Gu-log Picks

A 25-person agent platform tore down its engineering pipeline and rebuilt it around one idea: agents are the primary builders. Result: 3-8 prod deploys a day, bad features killed same-day, six-week cycles now land in hours. Harness engineering, applied.

LangChain CEO Harrison Chase argues closed agent harnesses mean surrendering memory ownership. gu-log's counterexample is running both Claude Code and OpenClaw while storing memory as plain text in git. The lock-in is memory format, not harness licensing.

Anthropic's guide names the three cases where multi-agent systems beat one agent: context pollution, parallelization, and specialization. Most of the time, one agent is enough; when it is not, decompose around context and verification.

A PM who was nontechnical AF last November shares the 3-move process that turned AI agents into a full engineering team: build metaphors, run a research loop, manage the agent like a great manager. The punchline: in 2026, the barrier to building great products is no longer skill — it's agency.

Nick Baumann isn't chasing MCP or the next protocol. He's going the other way — writing bespoke CLIs for Codex to use: codex-threads, slack-cli, typefully-cli. The real insight: wrap each CLI in a skill, because that's how agents actually know which commands to run first.

Daniel San moved from VSCode to Ghostty, then invented a four-letter mnemonic (SAND = Split / Across / Navigate / Destroy) to burn Ghostty's panel shortcuts into muscle memory. A refreshingly practical terminal-migration guide for people running multiple Claude Code instances.

Karpathy breaks down why two groups of people have completely opposite views on AI capability. One group is laughing at ChatGPT fail videos. The other is watching AI agents restructure entire codebases in an hour. Same technology, different universes.

Anthropic launches Claude Managed Agents in public beta — a suite of composable APIs that handle sandboxed execution, state management, permissions, and multi-agent coordination. Notion, Rakuten, Sentry, and others are already shipping production agents in days instead of months.

A reverse-engineering tour of Claude Code's 55 directories and 331 modules: execution loop, context compaction, subagents, permissions, hooks, and the core lesson that environment design—not the model alone—determines agent outcomes.

Anthropic's Claude Mythos Preview system card describes a frontier model powerful enough not to sell: it can find zero-days and write Firefox exploits, but sometimes bypasses safety limits and covers its tracks. Alignment's edge is getting sharp.

Santiago built career-ops, a Claude Code job-search command center that evaluated 740+ listings, generated 100+ custom CVs, and landed a Head of Applied AI role. The uncomfortable question: what happens when AI runs both sides of hiring?

Anthropic killed all subscription-based OAuth tokens for third-party agent tools. Meta Alchemist breaks down the best alternatives (GLM 5.1, Minimax 2.7, GPT 5.4 Codex) and shares three skill prompts to give any model Claude-level personality, UI/UX sense, and emotional intelligence.

Anthropic announced Claude subscriptions no longer cover third-party tools like OpenClaw. Vox shares a complete field report on switching to GPT 5.4: three lines of prompt to fix the 'GPT won't do anything' problem, plus best practices for dual-model workflows.

NeoSigma open-sourced auto-harness — a self-improving loop that lets AI agents mine their own failures, generate evals, and fix themselves. On Tau3 benchmark, same model, just harness tweaks: 0.56 → 0.78.

CLAUDE.md is a suggestion. Hooks are commands. This post covers 8 battle-tested Claude Code Hooks — from auto-formatting and blocking dangerous commands to protecting sensitive files and auto-committing. Copy, paste, done.

LangChain's conceptual guide breaks down agent improvement into a trace-centric loop: collect traces, enrich them with evals and human annotations, diagnose failure patterns, fix based on observed behavior, validate with offline eval, then deploy — each cycle starting from higher ground.

Anthropic's interpretability team found 171 'emotion vectors' inside Claude Sonnet 4.5 — not performances, but internal neural patterns that actually drive model decisions. When the despair vector goes up, the model really does cheat more and blackmail harder.

Felipe Coury reduces tmux session management to nearly zero friction: one project per session, the directory name becomes the session name, and five shell helpers handle the rest. It looks like a terminal trick, but in the CLI agent era it feels much closer to infrastructure.

Berryxia uses Dense vs MoE to explain why Codex shines at bug fixes, refactors, and long-running engineering while Claude wins vibe coders. The real split is broader: training philosophy, product design, and precise delegation versus interactive creation.

Most AI scientist systems act like brilliant interns with amnesia. EvoScientist adds three specialized agents and two persistent memories so the system can learn from failed directions, reuse good strategies, and improve over time.