Superforecasters Put AI Solving a Millennium Problem at 1.7% — and Keep Underestimating AI
emollick · x · 2026-09-24
Ethan Mollick notes that when asked in September 2025, top superforecasters gave only a 1.7% chance of AI resolving a Millennium Prize Problem by September 2026; optimistic industry experts said 4.6%.
The same forecasters also badly underestimated AI lab revenue, suggesting systematic underprediction of both AI capability and commercial progress.
Related event: Experts Have Consistently Underestimated AI Progress, Study Finds(4 posts)→
More from AGI Musings
- Models Scale Daily, But Human Comprehension Isn't Keeping Up — _jaydeepkarale · 2026-09-24
- Economists do the math: society may be underinvesting in AI safety by 30x or more — JimCripe · 2026-09-24
- New study: AI-adopting companies mostly hire for senior roles while cutting entry-level jobs — mariolefebvre · 2026-09-24
- AI's real danger isn't laziness — it's endless busyness, argues ryolu_ — ryolu_ · 2026-09-24
- A cloud PC just to manage email and calendar? AI bubble alarm bells — oran_ge · 2026-09-24
- Few investors are pricing in the consumer-vs-enterprise agent arms race, says investor — robleclerc · 2026-09-24