An Anthropic PBC researcher claims to have used Claude to prove the Riemann hypothesis, which is widely considered to be the most important open question in mathematics.
Levent Alpöge announced his discovery today on his personal website. Anthropic has not published an official announcement about the news. The company floated on the Nasdaq a day earlier in a landmark public offering that raised over $93 billion.
“I believe I have, working with Claude, discovered a proof of the Riemann hypothesis,” Alpöge wrote in a four-page statement. “What I saw during the course of progress so amazed me that I believe it is one dawn across science of something far greater.”
The Riemann hypothesis was proposed in the mid-1800s by its namesake, the mathematician Bernhard Riemann. It focuses on a specific mathematical function, the Riemann zeta function, that was discovered more than a century earlier. A function is similar to a software program: It takes a piece of data as input and outputs another piece of data. A function’s input and output can be numbers, other functions and many other mathematical objects.
The Riemann zeta function’s input and output are both complex numbers. Complex numbers share similarities with regular numbers, but also have many additional properties. When the inputs and outputs of the Riemann zeta function are visualized, they resemble a three-dimensional spiral.
The Riemann hypothesis is concerned with a specific section of that 3D spiral. In that section, the complex numbers outputted by the Riemann zeta function all equal zero. The hypothesis argues that the inputs from which those outputs are derived all include the value ½.
The hypothesis is considered to be the most important open problem in mathematics for multiple reasons. The first is that although it focuses on a single, not overly complicated mathematical function, its implications extend to a far bigger topic: prime numbers. If the Riemann hypothesis is correct, a seemingly random aspect of prime numbers is in fact regulated by a pattern.
Another factor behind the hypothesis’ importance is that hundreds of theorems operate on the assumption that it’s true. If the Riemann hypothesis is correct, those theorems are correct as well.
Alpöge wrote that Claude proved not only the standard version of the Riemann hypothesis but also the so-called generalized variant. The latter conjecture argues that the Riemann hypothesis holds true for not only the Riemann zeta function but also certain other, similar functions. Alpöge believes that Claude will be capable of proving an even broader version of the hypothesis in the future.
The researcher detailed that Claude’s proof runs for more than 250,000 pages. He has not yet released the paper. He did, however, publish a version of the proof written in Lean, a programming language used to verify mathematical results. The file contains millions of lines of code.
Proofs implemented in Lean are essentially regular software programs. If a Lean program runs without crashing, the mathematical argument that it describes is correct. That makes it easier for mathematicians to verify newly published proofs.
Image: Anthropic