An AI solves an 80-year-old geometry puzzle. What do mathematicians think? – Bitcoin News

An 80-year-old geometric conundrum finally changed when an OpenAI system assembled an improbable construction that exceeded long-held expectations. The unit distance problem, posed by Paul Erdős in 1946, asks how many pairs of points exactly one unit apart can exist among n points on the plane; the AI ​​found patterns that develop faster than the classic playbook allowed. Princeton mathematicians checked out the work, and heavyweights like Tim Gowers and Arul Shankar took note. Beyond bragging rights, the result hints at a new type of collaborator in mathematics, one that uses general inference to overcome human heuristics.

AI solves 80-year-old math mystery with revolutionary solution

Some problems continue to stretch the limits of human patience. The unit distance problem, posed in 1946 by Paul Erdős, posed a deceptively precise question: given n points on a flat plane, how many pairs can be exactly one unit apart. Generations have attacked it with grids, symmetry and courage. Progress was made in small steps, never in leaps. Then, discreetly, an AI intervened.

A decades-old problem finally solved

The classic approach arranged the points in square grids, adjusting the scale to attract more pairs at a distance of 1. This method suggested growth just above the linear line, roughly n multiplied by a factor that barely beats n as it grew. The field was based on the idea that the best lower bound hovered near n^(1+o(1)), a step above n, not a stride.

How AI beat guesswork

According to the researchers involved, an internal OpenAI model proposed a new family of point configurations that crosses a threshold long considered out of reach. The system produced constructions with at least n^(1 + δ) pairs of distance units, for a fixed δ greater than 0 that does not fade as n increases. This is a true polynomial improvement, not a rounding error.

The approach combined geometric understanding with advanced algebraic number theory, a surprising toolkit for a spatial counting puzzle. This doesn’t come from a specialized math engine. Instead, it emerged from a general inference model being evaluated, suggesting broader reasoning abilities capable of navigating across domains when the search space is large.

Confirmed by experts, celebrated by the field

Independent mathematicians at Princeton University examined the AI ​​constructs and confirmed the result, according to people familiar with the study. Esteemed voices, including Sir Tim Gowers and Arul Shankar, have hailed this advancement as a significant step for the field. This is the case where a new lower limit, long static, was finally moved because an AI found the right lens.

Implications for mathematics and beyond

What does it mean when a general model goes beyond entrenched conjectures? On the one hand, this alludes to a workflow in which machines surface candidate structures and humans stress test them. Besides geometry, disciplines such as combinatorics, coding theory and cryptography could see similar collaborations when proofs rely on rare constructs.