Home / Science / AI Claims Proof of Math's Toughest Problem
AI Claims Proof of Math's Toughest Problem
20 Nov
Summary
- Mathematician Levent Alpöge claims AI Claude helped prove Riemann Hypothesis.
- Proof uses Lean software; human-readable version to be published later.
- Riemann Hypothesis is a $1 million Millennium Prize problem from 1859.
Mathematician Levent Alpöge has claimed a proof for the Riemann Hypothesis, a renowned unsolved mathematical problem, stating that Anthropic's AI model, Claude, played a significant role in the discovery. Alpöge released a four-page statement detailing the potential proof and associated Lean software files, which are designed to verify each step of the argument.
He indicated that a complete human-readable proof would be prohibitively long, estimating a machine-generated draft at over 250,000 pages. Alpöge plans to work on understanding and properly documenting the argument, acknowledging that Claude significantly aided in the final crucial steps.
The Riemann Hypothesis, first posed in 1859, is one of seven Millennium Prize problems with a $1 million award. Its resolution has broad implications for number theory. Alpöge's work builds on prior mathematical techniques, incorporating a substitute for a calculus method that proved insufficient for whole numbers.
This announcement follows a prior dispute in September involving Alpöge and OpenAI over a different mathematical problem. He noted a constraint to respect regarding the timing of his announcement, suggesting coordination with Anthropic leadership.