An OpenAI model has disproved a central conjecture in discrete geometry
An OpenAI model has disproved the long-standing unit distance conjecture in discrete geometry, a problem mathematicians have studied for nearly 80 years. This marks the first time an AI has autonomously solved a prominent open problem in mathematics, showcasing advanced reasoning capabilities.
For nearly 80 years, mathematicians have pondered the planar unit distance problem: how many pairs of points can be exactly one unit apart when n points are placed on a plane? This problem, posed by Paul Erdős in 1946, is a cornerstone of combinatorial geometry, known for its simple statement yet profound difficulty.
An internal OpenAI model has now disproved a long-held conjecture regarding this problem. Previously, it was believed that square grid constructions were optimal for maximizing unit-distance pairs. The AI model provided an infinite family of examples that offer a polynomial improvement, a finding verified by external mathematicians.
This breakthrough is significant not only for its mathematical resolution but also for its origin. The proof emerged from a general-purpose reasoning model, rather than one specifically trained for mathematics. This outcome is part of a broader initiative to assess advanced models' contributions to frontier research, with this instance successfully resolving an open Erdős problem.
The achievement is a major milestone for both the mathematics and AI communities. It marks the first time AI has autonomously solved a prominent open problem central to a mathematical subfield, demonstrating the sophisticated reasoning capabilities these systems now possess. The solution also introduces unexpected and advanced concepts from algebraic number theory to address an elementary geometric question.
Leading experts have lauded this development. Fields medalist Tim Gowers described it as "a milestone in AI mathematics," while number theorist Arul Shankar noted that AI models are now capable of "original ingenious ideas" beyond merely assisting human mathematicians. These sentiments underscore the profound impact of this AI-driven discovery.
Related articles
The AI world is getting ‘loopy’
AI models are taking a significant leap forward with the adoption of "agentic loops," where AI agents continuously prompt each other to improve code and solve complex problems. This approach, though potentially resource-intensive, promises to unlock new levels of autonomous problem-solving and efficiency in AI applications.
Codex-maxxing for long-running work
Codex is increasingly being used by organizations to support long-running projects that go beyond a single prompt. This whitepaper by Jason Liu offers practical strategies for leveraging Codex as a persistent workspace, managing complex workflows and sustaining progress.
Nobel laureate John Jumper is leaving DeepMind for rival Anthropic
Nobel laureate John Jumper is departing Google DeepMind to join its competitor, Anthropic, after dedicating nearly nine years to DeepMind, where he led the AlphaFold team. Jumper, who shared a Nobel Prize for his work on AlphaFold, expressed gratitude for his time at DeepMind while looking forward to new endeavors.
