Advertisement

Anthropic Mathematician Claims AI Helped Solve 167-Year-Old Math Puzzle

An Anthropic researcher says he has proved the famous Riemann Hypothesis with help from Claude. The claim, if verified, could mark a historic moment for mathematics and AI.

Anthropic Mathematician Claims AI Helped Solve 167-Year-Old Math Puzzle
File photo

A mathematician and Claude took on a 167-year-old problem and apparently solved it. Levent Alpoge, a mathematician working at Anthropic, the company behind Claude says he has discovered a proof of the Riemann Hypothesis with Claude's help. The claim was made in a four-page statement, posted on his website.

"I believe I have, working with Claude, discovered a proof of the Riemann Hypothesis."

Riemann Hypothesis is considered one of the most famous unsolved problems in mathematics, first proposed by German mathematician Bernhard Riemann in 1859.

So what exactly is the Riemann Hypothesis? Prime numbers are numbers divisible only by themselves and one (2, 3, 5, 7, 11...). They appear unpredictable as we move along the number line. But mathematicians believe there is a deeper mathematical pattern governing how these prime numbers are distributed.

The Riemann Hypothesis essentially proposes a way to understand and predict that pattern with extraordinary precision. If proved, it would help mathematicians understand just how regularly or irregularly prime numbers are distributed, and strengthen a vast body of mathematical results that currently depend on this assumption.

The hypothesis is also one of the seven Millennium Prize Problems, each carrying a $1 million prize for a solution that meets the required standards.

How did Claude help? 

Alpoge says he had spent years preparing to tackle the problem and he describes working with Claude over recent months. He says he guided the AI towards the right mathematical approach, but Claude made extraordinary progress towards the finish line.

The mathematician specifically credits Claude with contributions that went beyond his own expertise. "No, Claude flew, and danced, and sparkled near that finish line," he said.

"I guided us to the right situation analogous to that toy setting, and then Claude began to leave me in awe," he added.
Has the proof actually been verified? 

Alpoge says the proof has been formalised using a computer language called Lean. In simple terms, Lean is a system that can check mathematical proofs step by step, helping establish whether the logical arguments hold.

He has released the formalised material, however, a readable, conventional mathematical paper is yet to follow.

Aploge himself acknowledges that the argument is incredibly long and says he hopes to simplify it. Thus an independent mathematical scrutiny is essential before the claim can be accepted.

"The argument is incredibly long at the moment..."

"I intend to do my best to understand and impart understanding of this argument and its certain simplifications, to speak about it, to write it properly."
Alpoge says the method also extends to the Generalized Riemann Hypothesis, an even broader mathematical conjecture.

Early verification checks by independent researchers seem to verify that the uploaded files match the release and that the top-level theorem statements properly target Lean's standard mathematical definitions, though the full line-by-line proof check is still running.

The bigger question here is whether AI is changing the way mathematics is approached. And this could very well be yet another example of AI moving beyond writing, coding and answering questions into advanced mathematical research.

To be sure, there is a distinction between AI producing a mathematical argument and the mathematical community independently establishing that it is correct.
However, if this claim holds up, it could be a landmark moment for AI-assisted scientific discovery.

Show full article

Track Latest News Live on NDTV.com and get news updates from India and around the world

Follow us:
Listen to the latest songs, only on JioSaavn.com