OpenAI announces solutions to 10 longstanding maths problems

OpenAI's prototype AI model Astra has reportedly solved ten longstanding mathematical problems. This disclosure comes in a series of recent AI-driven mathematical discoveries that are expected to reshape the field. In May, an OpenAI model solved a decades-old conjecture by Paul Erdős, and last month, Claude Fable 5 AI found a counterexample to the Jacobian conjecture, which had persisted for nearly a century.

These are among numerous AI-led mathematical discoveries made in recent months. The new solutions from OpenAI cover diverse problems, including the dense packing of spheres in spaces with more than three dimensions and quantum game theory. However, the one generating the most buzz is the discovery of a non-sofic group. Soficity, a property of group operations, has been a puzzle since 1999.

Mathematicians were uncertain if non-sofic groups existed until now, and OpenAI has finally proven their existence. This news has caused a stir among mathematicians, who are adjusting to this significant shift in their field. Some even speculate that we may have reached the "singularity" – the point at which AI becomes self-improving and moves towards general intelligence rapidly.

While OpenAI claims that all ten solutions have been unaddressed for at least a decade, some mathematicians believe that prior human work significantly contributed to these discoveries. They argue that AI alone could not have arrived at these solutions without human input, particularly concerning the new non-sofic group. Despite these concerns, the release of these ten solutions is still considered the most impressive AI mathematical achievement thus far.