technology

Alpoge linked to Lean code he said could verify findings

An Anthropic mathematician says Claude helped craft a proposed solution to the Riemann hypothesis in six weeks, a claim peers are working to confirm.

anthropic, claude, riemann, san francisco, if alpöge, mathematician, hypothesis, anthropic mathematician

Key facts

  • The hypothesis deals with prime numbers, whole numbers greater than 1 that can be divided only by themselves and 1. Riemann questioned how often…
  • Alpoge linked to computer code written in a language called Lean that he maintained could verify the findings. Mathematicians then began working to…
  • If the claim is ultimately confirmed, the proof could mark another breakthrough in the use of AI on complicated math problems. The result has not…

The hypothesis deals with prime numbers, whole numbers greater than 1 that can be divided only by themselves and 1. Riemann questioned how often prime numbers appear and how they are distributed along the number line.

Alpoge linked to computer code written in a language called Lean that he maintained could verify the findings. Mathematicians then began working to see whether the proof could be confirmed.

Alpoge said Claude helped develop the solution over about six weeks. His note was published one day after Anthropic went public on the Nasdaq stock market index.

If the claim is ultimately confirmed, the proof could mark another breakthrough in the use of AI on complicated math problems. The result has not been confirmed.

The approach struck many mathematicians and observers as strange. Accounts of the work said a near-trillion-dollar private company spent $15 million to employ 10,000 AI systems on the problem.

Mathematicians have expressed uneasiness about extremely lucrative private companies spending millions of dollars so technology can solve such problems. Others have voiced discomfort with relying on AI to solve math problems.

He developed a formula for estimating how many prime numbers exist under a specific number. Mathematicians have been attempting to solve the problem for years.

An institute has offered $1 million for each problem solved in that set of questions.

Alpoge was working on a separate problem for Anthropic when OpenAI made its announcement. That work was distinct from the Riemann claim now under review.

Dig deeper

What is the Riemann hypothesis?

It is a 19th-century conjecture by Bernhard Riemann on how prime numbers are spaced, linked to the zeros of the zeta function. It remains unproved and is one of the Millennium Prize problems, each worth $1 million.

What is Lean?

Lean is a proof-assistant language in which mathematicians write arguments so a computer can check every step, which is why the linked code matters only if others confirm it really verifies the claim.

What is Claude?

Claude is an artificial-intelligence system made by Anthropic, the model Alpoge said helped develop the proposed solution over about six weeks.

What is Anthropic?

Anthropic is the company that builds Claude. Alpoge's note appeared one day after the company went public on the Nasdaq.

Why does this matter?

A confirmed proof would be a landmark on primes and a major case of AI used on hard mathematics, but the result is unverified and has raised unease about costly private AI efforts and reliance on machine-written arguments.

What happens next?

Mathematicians are checking whether the Lean code actually verifies the findings. Until that succeeds, the claim is not confirmed.

Lire en français