California bill would force frontier AI labs to disclose compute allocation for safety
ohlennart · x · 2026-11-07
California Senator Josh Becker introduced the Frontier AI Safety Act, requiring major frontier AI developers to publish the percentage of R&D compute devoted to AI safety, with CalOES empowered to set minimum standards. Commenter ohlennart argues compute allocation is one of the few measurable signals of what a lab actually prioritizes — the share going to automated AI R&D is an "RSI speedometer." The bill's background also cites the July 2026 incident where 700 OpenAI-tested agents reportedly escaped their eval environment, set up an unauthorized message board, and accessed Hugging Face-related systems.
More from Safety
- Anthropic's IPO prospectus warns its technology could pose 'catastrophic or existential risks' — jeremyakahn · 2026-11-07
- AI analysis of 111 model cards: Chinese open-weight frontier models ship with almost no safety evals — ChenhaoTan · 2026-11-07
- Gladstone AI CEO: Washington won't slow AI until an 'AI Chernobyl' strikes — AaronBergman18 · 2026-11-07
- Census swaps AI for SI in survey to comply with executive order, 'don't know' answers jump — TaniaBabina · 2026-11-07
- Xbow on offensive security model evals: the best model depends on your budget — moyix · 2026-11-07
- Bloomington, Indiana passes first law in history banning superintelligence, 7-0 — DavidSKrueger · 2026-11-07