Erdős Problem #942

Partial progress by Gemini (23 Nov 2025): new proof found. 2 AI contributions are recorded in all.

Problem
Erdős problem #942 (number theory, powerful); current status: open
Credit
Gemini · AI building on known literature
Checked
Lean formalizationListed in Tao's wiki
Timeline
First AI contribution 2025-11-23 · wiki data as of 2026-06-30
Sources
Problem page (erdosproblems.com)Tao's AI-contributions wiki
PARTIALPer Tao's AI-contributions wiki

The problem

Let h(n)h(n) count the number of powerful (if p∣mp\mid m then p2∣mp^2\mid m) integers in [n2,(n+1)2)[n^2,(n+1)^2). Estimate h(n)h(n). In particular is there some constant c>0c>0 such that

h(n)<(log⁡n)c+o(1)h(n) < (\log n)^{c+o(1)}

and, for infinitely many nn,

h(n)>(log⁡n)c−o(1)?h(n) >(\log n)^{c-o(1)}?

Statement quoted from erdosproblems.com/942, maintained by Thomas Bloom. Original source: [Er76d,p.35]. Current status there: open.

AI contributions

As recorded in Terence Tao's wiki AI contributions to Erdős problems (data as of 30 Jun 2026). 🟢 full, 🟡 partial, ⚪ unverified, 🔴 incorrect.

Date Kind AI systems Humans Outcome Literature
23 Nov, 2025 AI building on known literature Gemini — 🟡 New proof found 🟡 Erdős (unpublished, 1976)
23 Nov, 2025 Literature search ChatGPT Deep research — 🟡 Partial results found —
13-14 Jun, 2026 AI working with humans Aristotle, Claude Opus 4.8, Codex, GPT-5.5 Scott Hughes 🟡 Partial result (Lean) —

Caveats

The wiki's classifications are provisional and are not a benchmark. Many Erdős problems have never had a thorough literature review, so a result that looks new may later turn out to be known; some problems were also stated ambiguously or incorrectly. See the problem page for the full discussion.

∎