Agentation

Claude Code gotcha. The Default option in /model is auto fallback. It runs Opus until you hit a usage threshold, then hands the session to Sonnet, and the switch is easy to miss. Pick Opus explicitly to pin it. Know which model wrote the code before you judge it.

"the AI reviewed itself" is not a review. 🚦 between working and done sit two deterministic gates — lint, types, tests, security. pure code, zero AI tokens. they either pass or they block. deterministic gates, explained → agentation.fr/deterministic-…

an AI coding team is two humans and a structure — not a swarm. 🤝 a tech lead encodes the rules once. product owners keep shipping. nobody blocks a developer. how the DUO works → agentation.fr/ai-coding-team…

multi-agent development without the swarm: 🛰️ one Lead that owns the rules. many Workers, each on one task, in its own worktree. every diff read before it ships. the architecture → agentation.fr/multi-agent-so…

your people are already vibe coding. 👀 ban it and you get shadow IT. allow it raw and you ship chaos. the third option: structure it — rules, gates, approved agents, audit. for IT leadership → agentation.fr/vibe-coding-in…

AI in your org either ships chaos or ships standards. 🔒 rules locked once. gates on every change. only approved agents. everything through your own github. governance, built in → agentation.fr/ai-coding-gove…

meet the citizen creator. 🧑‍🎨 the person who knows the need, the users, the business — but never wanted to fight the tech to make it real. now they ship, governed. a product owner without engineers → agentation.fr/product-owner-…

five columns. no human drag. 📋 todo → spec → working → review → done. humans can't move a card; only the Lead advances it, and only once the gates are green. from prompt to production → agentation.fr/from-prompt-to…

ship like a team of ten. ⚙️ annotation in, dispatcher routes, workers build in isolation, deterministic gates verify, your github ships. you never face the AI alone. the full pipeline → agentation.fr/ship-like-a-te…

sovereignty isn't about the models. it's about the tools. 🇫🇷 french team, EU hosting, your own github, GDPR by construction — and deterministic gates checking every line before prod. → agentation.fr/european-ai-co…

@thisguyknowsai @grok is that fake ? Analyse has a AI PHD from a Tech giant

Another WTF moment. A developer just open-sourced a coding agent harness that boots 245x faster than Claude Code. It's called jcode. You launch it and the first frame renders in 14 milliseconds. Claude Code takes 3,436. One active session uses 27.8 MB of RAM. Claude Code uses 386.6. Run ten sessions in parallel and jcode holds at 117 MB while OpenCode swells to 3.2 GB. Each agent has a semantic memory graph instead of a scratchpad. Every turn gets embedded as a vector. The graph is queried on every turn for related memories, and a sideagent verifies the hits before injecting them into context. Consolidation runs in the background to check for stale or conflicting facts. No manual /remember calls. No token burn on lookup tools. The provider list is 30+ deep. Claude, ChatGPT, Gemini, GitHub Copilot, Azure, OpenRouter, DeepSeek, Groq, Mistral, Perplexity, Fireworks, Ollama, LM Studio, and any OpenAI-compatible endpoint you point it at. Ran out of tokens on your first ChatGPT Pro sub? /account swaps to the second. Then there's Swarm. Spawn two agents in the same repo and the server manages them. When agent A edits a file agent B has been reading, agent B gets pinged and can check the diff. Agents can DM each other, broadcast to the room, or spawn their own worker teams for parallel tasks. Groups, channels, and completion statuses are handled automatically. The UI has live side panels that render mermaid diagrams inline. To make it fast, the author wrote a Rust mermaid renderer 1800x faster than the JavaScript one, then wrote a custom terminal called Handterm because no existing terminal could do smooth partial-line scrolling. Self-dev mode is where it gets wild. Tell your agent to enter self-dev and it starts editing jcode's own source code, rebuilds the binary, reloads it live, and keeps working across your existing sessions. You can also resume broken sessions from Claude Code, Codex, OpenCode, or pi directly inside jcode. Anthropic's cache goes cold at the 5-minute mark and you're staring down a big cache miss on your next turn? The UI warns you before you spend the tokens. Written in Rust. MIT licensed. Runs on macOS, Windows, Linux, and Termux. Sitting at 11.2k stars with a native iOS app coming. github.com/1jehuang/jcode

A rule of thumb for Claude Code automations: let the model write the content, but keep the schedule deterministic. Handing AI both the words and the timing is where instability creeps in.

you don't write tickets anymore. you point. 👆 annotate the live product, describe the fix, and a supervised agent ships it through your own github — same day. the annotate-to-ship workflow, explained → agentation.fr/annotate-to-sh…

the terminal walked so the workshop could run. you don't "code" here — you point, you sketch, you approve. on the live product. no black screen, no cryptic flags. this is what work is supposed to feel like.

No way opus 5 did design this in less than 2hours !

genuine question: who is STILL working in a terminal in 2026? staring at a wall of text to move one button. it's atrocious. building should be visual — point at the thing, watch it change. this is the workshop.

Say bye bye to vibe coding…

Forgot to send a prompt before living the office ? Continue on the transport ! That the app I Vibe create for @agentationai ! Nice liquid glass usage 🫠

Make GitHub simple for vibe creators. The simplicity of vibe coding finally happening 🫠

Fabeling all weekend until the 7 ❤️