この記事の日本語版はまだありません。英語版を表示しています。

Erdős Problem #951

Partial progress by GPT-5.2 Pro (27 Jan 2026): solution to variant problem. 3 AI contributions are recorded in all.

問題
Erdős problem #951 (number theory); current status: open
貢献
GPT-5.2 Pro · AI on its own
検証
Listed in Tao's wiki
経緯
First AI contribution 2026-01-27 · wiki data as of 2026-06-30
出典
Problem page (erdosproblems.com)Tao's AI-contributions wiki
部分的
結果
Per Tao's AI-contributions wiki

The problem

Let 1<a1<⋯1<a_1<\cdots be a sequence of real numbers such that

∣∏iaiki−∏jajℓj∣≥1\left\lvert \prod_i a_i^{k_i}-\prod_j a_j^{\ell_j}\right\rvert \geq 1

for every distinct pair of non-negative finitely supported integer tuples ki,ℓj≥0k_i,\ell_j\geq 0. Is it true that

#{ai≤x}≤π(x)?\#\{ a_i \leq x\} \leq \pi(x)?

Statement quoted from erdosproblems.com/951, maintained by Thomas Bloom. Original source: [Er69,p.82][Er77c,p.68]. 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
27 Jan, 2026 AI on its own GPT-5.2 Pro — 🟡 Solution to variant problem —
28 Jan, 2026 AI building on known literature AlphaEvolve — 🟡 New solution to variant problem 🟡 Barreto, GPT-5.2 Pro, and Price (2026) + GPT-5.2 Thinking and Sothanaphan (2026) (solution to variant problem)
28 Jan, 2026 AI working with humans GPT-5.2 Thinking Nat Sothanaphan 🟡 New solution to variant problem —

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.

∎