AI Safety Researchers Clash Over Lab Independence and Talent Tradeoffs
joshua_saxe · x · 2026-09-14
AI safety researcher Joshua Saxe and Halvar Flake argued over whether AI safety institutes can truly be independent from labs. Saxe says demanding independence ignores a real tradeoff: finding people who deeply understand LLMs and can do safety work, with funders, while being socially distant from labs is nearly impossible.
Flake strongly disagreed, arguing the nonprofit has built an exceptionally talent-dense safety team and that it "saved us" amid the dissolution of US CAISI under Trump. He concedes its security expertise is weak, but says security professionals should join rather than criticize.
Related event: Anthropic's Safety Plan Sparks METR Independence Backlash(31 posts)→
More from Companies & People
- Hugo Bowne launches Master Agentic Data Science workshop with PyMC's Thomas Wiecki — hugobowne · 2026-09-14
- Anthropic CEO Dario Amodei: the industry 'lied for too long' about AI risks — Polymarket · 2026-09-14
- Chart shows hyperscalers went on a CAPEX spree after DeepSeek — and Kimi broke the logic — StewartalsopIII · 2026-09-14
- CS enrollment trend sharply reverses, per FT — CSProfKGD · 2026-09-14
- After 20 years in Army Cyber Command, digital forensics lead joins AI evals group METR as advisor — ajeya_cotra · 2026-09-14
- Yoav Goldberg translates AI industry speak: 'third-party evaluators' are spies, 'pacing' means stop spending — yoavgo · 2026-09-14