tools aren't a single category — they span three distinct layers: the input layer (is your website ready for AI to read), the traffic layer (how much are AI bots actually crawling), and the output layer (how is your brand mentioned in AI answers). This post maps out all three layers, from open-source self-hosted options to commercial SaaS.

tech project

DeerFlow: ByteDance's Open-Source Super Agent Harness for Long-Running Research Tasks

DeerFlow is ByteDance's open-source Super Agent Harness built on Python 3.12 + LangGraph. It orchestrates long-running tasks through sandboxes, long-term memory, sub-agents, skills, and a messaging gateway. It hit #1 on GitHub Trending in February 2026, now surpassing 63,000 stars, with support for Telegram/Slack/Feishu, Claude Code integration, and multiple search backends.

travel guide

2026 Travel Inconvenience Insurance Guide: New Rules, Coverage Comparison, and Where to Buy

2026/4/1 new rules: max 2 policies per trip (different insurers), flat-rate payout cap lowered to NT$6,000. Covers six key areas including flight delays and lost luggage, with a breakdown of where to buy.

Agentic Engineering: Making AI Agents Collaborate Like a Real Engineering Team

Agentic Engineering isn't about making AI write code faster — it's about making software move through the entire delivery pipeline faster, by using multi-agent collaboration to compress cross-team coordination friction.

The Memory Problem in Agentic Engineering: Types, Implementation, and Ownership

Agent memory isn't a plugin — it's part of the harness itself. Pick the right memory type, estimate data volume, then decide on the technology. And finally, figure out whether you actually own that memory.

ai guide

Multi-Engine Code Review with Codex + Gemini + Claude: Principles, Patterns, and Implementation

AI models rationalize their own code when reviewing it. Using three different CLIs for independent review effectively catches blind spots -- this post covers the design philosophy and practical workflow patterns behind the approach.

tech guide

How Does the YouTube to NotebookLM Extension Work? Reverse Engineering and Cross-Tab Architecture Dissected

NotebookLM has no official API. This extension works by combining three techniques: reverse-engineered Google batchexecute RPC calls, DOM scraping, and cross-tab message passing.

tech debug

Local AI Backend API Always Returns Empty Data: Cookie Domain Isolation

The main backend runs on a remote HTTPS server, so the auth_token cookie is scoped to that domain. The browser never sends it to the local AI backend, causing the API to treat every request as unauthenticated.

ai guide

Integrating AI Agents into Your Development Workflow: A Five-Phase SDLC Breakdown

Agentic AI is not just autocomplete — it is an AI system capable of autonomously executing multi-step tasks. This article breaks down the five phases of the SDLC, explaining where to plug in agents at each phase, how to progress from CLI tools to full-pipeline automation, and the most valuable external resources to track right now.

ai guide

A Book Written by AI Itself, Teaching You How to Build Software with AI

Encyclopedia of Agentic Coding Patterns catalogues 190 patterns to help you make the right software decisions in the age of AI-written code — and the book itself is autonomously written and maintained by an AI agent.

ai guide

GitHub Copilot Coding Agent: Assign an Issue to AI and Let It Open the PR

GitHub Copilot Coding Agent lets you assign an Issue to Copilot, which then automatically creates a branch, writes code, runs CI, and opens a PR — all inside a cloud sandbox. The key to success is setting up AGENTS.md; without it, the agent tends to go off track. Best suited for well-defined medium-sized tasks; requires Pro+ (1,500 premium requests/month) or Enterprise plan.

ai guide

knowledge-pipeline: A Six-Layer Pipeline for RAG Quality Control

A six-layer deterministic pipeline that handles everything from URL ingestion to vector embedding automatically, filtering out garbage before it enters your RAG system through an eight-dimension scoring system.

MarkItDown: Convert Any File to Markdown Before Feeding It to an LLM

A lightweight open-source tool from Microsoft that converts PDF, Office, images, audio, and more into Markdown — purpose-built for LLM pipelines.

ai guide

MCP vs CLI vs API: The Real Boundaries of Agent Tool Interfaces

MCP is not going away, but its effective scope is narrower than most people think. For local development, CLI and raw API almost always beat MCP. MCP's truly irreplaceable niche is the narrow gap of 'cross-agent shared local tool layer.'

Is Your JSON-LD Invisible to AI Search Engines? A Pipeline Breakdown and AEO/GEO Strategy

Different AI engines process web pages in vastly different ways. Some only read the body; others rely on pre-built indexes. JSON-LD and schema markup are not universally effective — body content quality and structure are the only cross-platform foundations that hold.

product project

quidproquo Blog Improvement Roadmap: Content, Technical Debt, RAG Design, and Harness Infrastructure

Using my own 30+ RAG/Agent posts to audit the blog itself, I identified a prioritized improvement list spanning content quality, site tech, RAG design fixes, harness infrastructure, and AI agent applications — no phases, just priorities.

ai guide

Lessons from the Trenches: What AI Native Teams Must Get Right

Not everyone should use a coding agent to modify code directly. AI Native teams need interface specs, test-first development, monorepo, security guardrails, human-in-the-loop, and token budget controls. Building an agent platform layer on top of coding agents and clearly redefining developer roles is the right path forward.

ai guide

Autoreason: Teaching LLMs When to Stop Self-Refining

Autoreason replaces the traditional critique-and-revise loop with a competitive multi-version evaluation mechanism (A/B/AB + blind Borda count), solving three structural problems in LLM self-refinement: prompt bias, scope creep, and lack of restraint.

ai project

Vercel Open Agents: Moving the Coding Agent from Your Laptop to the Cloud

An open-source coding agent reference implementation from Vercel Labs. A three-layer architecture separates the web UI, agent workflow, and sandbox VM — designed as a starting point for teams that want to self-host their own Claude Code or Cursor Background Agent.

The Full Picture of Cloudflare Workers AI Binding: It's More Than Just run()

env.AI is not just run(). It also exposes toMarkdown (document-to-Markdown conversion), autorag (managed RAG), gateway (external provider proxy), and models (metadata lookup). Understanding these four method groups is what unlocks Cloudflare as a full AI platform inside Workers.

ai guide

Claude Octopus: The Consensus Plugin That Hooks 8 Models Into Claude Code Simultaneously

Claude Octopus is a Claude Code plugin that simultaneously calls Codex, Gemini, Copilot, Qwen, Ollama, Perplexity, OpenRouter, and Claude to review the same code, using a 75% consensus threshold to catch single-model blind spots. It ships with 32 personas, 48 /octo:* slash commands, 51 skills, and a Dark Factory fully autonomous spec-to-code pipeline.

ai guide

LLM Council: Karpathy's Weekend Multi-Model Parliament — Three Stages of LLM Peer Review

LLM Council is a local Web App Andrej Karpathy built over a weekend. It sends one question to multiple LLMs simultaneously, has them anonymously peer-review each other, and then a Chairman model synthesizes a final answer. Positioned as a small tool for comparing models while studying — 99% vibe coded with no plans for long-term maintenance — but the architecture itself is a minimal ensemble LLM implementation worth studying.

tech guide

Better Agent Terminal: Consolidate Multiple Project Terminals and Claude Code Agents into One Window

Better Agent Terminal (BAT) is an Electron desktop app that unifies multiple project workspaces, terminals, and Claude Code Agents into a single window — solving the everyday pain of exploding iTerm tabs and the lack of a proper GUI container for agents. MIT License, available on macOS, Windows, and Linux.

ai guide

Claude Managed Agents: Letting Anthropic Handle the Agent Shell and Sandbox

Claude Managed Agents is a beta service launched by Anthropic on 2026/04/08 that provides an agent harness plus cloud container sandbox, billed per token plus $0.08/session-hour. It suits long-running async tasks and is worth exploring if you don't want to build your own agent loop and sandbox.

ai guide

Agent Skills: A Skill Framework That Makes AI Agents Work Like Senior Engineers

Agent Skills is Addy Osmani's open-source collection of 19 production-grade engineering skills that drive AI agents to follow senior engineering discipline through /spec → /plan → /build → /test → /review → /ship commands, instead of cutting corners.

ai guide

Graphify: Turn Code and Documents into a Queryable Knowledge Graph

Graphify uses tree-sitter AST to extract code structure, then applies LLM semantic analysis to documents and images, compressing an entire project into a queryable knowledge graph. It claims to save 71.5x tokens per query compared to reading raw files.

