Anthropic researcher defends 50% doom estimate as critics call it fraud

TinfoilTricorn · x · 2026-09-11

Anthropic researcher Geoffrey Irving explains why he still puts p(doom) at 50%: despite years of thought, model uncertainty remains large. He previously converged on "10–90%, but I don't feel calibrated within that range" and decided the honest move was to take the mean. A critic fired back that the claim has "no proof, no math, no method to test" and should be sued for fraud — capturing ongoing controversy over falsifiability of doom estimates.

Related event: AI safety researcher Geoffrey Irving puts superintelligence extinction odds at ~50%(6 posts)→

Original post →

More from AGI Musings

AGI Musings channel →