[01]
OpenAI, Anthropic and Google are working to create an AI standards body
The three AI labs are forming a voluntary, FINRA-style self-regulatory standards body for frontier AI safety.
// daily-digest · mon 21 sep 2026 · morning refresh · last 72h focus
Ten items on AI policy, AI lab competition, coding-agent tooling, Rails World, and a genetics wildcard.
— refreshed for you, in 10 items.
The three AI labs are forming a voluntary, FINRA-style self-regulatory standards body for frontier AI safety.
Anthropic is considering a new model release as GPT-6 Astra gains enterprise share, days after urging an industry slowdown.
AMD shares crossed a trillion-dollar valuation for the first time, driven by AI data-center demand.
Claude Code now reads AGENTS.md when no CLAUDE.md exists, aligning with the cross-tool agent-instructions standard.
ChatGPT's built-in desktop browser now supports installing and pinning Chrome extensions such as 1Password.
Sean Goedecke argues shipping frequently beats endless polishing, framing perfectionism as an emotional hurdle.
Details on the free livestream for DHH's Rails World 2026 keynote in Austin, plus this year's sponsors.
Noah Smith weighs a Kennan-style containment strategy against China and doubts American political unity can sustain it.
Andrew Sharp argues AI-doom discourse has grown exaggerated and self-reinforcing even as some real risks remain.
A University of Bristol team identified the genetic cause of the AnWj blood antigen, defining new blood group MAL.