Ex-OpenAI/Anthropic pretraining researcher resigns, says labs are gambling on superintelligence

Miles_Brundage · x · 2026-09-10

Jacob (@hilbertspaess), who spent three years doing pretraining research at both OpenAI and Anthropic, resigned from Anthropic, saying neither company is acting responsibly and both are "racing straight to self-improving superintelligence and gambling with our lives."

Former OpenAI researcher Tomasz Korbak amplified the thread, vouching for Jacob as a thoughtful researcher and agreeing with the core claim: neither Anthropic nor OpenAI are on track to solve alignment to a degree sufficient for shipping superintelligence — "we need to slow down." Another public critique of frontier labs' safety commitments from within the safety community.

Related event: Anthropic researcher quits over safety fears as alignment lead puts extinction odds above 10%(133 posts)→

Original post →

More from AGI Musings

AGI Musings channel →