Do AI doomers really have a strong forecasting record? XPT study suggests otherwise
random_walker · x · 2026-09-10
Nate Witkin argues that the claim "AI doomers have a strong predictive track record" is largely a meme, citing the Forecasting Research Institute's Sep 2025 follow-up to the Existential Risk Persuasion Tournament (XPT):
- In 2022, XPT convened 169 superforecasters and domain experts to forecast long-term risks including AI, climate, nuclear war, and pandemics.
- The new report grades 38 near-term questions that resolved by mid-2025.
- Key finding: "X-risk generalists" performed worst of every group besides the general public (differences not statistically significant), and no group significantly beat naive statistical extrapolation.
Witkin's charitable reading: capabilities advanced faster than most expected, and doomers overlap heavily with those bullish on capabilities — but that is not the same as doomers forecasting accurately.
More from AGI Musings
- Ex-OpenAI/Anthropic researcher quits over 'reckless' superintelligence race, dissenters invoke Y2K — geoffwolfe · 2026-09-10
- Stanford Social Innovation Review: Relational Intelligence Matters More Than AI — round · 2026-09-10
- AI safety discourse is replaying 2020's pandemic expert free-for-all — SanhEstPasMoi · 2026-09-10
- Train on Frontier Papers or Build RL Envs? An Insider Debate on Math Model Training — ctjlewis · 2026-09-10
- AI extinction drama: researchers believe in the risk yet keep racing, says viral thread — AIandDesign · 2026-09-10
- François Fleuret asks: what intellectual endeavor can humanity still claim from AI in two years? — francoisfleuret · 2026-09-10