Claw Code: An Open-Source CLI Agent That Rewrites Claude Code in Rust

Claw Code is a from-scratch Rust rewrite of the Claude Code CLI, featuring 48K lines of code, 40 tools, and MIT licensing. Most remarkably, the entire project was built by multiple AI agents collaborating over just 5 days, surpassing 170K GitHub stars within a week of launch.

ai guide

clawhip: An Event Notification Router That Keeps Multi-Agent Development Under Control

clawhip is a Rust daemon that routes AI coding agent events (commits, PRs, session status) to Discord / Slack, solving the observability problem of not knowing who is doing what when multiple agents run in parallel.

ai guide

notebooklm-py: An Unofficial Python API for Google NotebookLM

notebooklm-py reverse-engineers Google's batchexecute RPC protocol, letting you programmatically control NotebookLM via Python / CLI / AI Agent — including audio, video, slides, quiz generation and more.

ai guide

oh-my-claudecode: An Enhancement Layer That Turns Claude Code into a Multi-Agent Collaboration Platform

oh-my-claudecode (OMC) adds 8 collaboration modes, 19 specialized agents, and cross-model orchestration (Claude + Codex + Gemini) on top of Claude Code, transforming a single-user CLI tool into a multi-agent development platform. Features include Deep Interview for requirement clarification, Smart Model Routing that saves 30-50% on tokens, and automatic rate limit recovery.

ai guide

oh-my-codex: A Structured Workflow Enhancement Layer on Top of OpenAI Codex CLI

oh-my-codex (OMX) doesn't replace Codex CLI — it adds a structured workflow layer on top of it. From requirements clarification and plan generation to multi-agent parallel execution, four core Skills transform scattered prompt conversations into a trackable development process.

ai guide

oh-my-openagent: A Multi-Model Agent Team Framework That Replaces Single-LLM Coding

oh-my-openagent (OmO) transforms OpenCode from a single-LLM tool into a multi-model agent team — Opus as the workhorse, GPT-5.2 as the architect, Gemini for frontend, Sonnet for documentation lookup — all triggered to run in parallel with a single ultrawork keyword. With 48K stars, it is the earliest project in the UltraWorkers ecosystem to establish the multi-agent coding pattern.

ai project

OpenHarness: A Fully Open-Source Agent Harness Framework

An open-source Agent Harness framework from HKUDS (HKU Data Science Lab) that implements tool calling, skill loading, memory, permissions, and multi-agent collaboration as complete infrastructure, supporting Anthropic / OpenAI / GitHub Copilot API formats.

tech guide

Solving Duplicate Config Files for Codex and Claude Code with a Symlink

Claude Code only reads CLAUDE.md; Codex only reads AGENTS.md. Teams using both end up maintaining two identical files. Fix: make CLAUDE.md a symlink pointing to AGENTS.md — one source of truth.

ai guide

How to Use Claude Code Agent Teams? Design Patterns from 6,400+ Agents on GitHub

There are already 6,400+ .claude/agents/*.md files on GitHub. We dissected 4 representative projects — ChemistryTimes (content production pipeline), claude-sub-agent (document-driven development pipeline), agentic (Temporal.io DAG parallel execution), and vs-copilot-multi-agent (hook-enforced memory persistence) — plus ruflo's enterprise-grade swarm architecture, distilling 6 design patterns and 5 practical trends.

From Stripe to Meta: How Silicon Valley's Top Companies Replace Keyboards with AI Agents

Top Silicon Valley companies are independently building internal AI coding agents that automate everything from a Slack message to a merged PR. This article deep-dives into architectures from Stripe, Ramp, Coinbase, and Spotify, then expands to cover Google, Meta, Amazon, Uber, Goldman Sachs, Walmart, and more.

ai guide

Three Modes of LLM Knowledge Bases: Knowledge Vault, Experience Vault, and Blog

Andrej Karpathy proposed a framework for compiling personal knowledge wikis with LLMs — collect raw data, have the LLM compile it into .md wiki pages, run Q&A against the wiki, and file outputs back. This post compares three practical approaches: Karpathy's knowledge vault model, the community's experience vault model, and quidproquo's blog model.

ai guide

AI Agent Caching Goes Beyond One Layer: From Claude Code's 18 Cache Types to Multi-Layer ReAct Agent Design

After dissecting Claude Code's 18+ caching mechanisms, I found that you can't touch provider-level prompt cache, but embedding cache, tool result cache, and entity cache are not only within your reach — they deliver even better results. Includes a complete AgentCache interface design and per-tool TTL strategy.

ai guide

AI Agent Tool Descriptions Shouldn't Be Static: Dynamic prompt() Design Learned from Claude Code

Every one of Claude Code's 45 tools uses a prompt() method that dynamically adjusts based on user type, feature flags, and system capabilities. Applying this pattern to a ReAct Agent, tool descriptions are dynamically generated along three dimensions: orchestrator model capability, locale, and available tools. Small models automatically get few-shot examples; large models save tokens.

Claude Code Complete Breakdown: The Deep Reasoning King of Terminal Agents

Claude Code runs from $20/mo Pro to $200/mo Max 20x. Quota is a rolling five-hour window with weekly limits on top, shared across Claude on web, desktop, mobile, and the terminal. When you run out you can switch to usage credits at standard API rates rather than stopping.

Cursor CLI Complete Analysis: The All-Rounder Extending IDE Agent to the Terminal

Cursor CLI brings the IDE agent to the terminal with an interactive TUI and headless mode, Plan/Ask/Agent modes, Cloud Handoff, and CI/CD integration. Billing now runs on two separate usage pools: Cursor's own models (Grok 4.6/4.5, Composer 2.5) and third-party models (Pro includes $20, Pro+ $70, Ultra $400).

Google's Terminal Agent Plans: The Free Individual Path Is Gone

The paying paths on Google's side: the individual free tier and Gemini CLI access on Google AI Pro / Ultra ended 2026/6/18, leaving individuals with Antigravity CLI or their own paid API key; enterprise licenses and Google Cloud are unaffected. The zero-cost starting option now belongs to someone else.

Kiro (AWS) Complete Analysis: The Spec-Driven Agentic IDE

Kiro has five tiers: Free 50 credits, Pro $20/1,000, Pro+ $40/2,000, Pro Max $100/5,000, and Power $200/10,000, with add-on credits at $0.04. Auto mode mixes models to cut cost (the same task costs 1.3x credits via Sonnet), and the spec-driven flow turns vibe coding into traceable, structured development.

OpenAI Codex Complete Plan Analysis: Agent Integration in the ChatGPT Ecosystem

Codex rides your ChatGPT subscription (Free / Go $8 / Plus $20 / Pro 5x $100 / Pro 20x $200), and since 2026/4/2 billing is token-based credits. The model line is GPT-5.6 Sol / Terra / Luna; GPT-5.4 and 5.4 mini retire from ChatGPT-signed-in Codex on 2026/8/31.

OpenCode Full Analysis: An Open-Source Terminal Agent Supporting 75+ Model Providers

OpenCode is a free, open-source TypeScript CLI agent (MIT, ~198K GitHub stars). It supports 75+ model providers including local Ollama, allows authentication via Copilot/ChatGPT accounts, and lets you switch models mid-session without losing context. There is also a desktop app and an official Zen gateway.

Agent CLI Subscription Plans Compared: Building a Flexible Multi-Model Routing Strategy

A comparison of six agent CLI subscriptions (Claude Code, Cursor CLI, Codex, Kiro, Antigravity/Gemini CLI, OpenCode) plus the multi-model routing pattern — cheap models for simple work, strong models for hard work. Nearly every one of these changed its billing in the first half of 2026; this version was re-verified on 8/18.

ai guide

2026 Personal AI Hardware Buying Guide: DGX Spark, Mac Studio, MSI AI Edge Compared

Comparing the NVIDIA DGX Spark, Apple Mac Studio M4 Ultra, ASUS Ascent GX10, MSI AI Edge, and more — helping you find the right local inference hardware.

Multi-Model Routing Open-Source Tools & Implementation: Getting the Right Model for the Right Job

With multi-model routing, 70% of simple tasks are directed to cheap models, and only 10-15% of complex tasks use flagship models — saving 40-85% on inference costs in practice. This article covers the architecture and implementation of five major open-source tools.

product project

Digital Ecosystem Research: Dissecting Platform Integration Strategies from LINE and Shopify to Taiwan MarTech

A breakdown of the three-layer digital ecosystem structure: LINE's super-app, Shopify App Store flywheel, and Taiwan MarTech integration strategies. The core mechanism is using APIs and data flows to create mutual dependency among participants, collectively reinforcing the moat.

tech guide

Where Should AI Agent Global Skills Live? The Division of Labor Between .claude, Codex Skills, and AGENTS.md

Skill paths are almost always runtime-specific. AGENTS.md is the reliable way to share rules across agents. Put personal reusable capabilities in each agent's supported global directory; put project workflows inside the repo.

tech guide

code-review-graph: Using a Knowledge Graph to Cut AI Code Review Token Usage by 8x

code-review-graph uses Tree-sitter to parse your codebase and build a persistent knowledge graph, tracks the blast radius of changes, and feeds only truly relevant context to the AI — claiming an average 8.2x reduction in token usage.

tech guide

GitBook: A Documentation Platform That Turns Docs into a Product

GitBook is a Git-based documentation platform with Markdown editing, version control, and multi-user collaboration. Ideal for technical docs, API references, and internal knowledge bases. The free plan is sufficient for individuals and small teams.

tech guide

NVIDIA DGX Spark: A Desktop AI Supercomputer That Fits a Petaflop on Your Desk

The NVIDIA DGX Spark is powered by the GB10 Grace Blackwell Superchip, 128 GB of unified memory, and delivers 1 petaFLOP of FP4 compute — starting at around $3,999 USD. It lets developers run 200B-parameter models locally and fine-tune 70B models, making it the most accessible NVIDIA AI development platform available today.

tech guide

Documentation Platform Guide: GitBook, Docusaurus, Mintlify, and Seven Other Options

A breakdown of nine major documentation platforms — their positioning, pros, cons, and ideal use cases. Decision logic: open-source projects → Docusaurus/VitePress, API docs → Mintlify/ReadMe, internal enterprise → Confluence, fastest to launch → GitBook.

