AI Solves 4 Math Problems That Stumped Experts for Years

When mathematicians Dawei Chen and Quentin Gendron hit a wall in their research five years ago, they had to publish their work as a conjecture instead of a proven theorem. Last month, an AI solved their problem overnight.

Chen spent hours trying to get ChatGPT to crack the puzzle with no success. Then he met Ken Ono, a mathematician who had just joined Axiom, an AI startup founded by his former student Carina Hong.

The next morning, Ono handed Chen a complete proof generated by Axiom's math-solving AI, AxiomProver. The AI discovered a connection to a 19th century numerical phenomenon that every human expert had missed.

"Everything fell into place naturally after that," says Chen, who worked with Axiom to write up the proof, now posted publicly for other mathematicians to review.

AxiomProver has solved three other problems in recent weeks that stumped experts for years. One involves formulas first discovered in the notebook of legendary Indian mathematician Srinivasa Ramanujan over 100 years ago.

The system combines large language models with specialized AI trained to reason through math problems and verify its own answers using a mathematical language called Lean. This means it's not just searching existing research but creating genuinely new solutions.

"Even as someone who's been watching the evolution of AI math tools closely for years, I find this pretty astounding," says Scott Kominers, a Harvard Business School professor. "It's not just that AxiomProver solved these problems fully automated, but also the elegance and beauty of the math it produced."

The Ripple Effect

The technology could reach far beyond abstract mathematics. The same verification techniques could help develop software that's more secure against cyberattacks by proving code is reliable and trustworthy.

"Math is really the great test ground and sandbox for reality," says Hong, Axiom's CEO. "We do believe that there are a lot of pretty important use cases of high commercial value."

Ono hopes AxiomProver will reveal something fundamental about how discoveries happen. "I'm interested in trying to understand if you can make these aha moments predictable," he says.

Chen isn't worried about AI replacing mathematicians. "Mathematicians did not forget multiplication tables after the invention of the calculator," he says. "I believe AI will serve as an intelligent partner, opening up richer and broader horizons for mathematical research."

The breakthroughs suggest we're entering an era where human creativity and artificial intelligence work together to solve problems neither could crack alone.