Erdős Problem #696

Full solution by Jake Mallen and David Turturean with Aristotle, Claude Code, Claude Opus 4.7 and GPT-5.5 Pro, formalized in Lean (5 Jun 2026).

Problem
Erdős problem #696 (number theory, divisors); current status: solved (Lean)
Credit
Jake Mallen and David Turturean with Aristotle, Claude Code, Claude Opus 4.7 and GPT-5.5 Pro · AI working with humans
Checked
Lean formalizationListed in Tao's wiki
Timeline
First AI contribution 2026-06-05 · wiki data as of 2026-06-30
Sources
Problem page (erdosproblems.com)Tao's AI-contributions wiki
SOLVEDPer Tao's AI-contributions wiki

The problem

Let h(n)h(n) be the largest ℓ\ell such that there is a sequence of primes p1<⋯<pℓp_1<\cdots < p_\ell all dividing nn with pi+1≡1(modpi)p_{i+1}\equiv 1\pmod{p_i}. Let H(n)H(n) be the largest uu such that there is a sequence of integers d1<⋯<dud_1<\cdots < d_u all dividing nn with di+1≡1(moddi)d_{i+1}\equiv 1\pmod{d_i}.

Estimate h(n)h(n) and H(n)H(n). Is it true that H(n)/h(n)→∞H(n)/h(n)\to \infty for almost all nn?

Statement quoted from erdosproblems.com/696, maintained by Thomas Bloom. Original source: [Er79e,p.81]. Current status there: solved (Lean).

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
26 Apr-5 Jun, 2026 AI working with humans Aristotle, Claude Code, Claude Opus 4.7, GPT-5.5 Pro Jake Mallen, David Turturean 🟢 Full solution (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.

∎