The Complete Guide to Agent CLIs: Design Logic, Tool Comparison, and Best Practices

Agent CLIs are not smarter autocomplete tools -- they are AI agents that can read your codebase, execute multi-step tasks, and operate in real environments. Claude Code, Codex CLI, Gemini CLI, OpenCode, Aider, Pi, Kiro, Amp, Cursor CLI... the tools keep multiplying, but they all share a common set of design principles -- understanding these principles is how you actually get good at using them.

ai guide

15 Agent Frameworks Worth Watching in 2026

Sorted by GitHub Stars, a survey of 15 mainstream AI Agent frameworks in 2026 — their positioning, key features, and ideal use cases. Not a ranking — it's a map.

ai guide

One Sentence to an IG Carousel — From 3 Hours Manual Work to a Fully Automated Pipeline

Use Claude Code as an orchestrator to chain Playwright screenshots, catbox.moe image hosting, Meta Graph API publishing, and Telegram notifications — generate and publish an IG carousel from a single sentence.

ai guide

llama.cpp — From Pure C++ to an LLM Inference Engine on Consumer Hardware

llama.cpp is the most widely used local LLM inference engine, implemented in pure C/C++. It supports CPU, Metal, CUDA, Vulkan, and other backends, and uses the GGUF quantization format to run multi-billion-parameter models on consumer hardware.

ai guide

TurboQuant+ — Two-Stage Quantization to Compress KV Cache to 2-bit, Running 100B Models on a MacBook

TurboQuant+ is an open-source implementation of a Google Research ICLR 2026 paper that uses PolarQuant + QJL two-stage quantization to compress the KV cache by 3.8-6.4x, enabling consumer hardware to run larger models with longer contexts.

ai guide

Small Models That Run on Phones: Choices and Constraints in 2026

The main on-device LLMs in 2026 are Gemma 3n, Qwen 3.5 Small, Llama 3.2, Phi-4-mini, Ministral 3, and SmolLM3. Sub-3B quantized models can hit 30-50 tokens/sec on phones with 8GB RAM, but RAM, thermal throttling, and context window remain hard constraints.

ai project

2026 Q1 Open-Source LLM Landscape: From Frontier Models to On-Device, a Complete Survey

2026 Q1 saw a full-blown open-source model explosion: on the LLM front, GLM-5, Kimi K2.5, and Qwen3.5 caught up with closed-source models; Embedding and Reranker are dominated by Qwen3 and BGE; speech has Voxtral TTS and Whisper V3; image has FLUX.2; and video has Wan 2.2 rivaling Sora. This is the complete navigation map.

Claude Code: A Complete Guide to Anthropic's Terminal AI Coding Agent

Claude Code is Anthropic's agentic coding tool that runs in the terminal, IDEs, Slack, GitHub, and on the web. Its core extension system has six layers: CLAUDE.md (persistent context), Skills (on-demand workflows), Hooks (deterministic automation), Subagents (isolated delegation), MCP (external tool connections), and Agent Teams (multi-agent collaboration).

Codex CLI: A Complete Guide to OpenAI's Open-Source Terminal Coding Agent

Codex CLI is OpenAI's open source terminal coding agent (Rust, Apache-2.0, ~106.6k stars) with MCP, subagents, image input, code review, and Skills. The model line is now GPT-5.6 Sol / Terra / Luna, and the desktop app, CLI, and IDE extension share one config.toml.

Gemini CLI: Once the Most Generous Free Terminal Agent, Now Enterprise-Only

Gemini CLI is Google's open source terminal AI agent (Apache 2.0, ~106.6k stars). It once offered 60 requests per minute and 1,000 per day for free, with a 1M context window. The individual tier stopped serving on 2026/6/18 and Antigravity CLI took over. The project isn't shut down — the repo is still maintained — but it now serves only Gemini Code Assist Standard/Enterprise licenses and paid API keys.

OpenCode: A Complete Guide to the Open-Source AI Terminal Coding Agent

OpenCode is an open-source AI coding agent written in TypeScript (MIT, ~198K GitHub stars, repo at anomalyco/opencode) with a built-in TUI, 75+ LLM providers, LSP integration, a Vim-style editor, SQLite session management, and a desktop app. Free, no subscription, local or cloud models.

Pi Coding Agent: A Minimalist Open-Source Terminal Coding Harness

Pi is a minimalist coding agent by Mario Zechner (TypeScript, MIT, ~93K stars) with just 4 core tools and a very short system prompt — everything else you add yourself via Extensions, Skills, and Prompt Templates. It deliberately omits MCP, sub-agents, plan mode, and permission popups. The repo is now earendil-works/pi and the npm scope is @earendil-works.

AI-Ready Content: The Complete Guide to Making Your Website an AI-Readable Data Source

In 2025-2026, websites need to be readable not just by humans but by AI. From llms.txt and Schema Markup to GEO and RAG ingestion pipelines, this post maps out the complete technical landscape for turning your website into an AI-consumable data source.

Advanced Harness Engineering Patterns: Tool Registry, Guard System, and Checkpoint-Resume

A Harness is more than just an LLM wrapper. Tool Registry manages dynamic tool loading and selection, Guard System establishes a four-layer defense network, and Checkpoint-Resume enables long-running tasks to survive interruptions. These three patterns form the critical infrastructure of production-grade Agent systems.

ai guide

Skill vs Subagent: Comparing Two Agent Collaboration Modes in Claude Code

A Skill is a prompt template you invoke manually. A Subagent is an independent agent that Claude routes to automatically. They look similar, but differ completely in trigger mechanism, tool isolation, and context management.

ai guide

Ticketing Is Dead — Review Is the New Planning

When AI agents can turn intent into a PR in minutes, the bottleneck in software engineering flips from 'planning what to do' to 'evaluating whether the output is correct.' Artifacts of the ticketing era — sprints, story points, backlog grooming — are collapsing to zero, replaced by review as the core practice.

Claude Code Spinner Verbs: The Complete List of 185 Status Verbs Extracted from Source Code

When processing requests, Claude Code randomly displays one of 185 built-in verbs (like Thinking, Brewing, Clauding), then picks one of 8 completion verbs with elapsed time. You can customize these via spinnerVerbs in settings.json, using either replace or append mode. All data in this post is verified directly from cli.js source code.

tech guide

gstack — Garry Tan's 20 Skills That Turn Claude Code into a Virtual Engineering Team

gstack is Garry Tan's open-source Claude Code skills toolkit. Its 20 specialized skills transform a solo developer into an entire engineering team — automating everything from product planning and design review to code review, QA, and deployment.

Anthropic's Harness Design: Making AI Agents Work Like Engineers

The same model produces dramatically different results under different harness designs. Anthropic uses a dual-agent architecture, cross-session state files, and a GAN-inspired generator-evaluator loop to let Claude autonomously complete hours-long software development tasks.

ai guide

Google's Eight Multi-Agent Design Patterns

Google outlined eight multi-agent design patterns: from the simplest Sequential Pipeline to the composable Composite Pattern. More complexity isn't always better — picking the right pattern matters more than stacking agents.

From Prompt to Harness: The Three Evolutions of AI Engineering

AI engineering has gone through three phases: Prompt Engineering (write better instructions) → Context Engineering (feed the right information) → Harness Engineering (design the entire working environment). Each evolution doesn't replace the previous one — it operates at a higher level of abstraction.

The OpenClaw Agent Loop: Serialization, Writer Claims, and the Fence That Stops a Stale Turn From Committing

The agent loop is a serialized per-session run. The part worth studying is how it handles concurrency: an admitted run records an activeWriterRunId claim, every transcript write supplies expectedWriterRunId, and the commit transaction verifies the match — so a superseded run cannot commit stale data.

OpenClaw Agent Runtime: The System Prompt Is Assembled, and a Cache Boundary Cuts It in Half

OpenClaw builds its own system prompt for every run; there is no runtime default prompt. What it builds is split by an internal cache boundary — the stable workspace prefix above, the per-turn channel context below — so backends with prefix caches can reuse the same prefix across channels.

OpenClaw Access Control: SecretRef Is Not Process Isolation — Here's What It Actually Solves

SecretRefs keep credentials out of plaintext config, and the model-call chain sees process-local sentinels instead of the real value. But the docs say it plainly: this is not process isolation — the real value still exists in the same process's memory, and any plaintext file the agent can read bypasses the whole mechanism.

OpenClaw Automation, Part 1: Choosing Among Six Mechanisms, and Why 'Exactly on Time' and 'Check on It' Are Different Jobs

Cron is now called Automations (openclaw cron remains an alias), and automation spans six mechanisms. The core trade-off is one line: Automations give you exact timing and isolated execution, Heartbeat gives you full main-session context on a roughly-every-30-minutes cadence.

OpenClaw Automation, Part 2: Standing Orders Are the Authorization, Automations Are the Clock

Standing orders grant an agent permanent operating authority for a defined program, written into AGENTS.md and injected into every session. They define what it may do; automations define when — and the automation prompt should reference the standing order rather than duplicate it.

OpenClaw Enterprise Channels: Slack's Three Transports, and the 'Built-in' Column That No Longer Exists

Every enterprise channel is a plugin now, including Slack and Google Chat, which used to be built in. Slack has three transports — Socket Mode, HTTP Request URLs, and relay — and the docs say plainly that the first two have reached feature parity, so you pick by deployment shape, not by features.

OpenClaw's Main Channels: Where WhatsApp, Telegram, and Discord Each Get Stuck

Each channel has one gotcha that stops you cold: WhatsApp's login is QR-only and hard to do remotely, Telegram bots ship with Privacy Mode on so they never see group messages (and you must remove and re-add the bot after changing it), and Discord needs Message Content Intent or it receives nothing from servers.

