VibeMathedMath problems solved by AI

Erdos-Graham Question on Averages of Unit Fractions

Erdos and Graham asked whether a positive-density subset of {1,,N}\{1,\ldots,N\} can avoid having any two distinct elements a,ba,b whose unit fractions average to a unit fraction. It can: there is a constant c>0c>0 such that for all large NN some A{1,,N}A \subseteq \{1,\ldots,N\} of size >cN> cN has that property, which also gives the best known lower bounds for related unit-fraction avoidance problems.

Result
Disproved
Status
Resolved
AI contribution
AI-assisted
Method
Construction
Field
Number theory
Posed by
Paul Erdos, Ronald Graham
Year posed
1980
Years open
46y
Solved
2026-07-16
Model
ChatGPT
Vendor
OpenAI
Collaborators
Will Sawin
Verification
Unreviewed
Publication
Preprint
Significance
15 / 100
Disclosed cost
Wikipedia
No dedicated article

What the AI did

The author gives a narrative rather than a blanket acknowledgement. Starting from a computation of Stijn Cambie, he asked ChatGPT to look for patterns in Cambie's extremal set that might suggest a generalization; it observed that in a pair with a given ratio the larger element is usually absent unless the smaller is absent for other reasons, and described a change of variables. Dropping the hedges in that observation gives the set the paper analyzes, which turns out to be essentially where Hooley's function takes its minimum value. The model was also used for reference search and proofreading.

Verification

Single-author arXiv preprint; not yet peer-reviewed.

Source

arXiv:2607.15419 - Sets of unit fractions without two members whose average is a unit fraction

Discussion