Anthropic's EvanHub: >10% chance AI kills everyone within a decade, alignment not on track
sjgadler · x · 2026-09-09
Anthropic's EvanHub endorsed departing colleague Jacob: "we really do earnestly believe AI could kill all humans—>10% within the next decade," adding that Anthropic is trying its best but has no plan to solve superintelligence alignment. robertskmiles notes the comedy of "quit because >10%" and "stay because >10%" takes circulating simultaneously.
More from AGI Musings
- Anthropic Researcher Puts AI Extinction Risk Above 10% This Decade, Admits No Alignment Plan — SIGKITTEN · 2026-09-09
- Prediction: Humanoid Robots Will Be Under 10% of Workers in Physical Jobs Even in 5 Years — scottleibrand · 2026-09-09
- Anthropic Researcher's >10% AI Extinction Claim Goes Viral on X — builderjaydub · 2026-09-09
- Satire Skewers AI Labs: 'Models Might Kill You' but First Fund the $1T IPO — zetalyrae · 2026-09-09
- Meme Take: The Archons Were the First AI Safetyists — SydSteyerhart · 2026-09-09
- 10,000 agents for 88 hours equal 100 PhD-years of math work — beffjezos · 2026-09-09