An OpenAI chatbot just solved an 80-year-old geometry problem that the legendary mathematician Paul Erdős thought was closed. This isn't just a math milestone — the algorithm behind it could revolutionize how we optimize health, from supplement timing to fasting protocols. Published in Nature on May 22, 2026, the finding has stunned mathematicians and biohackers alike, who see in this breakthrough a potential tool to push past human limits once considered insurmountable.
The Science Behind the Breakthrough
The problem, known as the "Erdős distinct distances problem," was posed in the 1940s. Erdős proved that certain point configurations in the plane couldn't exist, but left the possibility open for others. The OpenAI chatbot, trained with reinforcement learning and tree search, found a configuration Erdős considered impossible: a set of 7 points in the plane where all distances between them are integers, but not all points lie on a straight line. The result has stunned mathematicians, as for decades it was assumed such a configuration did not exist.
The chatbot's key wasn't brute force, but an algorithm combining random exploration with intelligent branch pruning. This approach mirrors methods used in health protocol optimization — for example, finding the optimal combination of supplements and fasting windows to maximize autophagy or insulin sensitivity. The algorithm evaluated trillions of possibilities in hours, a task that would have taken humans centuries. What's fascinating is that the same principle of intelligent search can be applied to biological problems where the space of possibilities is equally vast.
