OpenAI fires three key safety employees who drove frontier pacing and monitorability work
NathanpmYoung · x · 2026-10-08
Per WSJ, OpenAI fired three safety employees central to its biggest recent safety efforts: 'pacing the frontier', the METR-Redwood Hugging Face investigation, and cross-lab monitorability standards. Critics note they both pushed the company internally and defended it against unfair attacks. The fired employees are calling for external safety audits and preserved ability to monitor increasingly capable models.
More from Companies & People
- Why Apple, Controlling iOS, Still Fails at the AI Agent It's Best Positioned to Build — Yuchenj_UW · 2026-10-08
- SSI tops Paraform's Talent Density Index, edging out Anthropic and OpenAI — FinanceYF5 · 2026-10-08
- AGI House at OpenAI DevDay 2026: from models that answer to systems that act — agihouse_org · 2026-10-08
- How an unknown AI startup hit a $2.5B valuation in 13 days — abhijithneil · 2026-10-08
- AGI House and Coframe host a build day for the self-adapting 'living internet' — agihouse_org · 2026-10-08
- Jensen Huang and Satya Nadella Unveil Partnership to Reinvent Windows for Personal AI Agents — NVIDIAAI · 2026-10-08