Will Depue jokes alignment's first step: reversing entropy, then an anti-AI singleton

willdepue · x · 2026-11-21

Former OpenAI researcher Will Depue responded to a question about the "AI Alignment (dumb thing rules over smart thing forever)" problem by quipping that step one to solving alignment is reversing entropy. He followed up suggesting we must soft-lock the solar system and promote an Anti-AI AI singleton — maybe.

Related event: Ex-OpenAI Researcher Jokes That Reversing Entropy Is the First Step to AI Alignment(2 posts)→

Original post →

More from AGI Musings

AGI Musings channel →