Artificial intelligence may have solved one of the greatest mathematical problems of all time

Artificial intelligence may have solved one of the greatest mathematical problems of all time

Mathematician Levent Alpoge, employed by the American artificial intelligence company Anthropic, claims he has managed to solve Riemann’s hypothesis, one of the most well-known and difficult problems unresolved in mathematics, with the help of the artificial intelligence system Claude.

If his discovery is confirmed, this would make remarkable progress in mathematics but also an important evidence of artificial intelligence’s ability to help solve problems that for more than a century and a half have challenged the best scientists in the world, they write media.

Alpoge published a four – page document on Friday explaining how it came to the possible solution with Claude’s help. However, his claim has not yet been subjected to a thorough and independent verification from the mathematical community.

The mathematical problem no one solved for 167 years.

Artificial intelligence may have solved one of the greatest mathematical problems of all time


Riemann’s hippo was formed in 1859 by German mathematician Bernhard Riemann. It relates to the distribution of simple numbers, the numbers that are limited to number one and to themselves.

Examples of simple numbers are 2, 3, 5, 7 and 11. Although shown by some mathematical rules, the way they are distributed among the larger numbers has not yet been fully explained.

Riemann’s hypothetosis provides for a certain mathematical rule that relates to this distribution. Over the decades, scientists have collected many supporting evidence, but no one has so far managed to prove that it is always true.

Claudide passed me a long way ahead of final stage

Alpoge has shown that she has been working with Claude for several months in solving various mathematical problems. According to him, artificial intelligence played an especially important role in the final phase of work on Riemann’s hypothesis.

I’ve spent the past few months working on a wide range of problems, but this time Claude spent a lot of money in the final phase, Alpoge wrote in the document published on his page.

He added that Claude used to dance, dance and shine at the last working moments, thus describing the role of artificial intelligence in seeking to solve a problem that mathematicians have been trying to solve for over a century and a half.

Evidence contains 17 million lines of code

YouTube youtu.be

Alpoge published the alleged evidence in the form of some 17 million lines of the code written in Leah’s programming language. This system allows the computer to verify mathematical evidence step by step.

Unlike popular scientific works, where mathematicians explain their conclusions through formulas and text, in this case most of the evidence is written in order to be controlled by a computer program.

According to Alpoge, cooperation with artificial intelligence has produced more than 250,000 pages of mathematical material. He also acknowledged that he has not yet fully studied all parts of the published evidence.

Computer accepts test, but experts request verification

Alex Kontorovich, professor of mathematics at Rutgers University, explained that mathematicians usually have to read and verify every piece of evidence to ensure that all results are correct.

In this case, much of the computer does.

Kontorovich said the computer system accepts the evidence presented, but that does not mean he has fully understood it.

Does that mean I understand? Not yet. But the computer is extremely demanding, much more demanding than I am. So if you can convince the computer, you’re most likely to convince me, he said.

One million dollars is offered for the solution

Riemann’s hypothetosis is one of seven major mathematical problems that the Clay Institute for Mathematics offered in 2000 a $1 million reward for any problem it manages to solve through a scientifically accepted test.

Hours after Alpoge’s publication, the institute still classified Riemann’s hypothesis as an unresolved problem on its official page. A detailed review by experts is needed for accepting the test.

The president of the institute, Martin Bridson, said the publication could have historical significance and change the way artificial intelligence opportunities are understood. However, he warned that scientists still have much work ahead.

But the way from today’s publication to the true meaning will be long and difficult, Bridson said.

Another major mathematical problem allegedly solved with AI

This is the second case in recent months when an artificial intelligence company or an associate publishes a possible solution to one of the biggest unresolved mathematical problems.

In September, OpenAI announced that its technology had played a key role in the possible solution to the problem of the Navier-Stokes equations, which describe the movement of fluids and gases. Even this result is pending verification by experts.

‘%Navier-Stokes was resolved, and now Riemann’s hypothesis, Kontorovich said, describing the difficulty of believing that two so well-known mathematical problems could have taken solutions within just a few months.

The claim comes after the release of Anthropic to the stock market

The Alpoge publication comes shortly after the release of the Anthropic company on the US Stock Exchange. The company collected $81 billion through its initial public stock offering, reaching an estimated $1.5 trillion.

Mathematician claims he reached the possible solution Sunday evening and that prior to the publication he consulted with Anthropic Executive Director Dario Amodei to get confirmation if he could make the results public.

Anthropic did not answer the Associated Press’s questions about Alpoge’s claims.

Now it remains that the mathematical community will determine whether Claude has really helped solve one of the biggest scientific problems that has remained unresolved since 1859. /Telegraphy/