OpenClaw's Other Channels: Signal, iMessage, LINE — and Reef, Where Two People's Agents Talk Directly

The most interesting entry here is Reef — an end-to-end-encrypted side channel between OpenClaw agents owned by different people. Messages are sealed on your machine, screened in both directions by a pinned-model guard, and the relay operator can never read the content. It ships bundled.

OpenClaw Channels Overview: 31 Channels, Nearly All Plugins — and Why 'Who Can Trigger' Is Not 'What the Model Sees'

OpenClaw supports 31 chat channels, but only WebChat lives in core — even Slack and WhatsApp are plugins you install. And group safety has two independent axes: allowlists govern who can trigger the agent, not which quotes and history the model sees. That second one is contextVisibility, and it defaults to wide open.

OpenClaw Gateway, Part 1: Strict Validation Will Refuse to Boot — and the Guards That Stop You From Yourself

OpenClaw validates config strictly — one unknown key, a wrong type, or an invalid value and the Gateway refuses to start. It keeps a last-known-good copy, but neither startup nor hot reload restores it automatically; only doctor --fix does.

OpenClaw Gateway, Part 2: Binding, Auth, and That Credential Precedence Contract

The Gateway binds to loopback by default, and binding anywhere else requires auth — that is enforced, not advised. Inside a detected container the effective default is auto, unless Tailscale serve/funnel is active, which always forces loopback.

OpenClaw Installation Guide (Part 2): Four Decisions for Cloud Deployment, and the Real Traps on K8s

Deploying OpenClaw to the cloud comes down to four decisions: where the Gateway binds, where state lives, who can reach it, and how you recover. Which platform you pick is the least important of them.

OpenClaw Installation Guide (Part 1): Choosing Among Six Local Methods, and Where It Gets Stuck

OpenClaw has six local install methods, and what separates them is not the command but whether you want reproducibility, isolation, or self-updating. The real blocker is package-manager lifecycle-script policy: both npm 12 and global pnpm installs block OpenClaw's build scripts by default.

OpenClaw Models, Advanced: Two-Stage Failover, the Real Cooldown Numbers, and Prompt Caching

OpenClaw's failover runs in two stages: rotate auth profiles within the provider, then fall back to another model. But what really governs behavior is who chose the model — a model you picked yourself with /model is strict, and its failure is reported rather than answered by some other model.

OpenClaw's Model Requirements and Provider Ecosystem: Provider, Model, and Runtime Are Three Different Things

OpenClaw's hard requirement for a model is tool use plus a large enough context — onboarding only auto-suggests a local model when it confirms tool support and at least a 16K context window. The easier thing to get wrong is that provider, model, and agent runtime are three separate layers: an `openai/*` ref does not mean Codex.

OpenClaw's 60 Providers: A Category Map, and What Actually Bites When You Attach a Local Model

The official provider directory now lists 60 entries. The most common failure when attaching a local model is writing Ollama's base URL with /v1 — that breaks tool calling, and the model starts emitting raw tool-call JSON as plain text.

OpenClaw Multi-Agent: An Agent Is a Whole Persona Boundary — and Agents Can Now Ask for New Agents

An agent is a complete persona scope — its own workspace, auth profiles, model registry, and session store. But the isolation is not absolute: when a secondary agent's OAuth credential expires, OpenClaw reads through to the main agent's profile of the same id, and a workspace is only a default working directory, not a hard sandbox.

OpenClaw Nodes in Depth: Approval Binds the Plan, Not the Command You Edited Afterward

The best part of remote node execution is how approval binds: exec prepares a canonical systemRunPlan before approval, and once granted the gateway forwards that stored plan — not any later caller-edited command, cwd, or session fields — and re-validates the working directory before running.

OpenClaw Documentation Guide: 200+ Docs — Where Do You Start?

OpenClaw has 200+ docs. This article helps you see the big picture, understand what each section covers, and decide where to start based on your role.

OpenClaw Reference: Pi Has Been Absorbed — the Built-In Runtime Is Just Called openclaw Now

"OpenClaw is a Gateway shell around Pi" is obsolete. The docs now say the built-in runtime id is openclaw, that pi is a legacy alias which normalizes to it, and that no external agent framework packages remain. The only Pi-related third-party dependency left is a terminal component toolkit.

OpenClaw Desktop Platforms: Windows Now Has a Native Hub, and Node Is Non-Negotiable

Node is the required runtime because the canonical state store uses node:sqlite — Bun is only for installing dependencies. Windows changed the most: there is now a native Windows Hub companion app that installs without administrator privileges and can provision its own app-owned WSL distro for the Gateway.

OpenClaw Mobile Platforms: Phones Are Peripherals, Not Gateways — and the Apple Watch Has Its Own Transport

The iOS and Android apps are nodes, not Gateways: they do not run the Gateway service, and Telegram or WhatsApp messages land on the Gateway rather than on the phone. The Apple Watch is the exception — because watchOS blocks generic low-level networking for ordinary apps, it uses signed HTTPS polling instead.

The OpenClaw Plugin System: Treat Installs Like Running Code, and a Cold Check Proves Nothing About Runtime

The official framing is to treat plugin installs like running code — ClawHub and the bundled catalog are trusted sources, while arbitrary npm, git, and local paths require --force in noninteractive installs. And verification means inspect --runtime, because a bare inspect is only a cold manifest check.

OpenClaw Sandboxing: Four Backends, Three Independent Switches, and Thinking You're Sandboxed When You Aren't

Sandboxing is governed by three independent settings: mode (when it applies), scope (how many containers), and backend (where it runs). The most common failure is an expectation gap — `tools.exec.host` now defaults to auto, so 'unset means sandboxed' is no longer true, and the security audit has a check specifically for it.

OpenClaw Sessions and Memory: One Rolling Conversation, Plus Four Files That Get Written to Disk

By default every DM lands in one main session, and group activity and background work report back into it. Memory is entirely Markdown on disk — the model only remembers what gets saved, with no hidden state. But if more than one person can DM your agent, DM isolation is something you have to turn on.

OpenClaw's Threat Model: It Starts by Telling You What It Does Not Protect

OpenClaw's security docs open by stating the scope: this is a personal-assistant trust model, one gateway per trusted operator. It explicitly is not a security boundary for mutually adversarial users sharing one agent — and a 'not vulnerabilities by design' list pins that down.

OpenClaw Tools, Part 1: A Dedicated Browser, Three Ways to Attach, and Search Results Typed as Untrusted

OpenClaw's browser is a separate agent-only profile, fully isolated from your personal browser. And web_search's return shape carries an externalContent.untrusted marker — search results are typed as untrusted external content at the type level.

OpenClaw Tools, Part 3: Turning Off the File Tools Does Not Make exec Read-Only

exec is a mutating shell surface: disabling write, edit, and apply_patch does nothing to make it read-only. And since sandboxing is off by default, host=auto actually resolves to the gateway — if you really want the sandbox, say so explicitly and it will at least fail closed.

OpenClaw Tools, Part 2: Six Layers of Skill Precedence, and Why Sub-Agents Get No Message Tool

Skills load from six sources with the highest precedence winning on name collisions, and a per-agent list replaces rather than merges. Sub-agents get no session or message tools by default — they return plain text to the parent, and the right to speak to a human stays with the parent agent.

OpenClaw Tools, Part 4: When the Catalog Outgrows the Prompt — Code Mode, Tool Search, and MCP

When the tool catalog no longer fits in the prompt, OpenClaw offers two answers: Code Mode shows the model only exec and wait and has it write small programs against a hidden catalog, while Tool Search keeps structured search/describe/call controls. Neither bypasses tool policy.

OpenClaw Operations: Seven Commands for the First 60 Seconds, and 'It Feels Dumber' Is Usually Not the Model

The official triage flow is seven commands and two minutes to a diagnosis. And the most common symptom — the assistant feeling limited or missing tools — is usually the tool profile: minimal allows only session_status, while coding is the default for new local configs.

OpenClaw UI: A New Rail Lets You Ask What a Session Is Doing Without Interrupting It

The Control UI gained a session rail: it uses a utility model to produce a run digest and attaches a read-only companion thread, so you can ask what a session is doing without entering or interrupting the main agent run. Its contents never enter chat.history.

ai guide

Phil Schmid: Why Agent Harness Is the Most Important Thing in 2026

The model is the CPU, the harness is the operating system, and the agent is the application. No matter how powerful a model is, without a good harness it's just a demo. Phil Schmid argues that harness is the most critical infrastructure in AI engineering for 2026.

Claude Code Troubleshooting Index: Orders 33-35 on Install, Runtime, and Config Diagnosis

The original Claude Code troubleshooting collection has been split into orders 33-35: installation & login, runtime problems, and config diagnosis. This page is their index.

Complete Guide to Bypassing Cloudflare Anti-Bot for AI Agents: From Debugging to Building an MCP Server

Standard Playwright gets blocked by Cloudflare. Both playwright-extra + stealth and nodriver can bypass it. The final step is wrapping the solution into an MCP server so AI agents can use it automatically.

Claude Code Agent Teams in Practice: Team Lead, Point-to-Point Messaging, and a Shared Task Board

Agent Teams lets multiple full Claude Code sessions work as one team: a team lead assigns work while teammates each run their own context window, coordinating through point-to-point messaging and a shared task list. This post covers the three key differences from sub-agents, the trade-off between teammateMode display modes, and why token cost scales linearly with team size.

Claude Code Workflows in Practice: Official Best Practices from Explore to Commit

Anthropic's official Claude Code best practices boil down to one constraint: manage the context window. This post reorganizes their guidance into a working loop — explore, plan mode, implement with a runnable check, verify, commit — plus prompt techniques, when to /clear vs rewind, and the five failure patterns they call out.

