Quanta: AI is starting to crack legendary Erdős math problems
kylekabasares · x · 2026-08-04
Quanta reports that AI is starting to make real inroads on legendary Erdős problems, the deceptively simple but deep math questions posed by Paul Erdős.
- In May 2026, OpenAI said an internal model found a counterexample to the 1946 unit distance problem; humans later improved on the result.
- On August 1, OpenAI announced an unreleased model called Astra had made 10 more mathematical advances, including solutions to three additional Erdős problems.
- Mathematicians quoted in the piece describe this as a possible phase change in AI’s mathematical ability, and researchers are now studying what makes these problems uniquely tractable for models.
More from Research
- LeanReact 0.1: expressing composable, provably correct React components in Lean — hargup13 · 2026-09-18
- Cell Focus: Biology Needs World Models That Predict Interventions, Not Just Describe — marinkazitnik · 2026-09-18
- New Paper: Contrastive Noise Alignment Cuts FID Over 50% in Few-Step Generation — burny_tech · 2026-09-18
- DeepSWE-mini: a 16-instance subset that replicates the full DeepSWE leaderboard rankings — asankhs · 2026-09-18
- ICLR26 Paper Defines 'Interpretive Equivalence': Comparing Neural Network Algorithms Without Full Interpretation — burny_tech · 2026-09-18
- Tsinghua and ByteDance Seed unveil SMELT, the first fair comparison of Looped Transformers — jiqizhixin · 2026-09-18