Superhuman mathematical problem solving before 2030, assuming no AGI yet?
💡 What the odds say
The market puts this at about a 58% chance — more likely than not.
No money — just record your call and see if you were right. Yes is at 58% right now.
The market sees a better-than-even chance of AI solving a superhuman math problem before 2030 without AGI, driven by a string of recent breakthroughs in automated theorem proving and mathematical reasoning, though skeptics argue these successes remain narrow and don't match the creativity required for truly open problems.
📊 Base rate: Since 2019, AI systems have solved only a handful of long-standing math problems (e.g., the cap set problem in 2023, the moving sofa problem in 2024), with most major open problems still unsolved, giving a prior well below 50% for a breakthrough within a few years.
What's driving it
- • A July 23 report that a basic AI prompt cracked a decades-old math problem (New Scientist, Jul 23) suggests lower barriers to discovery, likely boosting Yes odds.
- • The WSJ (May 29) covering an 80-year-old math problem solved by AI reinforces momentum and public attention, swaying traders toward Yes.
- • Google's claim that its Gemini 3 agent Aletheia completed autonomous mathematical research (GIGAZINE, Feb 12) signals a major capability leap, supporting the Yes case.
- • Disagreement over AGI's definition (Medium, Mar 3) injects uncertainty into whether a solution would count as 'superhuman' without AGI, potentially leading to No bets.
The case for YES
- • Successive headline-grabbing feats—an 80-year problem solved (WSJ, May 29), a decades-old problem cracked by a basic prompt (New Scientist, Jul 23)—show that AI can already tackle problems long considered hard, and further improvements may yield a truly 'superhuman' result.
- • Google's Aletheia agent demonstrated autonomous mathematical research (GIGAZINE, Feb 12), implying that without needing AGI, very capable specialized systems can be built for pure mathematics.
- • The argument that 'superintelligence is already here' (Noahpinion, Mar 2) frames current AI as effectively superhuman in specific domains, making the threshold lower than many assume.
The case for NO
- • Despite recent wins, each solved problem was highly specific and often narrow; IEEE Spectrum (Feb 26) questioned how far AI can go, noting that open problems require genuine insight and abstract creativity that current systems lack.
- • Disagreement on AGI (Medium, Mar 3) could lead to a scenario where even a strong solution is deemed insufficient unless AGI is present, raising the bar for what counts as 'without AGI yet.'
- • The 58% odds may overprice recent hype; many prior AI math breakthroughs (e.g., the cap set problem) did not lead to a cascade of superhuman results within their timeframes, and the 2030 deadline is tight.
What to watch
- • If a major mathematical institute or journal (e.g., Annals of Mathematics) publishes a result claimed by an AI system before end of 2026, odds would spike toward 80%+ Yes, as peer review confers legitimacy.
- • A prominent critique from a Fields Medalist or leading mathematician in Q3 2026 that AI results are trivial or misclassified could drop odds to 40% or below.
- • Release of a new reasoning benchmark, such as the AI Mathematical Olympiad results in late 2026, if showing stagnation, would shift odds toward No.
AI-generated · grounded in recent news + odds · informational only, not advice. Verify on the source platform.
Data from Manifold’s public API, for informational purposes only. PredictPal is not affiliated with any platform and does not facilitate trading.
Discussion
Loading…
How it resolves
Resolved by whoever created the market, at their discretion per the question's description. It's play-money (Mana) and not tied to an official source — treat it as a community forecast.
ⓘ A market settles under its own written rules, which can lag what looks decided in the news — so the price may not move to 100% the moment an outcome seems obvious.
View the official rules on Manifold ↗Related markets
Will over 100,000 people be conceived with the help of advanced embryo selection techniques by 2030?
The market sees advanced embryo selection as still niche and ethically fraught, but recent coverage of genetic optimization startups and IVF's expanding role in family planning suggests the technology is gaining momentum, making the 100,000-conception threshold plausible if regulatory and adoption barriers fall.
Yes ≈ 26% chance