LEAP expert panel puts just 10% odds on AI cracking a Millennium Problem by 2027
Afinetheorem · x · 2026-09-09
The Forecasting Research Institute released Wave 2 of its Longitudinal Expert AI Panel (LEAP) report on AI for Science, aggregating forecasts from 277 experts, 58 superforecasters, and 1,022 members of the public collected between Aug 18 and Sep 15, 2025.
Key results:
- Millennium Prize Problems: Experts put only 10% probability on AI solving or substantially assisting with a Millennium Prize Problem by 2027; one expert cites the Clay Institute president's June 2025 claim that "we're very far away from AI being able to say anything serious about any of those problems."
- Diffusion of AI across sciences: forecasts for the share of Physics, Materials Science, and Medicine publications that will be 'AI-engaged' in 2030.
- Drug discovery: forecast shares of US recently-approved drug sales from AI-discovered drugs in 2027, 2030, and 2040.
- Electricity consumption: forecast US power consumption for training and deploying AI in 2027, 2030, and 2040.
- Cognitive limitations: what share of experts will agree, by end of 2030, that various shortcomings are serious limitations of state-of-the-art AI.
The report offers a systematic quantitative baseline for expectations about AI's scientific capabilities.
More from AGI Musings
- 'Half the sites I try via agents block access' — the anti-bot paradox holding back the agentic economy — XFreeze · 2026-10-27
- US chases AGI dominance while China, EU and ASEAN treat AI as public infrastructure — BubblyExtent1693 · 2026-10-27
- Lab monkeys at $30-80K each: a 4-year breeding cycle is the real preclinical drug bottleneck — zebird0 · 2026-10-27
- CJR: journalists should stop hyping AI's supposed humanity — just say 'rogue' — ShakeelHashim · 2026-10-27
- PhD applicant emails up 4x YoY due to AI — writing without AI now makes you stand out — AI4Code · 2026-10-27
- Weil's collapse shows BigLaw dies from eroding belief, not a broken business model — jkubicki · 2026-10-27