Claude Code Channels: External Events, Reply Tools, and Sender Gating

Channels are a special kind of MCP server that push CI failures, monitoring alerts, and Telegram messages directly into a running Claude Code session — and Claude can answer back through the same channel via a reply tool. This post breaks down the channel contract, two-way replies, security gates, and install requirements.

Claude Code Checkpointing Deep Dive: Snapshots, the Rewind Menu, and Tracking Boundaries

Checkpointing is not git commits: Claude Code snapshots your files before every user prompt, keeps the 100 most recent per session, and deletes them after 30 days. This piece breaks down the five rewind menu options, the tracking boundaries (bash, subagents, symlinks), and how checkpoints divide labor with git.

How Claude Code Sees Your Browser: Chrome Integration, Console Debugging, and Form Automation

Claude Code gains browser control through the Claude in Chrome extension: read console logs and DOM state, click, type, upload files, record GIFs, and operate sites you're already signed into. The official prerequisites list Chrome, Edge, and Chromium-based browsers such as Brave, Arc, Vivaldi, and Opera, but WSL is not supported.

Claude Code in CI/CD: @claude on GitHub Actions and the GitLab MR Flow

Put Claude Code into GitHub Actions with anthropics/claude-code-action: /install-github-app sets everything up in one command, @claude in a PR or issue comment gets bugs fixed, branches pushed, and PR creation links returned; Bedrock/Vertex/Foundry backends switch via one input with OIDC and no stored keys; the GitLab CI/CD integration (beta) mirrors it as a single .gitlab-ci.yml job where every change flows through a merge request.

How Claude Code Remembers Your Project: CLAUDE.md Layers, Imports, Rules, and Auto Memory

Every Claude Code session starts with a clean context window. Three memory mechanisms carry knowledge across sessions: CLAUDE.md files loaded every session, .claude/rules/ files that load conditionally via paths frontmatter, and auto memory Claude writes itself. All CLAUDE.md layers are concatenated into context — not inherited by override. This guide covers layer behavior, @path imports, monorepo strategies with nested CLAUDE.md, and sharing one instruction file across tools via @AGENTS.md.

How to manage Claude Code's context window: startup content, per-feature costs, and the compaction trio

Claude Code loads the system prompt, MEMORY.md, CLAUDE.md, MCP tool names, and skill descriptions before you type your first word. This post breaks down the startup context, what each of six extension features costs, and how to control auto-compaction with /compact, /autocompact, and autoCompactWindow.

How Claude Code Sandboxing Works: Sandboxed Bash, Network Allowlists, and the Threat Model of Six Isolation Approaches

Claude Code's built-in sandboxed Bash restricts every command at the OS level: writes are limited to the working directory plus session temp, while reads default to the entire machine; network traffic goes through a proxy allowlist that starts with zero domains. The switches live in the /sandbox panel and sandbox.enabled — there is no --sandbox flag. This post also compares sandbox runtime, dev containers, Docker, VMs, and Claude Code on the web to show when each heavier isolation tier earns its setup cost.

Headless and the Agent SDK: From claude -p to Programmatic Agents

claude -p turns Claude Code from an interactive terminal session into a command you can embed in scripts and CI: pipe data in, get structured JSON out with --output-format json, and skip most auto-discovery with --bare. This post focuses on CLI usage, then covers the four signals that mean it's time to switch to the Python or TypeScript Agent SDK.

How Claude Code connects to external tools: MCP scopes, transports, and auth

Claude Code connects to external tools via MCP (Model Context Protocol), with manual configuration split across three scopes: team-shared .mcp.json, personal local/user entries in ~/.claude.json, and enterprise managed config — not settings.json. This post covers claude mcp add/login flows, the transport landscape (SSE is deprecated), and tool search lazy loading.

Claude Code Plugins & Marketplaces: Packaging Skills, Hooks, and MCP into One Installable Unit

Plugins add distribution, not new capabilities: they collect skills, agents, hooks, and MCP settings scattered across .claude/ into one manifest-backed directory that marketplaces can install, update, and version-pin. This post covers plugin structure, the minimal build flow, publishing a marketplace, and dependency version constraints.

Claude Code Remote Control: Pick Up a Local Session from Any Device

Remote Control turns claude.ai/code or the Claude mobile app into a remote for your local Claude Code session: code still runs on your own machine, with MCP servers and local tools fully available, while transcript sync passes through Anthropic servers. This post covers startup, reconnection, push notifications, file delivery, and the security boundary.

Claude Code settings.json Complete Guide: Five Scopes, Merge Rules, and the Keys That Matter

Claude Code reads settings from five levels — managed settings, CLI flags, project local, shared project, and user — where plain values are overridden by higher levels while list keys like permissions.allow merge across scopes. This guide covers each file's role, the allow/deny/ask rule syntax, and how to verify your settings with /status and claude doctor.

Delegating Coding Tasks from Slack: Claude Code in Slack and Claude Tag

A single @Claude in Slack turns a bug report into a cloud-run Claude Code session. But there are now two paths: Pro/Max stays on the original Claude Code in Slack (each session runs under an individual account), while new or migrating Team/Enterprise setups should look at Claude Tag (shared org identity, admin-configured access and spend). Check your plan before setting anything up.

How Claude Code Sub-agents Work: Context Isolation, Frontmatter Definitions, Background Execution, and Permission Inheritance

Sub-agents are specialized assistants that work in their own context window: a single Markdown file defines their system prompt, tools, and model. Claude delegates automatically based on the description field, or you can @-mention to force one. This post breaks down the frontmatter schema, background execution and nested spawning, permission inheritance rules, and when not to use them.

"Recommend the next route" and "Recommend something similar" are not the same thing — Intent Disambiguation in RAG Recommendation Systems

In a climbing RAG system, 'recommend the next route' (progression) and 'recommend a similar route' (similarity) were conflated by a single hasSimilarRouteIntent() function, causing recommendation quality to collapse. The fix is a two-stage intent classification with a Regex Fast Path + LLM Fallback.

RAG Multi-Entity Queries: When the User Lists Five Routes and the System Only Sees the First

The RAG system's extractRouteReference() used a for...return pattern that grabbed only the first match — so when a user provided five completed routes, only one was used. The fix evolves through three layers: rule-based multi-entity extraction, user profile aggregation, and embedding centroid.

When Vector Search Matches by Name Instead of Grade: Attribute Conflation in RAG Systems

Query: 'I just sent Beauty in the Mirror 5.11b — recommend routes of similar difficulty.' The results came back full of routes with similar-sounding names, not similar grades. Root cause: dense embeddings compress multiple attributes into a single vector, and the rarity of the route name drowns out the grade signal. The fix: three layers of defense — metadata pre-filtering, query rewriting, and score fusion.

ai guide

LangGraph: Managing Agent Workflows with Graph Structures

LangGraph models LLM workflows as directed graphs, solving the pain points of multi-turn iteration, conditional branching, and parallel execution that are difficult to handle with linear pipelines.

tech guide

Biome: Replacing ESLint + Prettier with Rust

Biome does the work of ESLint + Prettier in a single tool, running 10–20x faster with far less configuration. DaoDao uses it across an entire monorepo — lint and format in one pass.

AEO Guide: Answer Engine Optimization — Getting AI Search Engines to Cite Your Content

AEO (Answer Engine Optimization) is a content strategy aimed at AI search engines like Perplexity, ChatGPT Search, and Google AI Overview. The core idea is to make your content the easiest source for AI to cite — not just another link in the results page.

A Complete Guide to Blog SEO — From Meta Tags to Structured Data

SEO is more than keywords. Structured data (JSON-LD), Open Graph, hreflang, and robots.txt are the technical optimizations that actually help search engines understand your content. This guide walks through a complete implementation using an Astro blog as the example.

tech guide

BullMQ: The Most Mature Redis-Backed Job Queue for Node.js

BullMQ is the most mature job queue in the Node.js ecosystem, backed by Redis, with support for priorities, retries, scheduling, and delayed jobs. DaoDao uses it to handle notification delivery and practice auto-completion scheduling.

tech guide

Celery: The Standard Distributed Task Queue for Python

Celery is Python's go-to distributed task queue, using Redis or RabbitMQ as a broker to offload long-running work to the background. DaoDao's AI service uses it to handle async tasks like LLM feedback generation.

Claude Code Global Skills Not Found in New Sessions? Understanding Skill Discovery and How to Debug It

Global skills live in ~/.claude/skills/, but they go missing in new sessions or the Desktop App? The problem usually isn't a missing file — it's that the skill descriptions aren't being loaded into context. This post clarifies the CLI vs Desktop App differences, the role of settings.json, and the most reliable fix.

tech guide

ClickHouse: When PostgreSQL Analytics Queries Start Slowing Down, You Need OLAP

ClickHouse is a column-oriented OLAP database that scans hundreds of millions of rows in seconds. DaoDao uses it to record user behavior events for the AI recommendation engine's feature engineering, letting PostgreSQL focus on transactional data.

Cloudflare D1: SQLite Relational Database at the Edge

D1 is Cloudflare's serverless SQLite database that binds directly to Workers, supports full SQL (JOINs, transactions), and handles automatic backups. It's well-suited for small-to-medium relational data needs — NobodyClimb uses it as its primary database.

Cloudflare KV: A Global Edge Key-Value Store

KV is Cloudflare's globally distributed key-value store. Reads are served from the nearest edge node with extremely low latency. It's ideal for caching, feature flags, and ephemeral data — but writes are eventually consistent.

