Ex-OpenAI/Anthropic Pretraining Researcher Quits, Says Labs Are 'Racing to Self-Improving Superintelligence'

sjgadler · x · 2026-09-11

A researcher who spent three years doing pretraining research at both OpenAI and Anthropic announced their resignation from Anthropic with a public accusation:

The researcher promised more thoughts to follow.

Related event: Anthropic Researcher Quits Warning of AI Doom, Sparking PR-Coordination Backlash(238 posts)→

Original post →

More from AGI Musings

AGI Musings channel →