Why the Legendary Erdős Problems Are Falling to AI
AI’s greatest mathematical successes have come from answers to problems posed by a mid-20th century iconoclast. By examining what makes the Erdős problems unique, mathematicians are trying to understand how AI might change the rest of math. The post Why the Legendary Erdős Problems Are Falling to AI first appeared on Quanta Magazine
On May 20, 2026, OpenAI made an announcement that sent shockwaves through the mathematical community. An internal AI model, not accessible to the general public, had generated a counterexample to the "unit distance" problem, a conjecture proposed by mathematician Paul Erdős in 1946. This problem, simple in explanation yet mathematically profound, marked the first significant proof to emerge from an AI model.
Initially, the model's result was not conclusive; however, within days, improved techniques were employed, leading to the solving of additional Erdős problems. On August 1, OpenAI revealed that another unreleased model, named Astra, had achieved 10 more mathematical breakthroughs, including solutions to three more Erdős problems.
Many mathematicians viewed these developments as a watershed moment in AI's mathematical capabilities, dramatically altering the landscape of mathematical research. Erdős, known for his itinerant lifestyle and gift of posing questions to the mathematical world, often attached rewards for solutions. Despite his death in 1996, a foundation in Iowa continues to honor his bounties.
He was a controversial figure, preferring silk and avoiding physical contact, and fueled his prolific output with amphetamines. The problems he posed have now become a central proving ground for the world's leading technology firms, often generating PR triumphs for these companies. This shift was largely due to Thomas Bloom, an English mathematician with a keen interest in arithmetic combinatorics.
Bloom, recognizing the need for a comprehensive list of Erdős problems, created erdosproblems.com in early 2023, initially for his personal use but later sharing it widely. Bloom's meticulous curation of the list facilitated the solving of many problems by applying modern techniques. His website quickly gained traction, with nearly 1,000 problems cataloged by August 2025.
The addition of a comment feature further enriched the site, fostering nuanced mathematical discussions among a diverse group of participants, including non-professional mathematicians. Bloom's foresight in creating this digital platform played a crucial role in the AI-powered breakthroughs, transforming Erdős' legacy into a catalyst for mathematical innovation.
Written by urgent.news from Quanta Magazine's reporting — not their text. Machine-written; read the original for the full account.
Also reported by 1 other outlet
- Why the Legendary Erdős Problems Are Falling to AI quantamagazine.org