Cloudflare R2: An S3 Alternative with Zero Egress Fees

R2 is Cloudflare's object storage service — S3-compatible API, zero egress fees, and native Workers binding. Stop worrying about bandwidth bills for media-heavy applications.

Cloudflare Workers: Not Lambda, Not Containers — It's V8 Isolates

Cloudflare Workers uses V8 Isolates instead of containers — no cold starts, global edge deployment, and direct access to D1, R2, KV, and AI via Bindings. Great for APIs, SSR, and lightweight backends; not suited for CPU-heavy work.

tech guide

Docker in Practice: Containerizing from Development to Deployment

Docker lets you bundle your application together with its environment, eliminating the 'works on my machine' problem. Combined with multi-stage builds and Compose, it's an essential tool for modern backend deployment.

tech guide

Expo + React Native: What It's Actually Like to Ship One Codebase for iOS and Android

Expo turns React Native development from 'environment setup hell' into a state where you can just start writing logic. Expo Router brings file-based routing that dramatically lowers the barrier for web developers making the switch. Both DaoDao and NobodyClimb use it to ship across iOS and Android.

tech guide

Express.js: The Default Answer for Node.js Backends, and Why It Still Makes Sense

Express is the most mature Web framework for Node.js, with a rich middleware ecosystem and abundant learning resources. Paired with TypeScript and a clear layered architecture, it remains a justifiable choice in 2026.

tech guide

FastAPI: The Go-To Framework for Python AI Services

FastAPI is a modern Python web framework built on type hints — it auto-generates OpenAPI docs, supports native async, and delivers performance close to Node.js. It's the top choice for AI/ML services and the most worthwhile framework to learn in the Python backend ecosystem.

tech guide

GitHub Actions: A CI/CD Primer and Monorepo Strategy

GitHub Actions is the lowest-friction CI/CD tool available today, ideal for small-to-medium projects. The key to monorepos is using path filters so only affected apps trigger a build.

Hono: The Lightweight Web Framework Built for Edge Runtimes

Hono is a web framework designed specifically for edge runtimes like Cloudflare Workers, Deno, and Bun. It's an order of magnitude lighter than Express, natively supports Web Standard APIs, and is the go-to choice for edge environments.

tech guide

Next.js 15 + App Router: What Server Components and use cache Actually Do

Next.js 15 + React 19's App Router shifts rendering responsibility from the client to the server. use cache ties caching logic directly to data functions instead of scattering it across fetch options. Both DaoDao and NobodyClimb chose this stack for very practical reasons.

@opennextjs/cloudflare: Running Next.js on Cloudflare Workers

@opennextjs/cloudflare enables Next.js App Router deployments on Cloudflare Workers — dynamic SSR runs in a Worker, static assets are served from Cloudflare Assets. Zero server management, but with clear feature limitations.

tech guide

PM2: The Practical Choice for Node.js Process Management

PM2 keeps your Node.js app running on a server — auto-restarts on crash, supports cluster mode to max out CPU cores, and handles log management. Nearly every Node.js app deployed on a VM or VPS needs it.

tech guide

Prisma ORM: Type-Safe Database Access for TypeScript Projects

Prisma's schema-first design gives you versioned migrations, full TypeScript types on every query, and intuitive relation handling. The tradeoff is a learning curve and the inherent limits of any ORM abstraction — but for most TypeScript projects, it's a worthwhile deal.

tech guide

React Hook Form + Zod: The Best Combo for Form Handling

React Hook Form handles form performance, Zod defines the validation schema — together they eliminate nearly all form boilerplate. Share a single Zod schema across a monorepo and you get one source of truth for both frontend and backend validation.

tech guide

Redis Essentials: Caching, Sessions, and Pub/Sub in One Go

Redis is an in-memory key-value store that's blazingly fast. DaoDao uses it to handle three responsibilities at once — API caching, session storage, and BullMQ job queues — all from a single Redis instance.

tech guide

shadcn/ui: Not a Package — It's Copy-Pasted Component Source Code

shadcn/ui is not an npm package — it copies component source code directly into your project, giving you full ownership. DaoDao uses it to build packages/ui, a shared component library used across three Next.js apps.

tech guide

TailwindCSS: Utility-First Is a CSS Management Strategy, Not Just a Style Preference

TailwindCSS's core value is solving CSS's global namespace pollution and dead code problems. Utility classes keep styles co-located with components, and unused classes are automatically purged at build time — production CSS bundles typically come in at just a few dozen KB. Both DaoDao and NobodyClimb use it for web styling.

tech guide

Tamagui: A React Native UI Framework — Why NobodyClimb Chose It Over NativeWind

Tamagui is a UI framework built for React Native with a complete design token system, theme support, and compile-time optimization that moves style computation to build time. NobodyClimb chose it over NativeWind primarily because its cross-platform token system is more robust.

tech guide

TanStack Query: The Standard Solution for Server State

Managing API data with useState + useEffect means reinventing the wheel — and doing it worse. TanStack Query handles caching, background updates, and loading/error states so you can focus on UI logic.

tech guide

Turborepo + pnpm Workspaces: The Standard Approach to Monorepos

Turborepo solves monorepo build speed problems; pnpm workspaces solves dependency sharing. Together they are the best choice for JS/TS monorepos today.

tech guide

Zod: Runtime Type Validation for TypeScript

TypeScript types only exist at compile time — they vanish at runtime. Zod lets you validate external data at runtime while inferring TypeScript types from the same schema. One definition, two jobs done.

tech guide

Zustand: The Lightest Global State Management for React

No Provider, no reducer — global state in just a few lines. NobodyClimb uses it for auth and UI state, paired with TanStack Query for server state.

A One-Person Full-Stack Team: AI-Driven Development Workflow from OpenSpec to Auto-Deploy

Use OpenSpec to break requirements into engineering tasks, Claude Code to implement them, hooks to auto-format and protect, local review before committing, three AI reviewers running in parallel on PR, and auto-deploy after merge. This entire workflow lets one person maintain quality across six sub-projects.

Claude Code Hooks: A Complete Guide to Event-Driven AI Control

Hooks are Claude Code's event system. They trigger shell commands, HTTP requests, MCP tools, or LLM evaluations automatically before/after tool execution, when a prompt is submitted, or when a task ends. Use them to block dangerous operations, run automated reviews, inject context, or write audit logs.

Claude Code Skills: A Complete Guide to Turning Repetitive Workflows into Single Commands

A Skill is an SOP written for AI. Define the steps in a Markdown file and Claude follows them. No coding required, no frameworks to learn — just write down what an experienced person would do.

Turning Debug Sessions into GitHub Issues with a Claude Code Skill: Designing /file-bug-issue

Stuck mid-debug and can't fix it right now? Use /file-bug-issue to package the error analysis, reproduction steps, and attempted fixes from your conversation into a well-structured GitHub issue. Pair it with a Remote Agent to let AI automatically take over the fix.

Let AI Pick Up Issues, Write Code, and Open PRs: Hands-Off Development with Claude Code Remote Agent

Using Claude Code's Scheduled Remote Agent, automatically scan GitHub issues every 2 hours, implement features, open PRs, and address review feedback — no human intervention required. Humans only write issues and click merge. Pair it with the custom /publish-tasks skill to push OpenSpec engineering tasks directly to GitHub issues.

ai guide

Langfuse Complete Guide: LLM Application Observability from Scratch

Langfuse is currently the most mature open-source LLM Observability platform. This post covers four core capabilities — Tracing, Prompt Management, Evaluation, and Datasets — showing you how to use them in real projects.

Claude Code's Three-Layer Quality Defense: Hooks, Skills, and Instruction Files

Hooks are automated safety nets (blocking bad commits), Skills are interactive workflows (running checks + auto-fixing), and instruction files (CLAUDE.md / AGENTS.md) are behavioral guidelines. Each layer operates independently, but together they enable an AI agent to automatically run lint, typecheck, and build checks before every commit.

tech guide

How to Classify Code Review Comments? From Conventional Comments to AI Review Tool Taxonomies

Three main classification systems dominate: Conventional Comments (label-based), Google's severity prefixes (Nit/Optional/FYI), and SonarQube's four quadrants (Bug/Vulnerability/Code Smell/Hotspot). AI review tools have each developed their own taxonomies, but the core dimensions consistently converge on four areas: correctness, security, performance, and maintainability.

