California proposes kill switch letting officials shut down AI models before court review
mark_k · x · 2026-11-20
New recommendations from Governor Newsom would let California officials order an immediate shutdown of any AI model they believe poses an imminent catastrophic threat. The order takes effect right away, before any court review — companies can only challenge it afterward.
Additional measures include mandatory state-approved testing at every training stage, 48-hour incident reporting, and the power to cap AI used to accelerate AI research. The author argues the last one is the most alarming: using AI to speed up AI research is how progress compounds, and California wants permission to slow it. It's still a proposal, but Governor-elect Becerra says he'll push it into law quickly.
Related event: California Proposes State Power to Shut Down Dangerous AI Models(3 posts)→
More from AGI Musings
- Judging on limited info is hardest, so people defer to their in-group — gabriel1 · 2026-11-20
- Perception runs on "others think it matters", not independent judgment — gabriel1 · 2026-11-20
- Claude builds its own tooling mid-task, cutting overnight jobs to one hour — DaniyarQQQ · 2026-11-20
- RSI ban is a 'Trojan horse for an AI ban', argues engineer: any competent engineer can already self-improve agents — nptacek · 2026-11-20
- Pedro Domingos on data center backlash: no one who has looked at the data opposes them — pmddomingos · 2026-11-20
- UChicago study: AI-generated research ideas converge, rarely challenging assumptions — _akpiper · 2026-11-20