OpenAI safety report lead resigns, warns 'the time for trial and error is over'
ryan_t_lowe · x · 2026-10-04
David Robinson, one of OpenAI's longest-tenured employees who oversaw safety reports on 12 frontier model launches, resigned this week and penned an Atlantic essay: "The time for trial and error is over."
Former OpenAI safety researcher jachiam0 vouched for Robinson as a sober, non-ideological thinker (not EA-aligned), and endorsed his core critique: professional safety engineering practices from other fields haven't taken root in AI safety. He argues the field has hit a phase shift — organizational practices sufficient for 2025-level models are no longer adequate to prevent serious incidents, and every lab must uplevel.
Robinson's essay calls for cultural change beyond specific rules or new laws; Miles Brundage also endorsed it, expressing regret over pushing "iterative deployment."
Related event: OpenAI Safety Veteran David Robinson Resigns, Blasting Company Culture(17 posts)→
More from Companies & People
- Rumor: Microsoft cut internal Claude spend by 33%, slashed per-employee token budget from $100k to $10k/month — ns123abc · 2026-10-06
- Mercor CEO on 60 Minutes: AI is 'changing human work', 100k freelancers now train models — briannekimmel · 2026-10-06
- AI Labs' Takeover of Mathematics Sparks Backlash From Mathematicians — The Verge AI · 2026-10-06
- Paris community hackathon ahead of OpenAI DevDay: Oct 24-26, winners get 10 DevDay tickets — paw_lean · 2026-10-06
- PyTorch Conference: IBM to keynote Spyre Accelerator and distributed inference work — PyTorch · 2026-10-06
- David Sacks: Elon Musk was still working past 2 a.m. while running five companies — XFreeze · 2026-10-06