AI-assisted construction refutes a Naslund bound over F₃

A ten-word code in F₃⁴ and an eight-degree lift give square-difference-free polynomial sets of size 810ᵉ below degree 8e and 27·810ᵉ below degree 8e+4. They refute Naslund’s Conjecture 13 at q=3, k=2 for every n divisible by four with n≥8, and imply a growth-exponent lower bound of 16/21.

Sources for this record

What was already known

Naslund’s Conjecture 13 predicted the upper bound 3^(3n/4) for these parameters.

AI involvement AI-centered

Jones directed the task; the disclosure credits Claude, GPT-6 and Grok with work on the construction and proof. It does not identify one model as the sole originator.

The disclosure credits multiple AI systems with the central construction under Jones’s direction.

Importance Notable

The infinite family disproves the stated bound and raises the known growth exponent in this setting.

Limitations

The construction gives no optimal value and makes no claim for other q. No independent human mathematical check is documented.

Technical verification Formal (Lean 4.33.0/Mathlib with exact kernel computation)

The released verification report lists Lean proofs of the infinite families and growth bound; finite Python checks are corroboration.

Formal checking

Complete Lean proof build, trust/axiom audit and source guard; exact finite code property by kernel decide

Checker
Lean 4 kernel in Jones’s reported local build
Scope
Six selected polynomial-set, conjecture-refutation and liminf statements over F₃
Basis
Reported by the linked checker
Evidence source

Checking a formal statement does not establish that it matches the original problem.

Human understanding Few

No named human explanation of the code-and-lift mechanism is documented; Few is the evidence default.

Correction and publication history

2026-10-18 · publication

The q=3, k=2 infinite-family refutation is distinct from other coded-set and Lean publications.

Read this published revision