Scott Alexander challenges AI skeptics to name a Metaculus score AIs will never hit
slatestarcodex · x · 2026-10-26
Scott Alexander (slatestarcodex) publicly challenged AI skeptics like Sayash Kapoor and Arvind Narayanan to make a falsifiable prediction: name a specific Metaculus score that AIs will never achieve, or a specific point where the trend of AI's forecasting scores rising relative to humans will break. The move pushes the "AI forecasting is overhyped" debate toward testable claims.
More from AGI Musings
- Harnesses for output variance may matter as much as model training — Afinetheorem · 2026-10-26
- Vinge's 1993 Singularity essay resurfaces: technological unemployment and tuning surprises — akbirthko · 2026-10-26
- LessWrong post argues FDT is difficult-to-impossible to specify due to logical counterfactuals — jessi_cata · 2026-10-26
- Bitter Lesson misread: scaling won't brute-force biology, researcher argues — anshulkundaje · 2026-10-26
- Brynjolfsson: AI benchmarks soar but economic productivity gains lag like early electricity — luisdans · 2026-10-26
- Venkatesh Rao's long essay argues AI panic is a form of "Cognitive Mercantilism" — ianand · 2026-10-26