Anthropic researcher: I earnestly believe AI could kill all humans, >10% odds within a decade
JFPuget · x · 2026-09-09
Anthropic alignment researcher Evan Hubinger (EvanHub) publicly stated that he earnestly believes AI could kill all humans, personally estimating the probability at >10% within the next decade. He said Anthropic is trying its best, but there is not yet a plan to solve alignment for superintelligence, and it's not clearly on track. The retweeter added a sarcastic take that people delegating all thinking to LLMs risk becoming 'brain dead'.
More from AGI Musings
- "Believing in 10% doom is more dangerous than doom itself": AI labs' p(doom) debate reignites — kuza55 · 2026-09-09
- Anthropic launches economic futures explorer: fast-AI scenarios hit knowledge-worker wages — eherrerosj · 2026-09-09
- Anthropic's first economics paper models transformative AI scenarios with 15% annual GDP growth — soumitrashukla9 · 2026-09-09
- Investor: the existential threat to app-layer software is the emerging agent layer, not stock swings — matt_slotnick · 2026-09-09
- Nat Lambert: AI labs are 'brainwashing' people into evidence-free doom beliefs — natolambert · 2026-09-09
- Matt Slotnick: nothing in 3 months invalidated the agent-layer threat to app software — matt_slotnick · 2026-09-09