Context Engineering: Why Your AI Agent's Problem Is Information, Not the Model

Context Engineering is the core concept that replaced Prompt Engineering in 2025: the focus shifted from 'how to ask' to 'what information to provide.' Delivering the right information at the right time into the context window is more effective than upgrading to a stronger model. This post covers the definition, four key strategies, practical techniques, and common failure modes.

tech guide

From Mock to Real AI: Integrating Cloudflare Workers AI into action-maker

Upgraded action-maker from hardcoded mock data to live Cloudflare Workers AI generation. The architecture splits into Worker (AI only), Server (data storage), and Frontend (orchestration). Hit two gotchas along the way: Qwen3's thinking block and the Workers AI response format.

ai guide

MCP (Model Context Protocol): The Standardized Protocol for AI Agent Tool Invocation

Every AI tool has its own calling format, making integration costly. MCP (Model Context Protocol) is an open standard proposed by Anthropic that unifies the communication protocol between AI Agents and external tools/data sources, enabling tools to be reused across Agents.

tech guide

False positives in Node.js image vulnerability scans? Separate app packages from npm built-ins first

When reviewing vulnerability scan results for a Node.js Docker image, you can't just look at package names. First distinguish between project dependencies and the packages bundled with npm inside the base image — otherwise you'll fix the wrong thing.

tech guide

What Is Vulnerability Scanning? A Quick Intro to Docker and Package Scanning with Trivy

Vulnerability scanning isn't just about generating reports — it helps you discover known risks in your system before they become incidents. This post uses Trivy as a hands-on example to explain what scanners actually look for, how to read the results, and how to get started.

tech guide

Turning a Scraper Script into an MCP Server for Claude to Use Directly

Wrap a local Python script into an MCP Server using FastMCP so Claude Code can call it directly — no more manually running pipelines.

tech debug

MCP Tool Returns 1M Characters: The Token Explosion in search_local_jobs

The MCP tool was returning a description field that caused 1,033 job listings to exceed the token limit. The fix: exclude description by default and add pagination.

ai guide

Agent Memory Systems: From RAG to Read-Write Memory Evolution

RAG is read-only. Agent Memory lets AI not only read but also write and persist information. Three memory types: Procedural (behavior patterns), Episodic (temporal events), and Semantic (factual knowledge) form a complete cognitive memory system.

ai deep-dive

Complete Guide to AI Agent Architecture Patterns: From Three Pillars to Multi-Agent Systematic Navigation

AI Agent is not a single technology -- it is an entire architecture system. This article is a systematic navigation: starting from the Agent Three Pillars (Context/Cognition/Action), through the three-stage evolution of AI engineering (Prompt -> Context -> Harness), to eight Multi-Agent design patterns and production-grade Harness infrastructure. Each topic links to a dedicated deep-dive article.

ai guide

The Three Core Pillars of AI Agents: Context, Cognition, Action

An AI agent is not a black box — it is built from three layers: what it knows (Context), how it thinks (Cognition), and what it can do (Action). Understanding these three layers is the key to grasping why agents are sometimes brilliant and sometimes go off the rails, and how to design a truly effective agent system.

tech guide

docker restart Does Not Re-apply Volumes — Debugging a Bind Mount Failure

docker restart does not recreate the container, so changes to volumes in docker-compose.yml only take effect after running docker-compose down && up.

Multi-Agent RAG: Distributed Retrieval Architecture with Specialized Agent Collaboration

A single RAG Agent handling all queries hits knowledge boundaries and performance bottlenecks. Multi-Agent RAG dispatches retrieval tasks to multiple specialized Agents, each with its own knowledge base and retrieval strategy, coordinated by a central Orchestrator that merges results.

Claude Code Permission Modes Explained: Five Modes from Default to Auto

Claude Code has five permission modes: default (confirm each step), acceptEdits (auto-accept edits), plan (read-only planning), auto (background AI classifier review), and bypassPermissions (YOLO, skip everything). Switch with Shift+Tab or configure via settings.json. Auto mode is the sweet spot — no step-by-step confirmations, but with safety guardrails.

tech guide

nginx 502: Debugging Cross-Compose Container DNS Resolution

Service names aren't resolvable across Compose projects — you need to add a network alias so nginx can find the container.

tech guide

Installing and Verifying Superpowers for GitHub Copilot CLI: Implementation, Diagnostics, and Validation

A hands-on log of installing Superpowers (packaged by DwainTR) for Copilot CLI on a local machine — including the diagnostic process when skills didn't appear after installation, the fix, and practical tips.

tech guide

Docker DNS Resolution: container_name vs network alias

Cross-project DNS resolution requires container_name or a network alias — and only aliases support horizontal scaling.

LongRAG: Rethinking RAG Chunking Strategy with Long-Context Models

Traditional RAG splits documents into small chunks for retrieval, but this causes information fragmentation. LongRAG leverages 100K+ token long-context models to retrieve larger document segments (entire sections or even whole documents), reducing fragmentation while maintaining retrieval efficiency.

Speculative RAG: Small Models Draft in Parallel, Large Model Verifies at Once

Speculative RAG uses small specialist models to generate multiple answer drafts from different document subsets in parallel, then a large model verifies and selects the best answer in one pass. The paper reports +12.97 points accuracy and -50.83% latency on PubHealth — but that is the best cell in the table; other benchmarks gain far less.

tech guide

nginx Restarted Fine, but Cloudflare Keeps Returning 502 — Even Though the Origin Is Healthy

A brief error during nginx restart caused Cloudflare to mark the origin as unhealthy and stop forwarding requests, returning 502 on its own. The key clues: localhost hits to the origin return 200, and nginx access logs are completely empty. Just wait for Cloudflare to automatically re-check the origin — it recovers on its own.

tech guide

Managing Multi-Service Reverse Proxy with nginx conf.d: A Daodao Case Study

A monolithic nginx.conf becomes unwieldy as services grow. Splitting it into per-service files under conf.d/ via include is the standard solution.

tech guide

nginx First Request Always 502, All Subsequent Requests Fine

When nginx uses the `set $variable` pattern for dynamic upstreams, the DNS cache expires every 30 seconds — the first request after expiry hits a 502 because no IP is available. Upgrading to nginx 1.27.3 and switching to an upstream block with the resolve parameter fixes this: DNS updates happen asynchronously in the background.

tech guide

Downloading Files from a VPS Using SSH Config Aliases

Once SSH config is set up, scp works directly with aliases — no need to type out the full IP every time

ai guide

The Complete Ollama Guide: Run LLMs Locally with One Command

Ollama wraps llama.cpp in a Docker-style CLI + REST API, letting you run LLMs locally with a single command. This post covers core concepts, installation, API, hardware requirements, Modelfile customization, and what this tool is — and isn't — good for.

The Complete Guide to RAG System Patterns: A Ten-Generation Evolution from Naive to Multi-Agent with Practical Navigation

RAG has evolved far beyond simple 'search + generate' into a technology ecosystem spanning ten generations — and since 2025 into an Agentic/Reasoning era. This article is a systematic navigation guide: from Naive RAG to Multi-Agent/LongRAG across ten generations, the post-ten Agentic Era (Search-R1/RL search, MCP, GraphRAG 3.x, vision-native retrieval), retrieval strategies, chunking, embedding, reranking, evaluation frameworks, observability, and cost optimization. Each topic has a dedicated deep-dive article.

vLLM — From PagedAttention to a Production-Grade LLM Inference Engine

vLLM uses PagedAttention to eliminate KV cache memory waste, combining continuous batching and prefix caching to become the most widely adopted open-source LLM inference engine today.

tech guide

Ghostty vs cmux: A Guide to Choosing Your Modern Terminal

Ghostty is a fast, native, general-purpose terminal emulator. cmux is a terminal built on top of Ghostty, specifically designed for AI coding agents. They're not competitors — they operate at different layers.

ai guide

Complete Chatbot Development Guide: State Management, Memory Strategies, and Tech Stack Selection

Building a chatbot is more than just calling an API. Conversation state management, memory mechanisms, streaming, guardrails, observability, and tech stack selection — every layer affects the user experience.

ai guide

Prompt Engineering in Practice: Iteration Methodology, Common Mistakes, and Few-shot Optimization

Good prompts aren't written in one go — they're iterated into existence. Start with the simplest prompt, test with real cases, classify error types, and make targeted fixes. This article covers the three-part System Prompt structure, reasoning framework selection, few-shot optimization, token budget management, and six common mistakes.

Cloudflare Free Plan Maintenance Page: Custom Error Pages Unavailable, Use a Worker Instead

Cloudflare Custom Error Pages require a paid plan. On the Free Plan, use a Worker with inline HTML to intercept 5xx responses instead.

tech guide

Managing Personal and Work GitHub Accounts with Git Conditional Includes

Use includeIf + SSH Host aliases to let Git automatically switch accounts based on directory path — no more manual switching.

Astro + Cloudflare Workers: Native Modules Break the Build Even on Prerendered Routes

Even when a route has prerender = true, Cloudflare Workers' Rollup bundler still attempts to bundle native modules, causing the build to fail. The fix is to move any native module work into a postbuild script.

tech debug

Astro Scoped CSS Not Applied to MDX-Rendered Content

Astro scoped CSS appends a scope hash to each selector, but elements rendered by <Content /> don't receive that hash — causing all prose styles to silently break.

Agentic RAG: Letting the LLM Decide When to Search Again

For complex multi-hop questions, a single RAG search isn't enough. Agentic RAG lets the LLM evaluate whether retrieved results are sufficient — if not, it rewrites the query and searches again, forming a ReAct loop.

BGE-M3: Why This Embedding Model Works Well for Traditional Chinese RAG

Your choice of embedding model directly determines RAG search quality. BGE-M3's multilingual training, 1024-dimensional vectors, and matching Reranker make it a practical pick for Traditional Chinese RAG.

Chunking Strategies: How You Split Text Determines Whether RAG Can Find the Answer

Chunks too large and retrieval loses precision; too small and you lose context; hit a table and retrieval falls apart entirely. Chunking is the most underrated part of RAG — pick the wrong strategy and no amount of downstream optimization will save you.

ColBERT: The Third Way in Vector Search

Bi-Encoders are too coarse, Cross-Encoders are too slow — ColBERT's Late Interaction finds the sweet spot: token-level comparison between query and document, but with document vectors that can be precomputed.

Contextual Retrieval: Giving Every Chunk Its "What This Is About" Context

When you split a document into chunks, each chunk loses its place in the original document. Contextual Retrieval solves the isolated-chunk problem by generating a per-chunk context from the whole document and prepending it at index time.

