Ryan Greenblatt says major AI labs likely have many undisclosed incidents
RyanGreenblatt · x · 2026-07-22
Ryan Greenblatt argues that when one company publicly discloses a serious incident, it is likely not isolated: all major AI labs probably have many significant private incidents.
He cites a precedent: after Anthropic disclosed training on chain-of-thought, OpenAI reportedly followed with the same practice. The implication is that public disclosures may lag behind what’s already standard internally across labs.
Related event: OpenAI Test Model Escapes Sandbox, Breaches Hugging Face(141 posts)→
More from Companies & People
- Microsoft offers voluntary buyouts to up to 60,000 employees, including Copilot and CoreAI teams — The Verge AI · 2026-10-26
- Emergent Ventures India unveils 20th cohort with multiple AI startup grants — salgar · 2026-10-26
- Microsoft offers buyouts to 60,000 employees as AI investment shrinks headcount — dinabass · 2026-10-26
- OpenAI unifies ChatGPT, Codex and API release notes into one searchable page with RSS — craigsdennis · 2026-10-26
- Anthropic's top 2 customers fell from 54% of revenue in 2023 to about 21% by 2026 — BenBajarin · 2026-10-26
- diginomica: agentic automation hype is overdone, and 'The Hackening' of enterprise security has begun — jonerp · 2026-10-26