AI to Break a Thousand Math Conjectures? Predicting a Decade of Counterexamples

jasondeanlee · x · 2026-07-21

The recent AI-assisted discovery of a counterexample to the Jacobian conjecture has sparked industry interest in AI's potential in fundamental sciences.

Commentators predict that while monumental problems like the Riemann hypothesis or P vs NP will likely remain untouched, there are numerous smaller conjectures in mathematics. Historically, the limiting factor in disproving them hasn't been a lack of problem-solving ability, but rather researchers' over-investment in proving them right.

With AI's intervention, breaking down old assumptions will become a powerful way to open up new fields. Over the next decade, AI could uncover a thousand little broken conjectures, which could be a massive deal and reframe much of our current knowledge.

Original post →

More from AGI Musings

AGI Musings channel →