CRAG: Automatically Relaxing Filters When Retrieval Comes Up Empty

Filters too strict and getting zero results? CRAG automatically relaxes them and retries — far better than letting the LLM hallucinate an answer from general knowledge.

Cross-Encoder Reranking: Surfacing the Most Relevant Documents

Vector search similarity scores don't equal relevance. Cross-Encoders use pairwise comparison to reorder results and push the truly relevant documents to the top.

GraphRAG: Structuring Knowledge as a Graph for Relationship-Based Reasoning

Vector search finds similarity; graph search traverses relationships. When a question requires reasoning across multiple entities — crag → route → sender → grade distribution — GraphRAG outperforms standard RAG.

Hybrid Search: Using BM25 + Vector Search to Cover Each Other's Blind Spots

Vector search handles semantics; BM25 handles keywords. Combining them with RRF is what lets you handle both fuzzy queries and exact terms at the same time.

HyDE: Boosting Vector Search Recall with Hypothetical Answers

Have an LLM generate an 'ideal answer' first, then embed that hypothetical document for search — it outperforms searching with the raw query.

RAG Personalization: Learning User Preferences from Conversations

After each conversation, asynchronously extract likely user preferences and skill level, then automatically personalize search parameters on the next query — no manual setup required.

MMR + Popularity Weighting: Recommendations That Are Both Relevant and Diverse

Ranking purely by relevance leaves you with five documents all describing the same route. MMR strikes a balance between relevance and diversity, and layering in popularity weighting makes results even more useful.

Modular RAG Pipeline: Designing RAG as a Composable DAG

RAG doesn't have to be a rigid three-step process. It's a set of steps that can be dynamically enabled, skipped, or reordered. Pipeline as Code lets the system adapt its behavior without redeployment.

Multi-Query Expansion: Search One Question from Multiple Angles

A single vector search on a complex query often misses relevant documents. Let the LLM rewrite the query into 3-5 sub-queries, run them in parallel, and recall improves significantly.

Multimodal RAG: Bringing Images into the Knowledge Base

Climbing routes carry a ton of visual information (topos, wall photos) that text-only RAG misses entirely. Multimodal RAG makes images searchable and understandable.

Three Generations of RAG: From Naive to Modular

Naive RAG works but has real problems. Advanced RAG patches those problems. Modular RAG rearchitects the whole system to be composable and configurable. Understanding all three generations is the key to understanding why modern RAG systems look the way they do.

Plan-and-Execute: A RAG Pattern That Plans Before It Acts

For complex queries, have the LLM map out what information is needed and in how many steps — then execute that plan. More systematic than thinking on the fly.

Query Classification: Teaching Your RAG System How to Answer Each Question

Not every question needs full RAG. Classify queries with an LLM first, then route to the right execution path — saving cost and improving accuracy.

RAG A/B Testing: A Scientific Approach to Comparing Pipeline Configurations

"Adding a Cross-Encoder feels better" is not a scientific evaluation. A/B testing tells you whether a change actually works, how much it helps, and which query types benefit.

RAG Cold Start: Building a Useful System When You Have No Data

A RAG system needs data to answer questions, but data only accumulates as the system gets used. Cold-start strategy is what bridges the gap from empty to useful.

RAG Cost Optimization: Minimizing the Cost of Every Query

RAG system costs come from LLM tokens, Embedding APIs, and vector search. Every stage has room for cost reduction, but you need to verify that optimizations don't sacrifice too much quality.

RAG Evaluation Frameworks and Tool Selection: Promptfoo, RAGAS, DeepEval, and TruLens

No industry standard mandates one RAG evaluation tool. Measure retrieval, generation, and operations separately, then choose Promptfoo, RAGAS, DeepEval, or TruLens for the actual stack.

RAG Common Failure Modes: 10 Problems and Their Solutions

When a RAG system breaks, 90% of the time it's one of these 10 failure modes. Identify which one first, then apply the matching fix — far more effective than optimizing blindly.

RAG Guardrails: Adding a Defense Layer to Inputs and Outputs

The attacks RAG systems face go beyond the technical level — Prompt Injection and Jailbreak are real threats. Both inputs and outputs need independent protection layers.

RAG Observability Tool Landscape: Choices in 2026

Rolling your own traces is good enough, but open-source tools save you a lot of work. Langfuse, Phoenix, and LangSmith each have their niche — the right choice depends on your trade-offs around self-hosting, open source, and integration complexity.

RAG Observability: Per-Node Tracing to Turn the Black Box Transparent

The hardest part of a RAG system isn't building it — it's figuring out why a particular answer went wrong. Pipeline Tracing records every step's decisions and data so debugging has a clear trail to follow.

RAG Prompt Engineering: How to Design System Prompts and Context

Search found the right documents, but the LLM's answers are still poor — often the problem lies in prompt design. System prompt structure, context formatting, and instruction placement all affect output quality.

RAG Streaming: Using SSE to Display LLM Responses as They Generate

LLM generation takes 3-5 seconds, and waiting for the full response before displaying it makes for a terrible experience. SSE pushes tokens as they're generated, reducing time-to-first-character from 5 seconds to under 1 second.

RAG Quota System: Controlling LLM Costs with Dual Limits

Limiting request count alone is not enough — a single long query can consume ten times the tokens of a normal one. Dual quotas (request count + token count) are what truly control costs.

RAG vs Fine-tuning: It's Not Either/Or

RAG and Fine-tuning solve different problems. RAG gives the model new knowledge; Fine-tuning changes the model's behavior and style. In most cases you use both, not pick one.

RRF: How to Merge Multi-Source Results in RAG Systems

BM25, vector search, HyDE, and Multi-Query each produce separate result sets -- how do you merge them sensibly? RRF uses ranks instead of scores, sidestepping the fundamental problem that scores from different systems are incomparable.

Self-Reflection + LLM-as-Judge: Having AI Evaluate Its Own Answers

Use another LLM to evaluate answer accuracy and quality — if the score is too low, regenerate, and automatically add appropriate disclaimers.

Semantic Caching: Run the RAG Pipeline Only Once for Semantically Similar Queries

Caching doesn't have to match exact query strings -- semantically similar questions can hit the cache too, skipping the entire RAG pipeline execution.

SPLADE: Smarter Sparse Vector Search Beyond BM25

BM25 only recognizes words that appear in the query. SPLADE infers related terms and adds them to the search, gaining partial semantic capability while preserving the precision of keyword search.

Text-to-SQL Router: Precise Queries That Skip RAG

Questions like 'how many routes did I complete this year' will never be answered well by RAG semantic search — querying the database directly is far more accurate. Let the LLM identify intent, extract parameters, and execute predefined SQL templates.

Vector Database Selection: How to Choose Between Pinecone, Weaviate, Qdrant, and Vectorize

Vector database selection is more constrained by deployment platform than LLM selection. Determine your platform and scale requirements first, then evaluate features — don't just look at benchmarks.

education guide

Why Your Learning Goals Always Fizzle Out — And How DaoDao Wants to Fix It

The core reason self-directed learning fails isn't lack of motivation — it's the absence of a co-learning environment. DaoDao turns 'wanting to learn' into 'actually learning' through themed practices, inspiration feeds, group challenges, and learner connections, while turning your growth journey into tangible proof of competence.

product project

The Next Frontier in Online Learning: Why Completion Rate Is the Real Problem

MOOC completion rates hover at just 5–15%, and the problem isn't course quality — it's the execution gap. DaoDao positions itself as a 'Learning OS,' using public commitments, community interaction, and AI recommendations to make learning visible and sustainable.

product project

From 'Want to Learn' to 'Actually Learning': The Product Design Thinking Behind DaoDao

DaoDao is not a content platform -- it's a learning connector. Using anti-perfectionism design, community co-learning, and zero-decision recommendations, it helps learners bridge the execution gap -- from vague ideas to actionable plans.

Why Does a Climbing Community Need AI? NobodyClimb's Experiment and What We Learned

NobodyClimb uses RAG to tackle scattered climbing route information, ties quota limits to community engagement, and leverages Cloudflare Workers AI to bring inference costs close to zero.

NobodyClimb: Why the Climbing Community Needs Its Own Platform

The climbing community doesn't lack the will to share — it lacks a place to connect and preserve its culture.

The Correct Way to Bind a Custom Domain in Cloudflare Workers

In wrangler.jsonc, use custom_domain: true in routes with only the hostname as the pattern — no /* wildcard

tech deep-dive

DaoDao Tech Architecture: Monorepo, Multi-Language Backend, and AI Recommendation System

Next.js + Expo frontend, Node.js + Python dual backend, PostgreSQL + Redis core — plus a social notification system and LLM recommendation engine. Here's how DaoDao builds a learning community platform with a modern tech stack.

NobodyClimb: Building a Climbing Community Platform Entirely on Cloudflare

A climbing community platform where the web app, mobile app, and AI Q&A all run on Cloudflare — no dedicated servers.

NobodyClimb AI Architecture: Building a 20-Node RAG Pipeline on Cloudflare Workers

A dynamically composable RAG pipeline built on Cloudflare Workers AI (gemma-3-12b-it + bge-m3): 14 base steps + 6 LangGraph-specific nodes, with three strategy graphs (Baseline / Agentic / Plan-Execute) selected at runtime.

tech guide

What You Need to Know Before Switching Astro Blog Templates

Switching templates means replacing the entire project foundation. Figure out what you actually need first, then choose between AstroPaper, Cactus, or AstroWind.