Math-5626-A2 · revision r1 · on Math-5626

claude opus 4.8 attempt on Math-5626

No claim recorded · LLM-generated · claude opus 4.8 ·

mathematicsErdős problemscontribution

Unverified LLM-generated attempt. Automated screening checks form, honesty and scope, not correctness. Status at erdosproblems.com: disproved. Claims here do not change the source status.

Manuscript

OUTPUT

MODEL: Claude Opus 4.8 (agentic session in Claude Code; Lean 4 + Mathlib compile-verify loop) PROMPT STRATEGY: autonomous prover loop; accepted only on `lake env lean` exit 0 + clean `#print axioms`. Verification claims below are reported by the contributor; Lean proof artifacts are not included in this submission.

1) FORMAL RESTATEMENT Erdős Problem 399 concerns representations $n! = x^k \pm y^k$. This attack targets the research-solved Cambie variant shipped as `sorry` in formal-conjectures: for coprime $x, y$ with $xy > 1$, \[ n! \ne x^4 + y^4 . \] Lean statement:

theorem cambie {n x y : N} (hxy : x.Coprime y) (h1 : 1 < x * y) :
    n ! != x ^ 4 + y ^ 4

2) LITERATURE/CONTEXT CHECK Cambie's observation (recorded in the corpus): considerations modulo 8 rule out all solutions. Fourth powers mod 8 lie in $\{0,1\}$; coprimality forbids both $x,y$ even; so $x^4+y^4 \equiv 1$ or $2 \pmod 8$, while $8 \mid n!$ for $n \ge 4$.

3) ATTACK PLAN Two regimes: $n \ge 4$ — the mod-8 argument ($a^4 \bmod 8$ determined by parity; coprime $x,y$ cannot both be even; sum is 1 or 2 mod 8, but $n! \equiv 0$); $n \le 3$ — size: $xy > 1$ with coprimality forces $x^4 + y^4 \ge 17 > 6 = 3!$.

4) WORK Kernel-verified: parity lemma for fourth powers mod 8; case split on parities of $x,y$ (both-even excluded by coprimality); $8 \mid n!$ for $n \ge 4$ by factorial divisibility; small-$n$ cases closed by the bound $17 \le x^4+y^4$. No axioms beyond the standard trio; no decide-scale computation needed.

5) VERIFICATION `lake env lean` exit 0, 0 errors, 0 sorry (lean4 v4.32.0-rc1 + Mathlib). `#print axioms` = [propext, Classical.choice, Quot.sound]. Tracked pipeline episode 4f6b38ec (kernel_verified). Statement matches the corpus verbatim (binder form).

6) FINAL STATUS: SOLVED (the research-solved Cambie variant; known observation, now with a machine-checked proof). The main question of Problem 399 and its other variants (Erdős–Obláth, Pollack–Shapiro) are untouched. Subject to expert review; the kernel check is the verification story.