Author beats experts in forecasting FrontierMath accuracy
peterwildeford · x · 2026-08-29
Peter Wildeford shared that his personal forecast on AI capabilities outperformed expert groups. Regarding the max accuracy on FrontierMath by end of 2025, he predicted 38%, while experts, superforecasters, and the public averaged 31%, 30%, and 27% respectively. The actual result was 40.7%. This question comes from the newly launched Longitudinal Expert AI Panel (LEAP).
Related event: Forecasters beat expert panel on FrontierMath prediction accuracy(2 posts)→
More from AGI Musings
- "An LLM doesn't understand anything" gets the perfect parody reversal: neither do you — VoidStateKate · 2026-10-16
- Nina Schick: Europe's Policy Circle Still Doesn't Believe FSD, Orbital Compute Are Real — NinaDSchick · 2026-10-16
- Meta's Superintelligence Lab Runs 200+ Direct Reports—A Lesson for Flat Multi-Agent Systems — peterjliu · 2026-10-16
- Andy Masley lays out his version of effective altruism, pluralism over utilitarian math — RobbWiller · 2026-10-16
- We're forgetting how much of a miracle GPT-3.5-level intelligence was, and it may cost us — EigenGender · 2026-10-16
- Ex-Anthropic's McDonagh Builds Compute Intelligence Agency to Study Frontier AI Costs — McDonaghMatthew · 2026-10-16