Today, we share a breakthrough on the planar unit distance problem, a famous open question first posed by Paul Erdős in 1946. For nearly 80 years, mathematicians believed the best possible solutions looked roughly like square grids. An OpenAI model has now disproved that belief, discovering an entirely new family of constructions that performs better. This marks the first time AI has autonomously solved a prominent open problem central to a field of mathematics.

New NOVA Blueprint update We are making our chemical search competition significantly more exploration-focused. Starting Wed May 20, 2026 00:00 UTC, submissions containing molecules that are too chemically similar to each other will be rejected entirely. Why this matters: In real drug discovery campaigns, the most valuable search strategies are not the ones that generate many near-identical molecules around the same idea. The real challenge is finding multiple distinct scaffolds that perform well. Once you discover a promising scaffold, generating variations around it is relatively easy. What is hard — and far more valuable — is exploring chemical space broadly enough to uncover entirely new starting points. That is why we originally introduced entropy thresholds in Blueprint, and this new update pushes that philosophy further. We are explicitly rewarding algorithms that: - explore broadly - surface diverse chemical ideas - avoid collapsing into narrow clusters of near-duplicates The goal is not just higher scores but rather better chemical navigation strategies for real-world discovery. This update should lead to: * more novel molecular starting points * stronger search algorithms * and ultimately better drug discovery systems #Bittensor #SN68 #DrugDiscovery #AI #DeSci