Analysis · Security
An official package name is not proof of an official release. Here is how hash pins, PyPI attestations and a .pth startup audit verify a Python package before pip runs it.
✎By Guillaume · Tech
From October 25 to November 9, 2026, 37 releases of my Git client: a new theme, blame inside the diff, GitLab, Bitbucket and Gitea, Jira sprints in the analytics, profiles, local dictation, and agents you can message while they work.
Analysis · Security
Marketplace badges and star counts won't stop a compromised update. Here is how to check a VS Code extension's publisher, diff its versions, grep its code and watch what it writes to disk.
Brief · AI
Bluesky expanded Attie so users can make games, polls and dashboards with AI and share them. Bluesky CEO Toni Schneider discussed the launch at TechCrunch Disrupt.
Brief · AI
Bluesky is letting anyone use its experimental Attie AI platform to build websites, games and quizzes that draw on the social network's audience.
Analysis · AI
MCP is the JSON-RPC protocol that lets AI hosts discover and call tools. Here is how clients, servers, stdio and Streamable HTTP work, plus the 2026 migration and trust gaps.
Analysis · AI
Haiku 5.5's 1M-token window is cheap to try and expensive to repeat. Here are the tier math, a break-even formula for long context vs RAG, and what remains unconfirmed.
Analysis · Tech
The command line has been stagnant for decades. Now, native GPU-accelerated emulators like Ghostty and Warp are bringing the 50-year-old CLI into the modern silicon era.
Analysis · AI
Transformers conquered the world, but their quadratic compute penalty is unsustainable. Here is why State Space Models like Mamba process data linearly and threaten the attention mechanism.
Analysis · AI
Gemini 1.5 Pro can ingest 2 million tokens at once. Here is why the era of complex vector databases and chunking might be coming to an end.
Analysis · Tech
VS Code has held a monopoly on developers for nearly a decade. Now, AI-native forks like Cursor and Windsurf are ripping up the editor playbook and threatening Microsoft's crown.
Analysis · AI
Developers are tired of API costs, privacy risks, and rate limits. Ollama and Llama 3 just made running local AI models offline as easy as using Docker.