Skip to content
daily-hour-news·

🔬OpenAI Model Cracks 80-Year-Old Erdos Math Problem

TL;DR

A general-purpose OpenAI reasoning model produced an original proof disproving Erdos's 1946 planar unit-distance conjecture. It found a new family of point constructions that beats the square grid, verified by mathematicians including Princeton's Noga Alon.

A general-purpose OpenAI reasoning model produced an original proof disproving Erdos's 1946 planar unit-distance conjecture. It found a new family of point constructions that beats the square grid, verified by mathematicians including Princeton's Noga Alon.

OpenAI Model Cracks 80-Year-Old Erdos Math Problem — daily-hour-news

Key Points

1

Disproves the planar unit-distance conjecture Erdos posed in 1946

2

Used Gaussian integers, class field towers, and Golod-Shafarevich theory

3

Model was not fine-tuned for math; given only the written problem statement

4

Independently checked by Noga Alon, Melanie Wood, and Thomas Bloom

Why It Matters

A general reasoning model connecting algebraic number theory to geometry on its own suggests AI can now sustain long proof chains across fields humans never bridged.

Quick Facts

OpenAIreasoning modelsmathematicsErdos conjectureAI researchproofs

Frequently Asked Questions

Why does this matter?

A general reasoning model connecting algebraic number theory to geometry on its own suggests AI can now sustain long proof chains across fields humans never bridged.

What happened?

A general-purpose OpenAI reasoning model produced an original proof disproving Erdos's 1946 planar unit-distance conjecture. It found a new family of point constructions that beats the square grid, verified by mathematicians including Princeton's Noga Alon.

Comments

Subscribe to join the conversation...

Be the first to comment

Enjoyed this article?

Get it daily. 7am. Free. Reads in 5 minutes.

Join 3,461 builders reading daily.

Also get