AI Model Cracks 80-Year-Old Math Problem That Stumped Humans
An OpenAI model has disproved the Erdős unit distance conjecture, a famous problem in discrete geometry that had resisted human mathematicians for eight decades.

In a breakthrough announced in mid-May, OpenAI revealed that an internal AI model had successfully disproved the Erdős unit distance conjecture, a renowned problem in discrete geometry that had baffled human mathematicians for 80 years. The achievement was met with enthusiasm from the mathematical community. OpenAI provided several mathematicians with early access to the result, and their reactions were overwhelmingly positive.
Tim Gowers, a Fields Medal winner, the most prestigious prize in mathematics, wrote that “there is no doubt that the solution to the unit-distance problem is a milestone in AI mathematics.” Daniel Litt, a professor at the University of Toronto, echoed Gowers' sentiments, noting that “this is the first example of a result produced autonomously by an AI that I find exciting in itself, as opposed to as a leading indicator.” This accomplishment marks a significant milestone in the application of AI to mathematical research, demonstrating the potential for machines to make meaningful contributions to the field. The Erdős unit distance conjecture, named after the Hungarian mathematician Paul Erdős, has been an open problem in discrete geometry since its formulation decades ago. The conjecture deals with the maximum number of unit distances that can be achieved by a set of points in a given space.
While human mathematicians have made significant progress on related problems, the Erdős unit distance conjecture had remained unsolved – until now. The success of OpenAI's model highlights the growing capabilities of AI in tackling complex mathematical challenges. As AI continues to evolve and improve, it is likely to play an increasingly important role in advancing our understanding of mathematics and the natural world.
The implications of this achievement extend beyond the mathematical community, with potential applications in fields such as computer science, physics, and engineering. As researchers continue to explore the possibilities of AI-assisted mathematics, we can expect to see new breakthroughs and innovations emerge in the years to come.
Source: Ars Technica