VibeMathedMath problems solved with AI

The volume conjecture for all knots (Murakami and Murakami): does 2πNlog⁡∣JN(K;e2πi/N)∣\frac{2\pi}{N}\log|J_N(K;e^{2\pi i/N})| converge to the simplicial volume?

Let JN(K;q)J_N(K;q) be the NN-colored Jones polynomial of a knot KK, normalized to be 11 on the unknot, and put RN(K)=2πNlog⁡∣JN(K;e2πi/N)∣R_N(K)=\frac{2\pi}{N}\log|J_N(K;e^{2\pi i/N})|. Kashaev conjectured that for a hyperbolic knot this growth rate tends to the hyperbolic volume of the complement. Murakami and Murakami identified Kashaev's invariant with this evaluation of the colored Jones polynomial and proposed a stronger volume conjecture for any knot KK: lim⁡N→∞RN(K)=v3 ∥K∥,\lim_{N\to\infty}R_N(K)=v_3\,\|K\|, where ∥K∥\|K\| is the simplicial volume (Gromov norm) of the knot complement and v3v_3 is the volume of the regular ideal tetrahedron. The right side is the sum of the hyperbolic volumes of the JSJ pieces of the knot exterior. Does the limit exist and equal this volume for every knot, satellite knots included?

Result
Disproved(see note)
Status
Candidate (review pending)
AI contribution
AI-assisted
Method
Construction
Field
Quantum topology; colored Jones polynomial, volume conjecture
Posed by
Hitoshi Murakami and Jun Murakami, The colored Jones polynomials and the simplicial volume of a knot, Acta Math. 186 (2001), 85-104, Conjecture 5.1
Year posed
1999
Years open
27y
Solved
2026-11-06
Model
OpenAI Codex
Vendor
OpenAI
Collaborators
Haibin Hang
Verification
Unreviewed
Publication
Preprint
Significance
40 / 100
Disclosed cost
—
Wikipedia
No dedicated article

What was actually shown

Theorem 1 of arXiv:2611.09063 (Haibin Hang, v1 6 November 2026): let EE be the figure-eight knot with volume VV and L=Cp,w(E#d)L=C_{p,w}(E^{\#d}) the (p,w)(p,w)-cable of the connected sum of d≥2d\ge2 copies of EE, with w≥2w\ge2 and gcd⁡(p,w)=1\gcd(p,w)=1. Then the JSJ volume of LL is dVdV, RN(L)→dVR_N(L)\to dV along NN coprime to 2w2w, and lim sup⁡RN(L)≤V\limsup R_N(L)\le V along NN with gcd⁡(N,w)>1\gcd(N,w)>1, so lim⁡NRN(L)\lim_N R_N(L) does not exist. For the (1,2)(1,2)-cable of E#EE\#E the odd colors give 2V2V and the even colors at most VV; for the (1,3)(1,3)-cable the limit fails even along odd NN. What survives: lim sup⁡NRN(L)=dV\limsup_N R_N(L)=dV still equals the volume for these knots, and Kashaev's conjecture for hyperbolic knots is untouched, since the examples are satellite knots. The paper is 5 pages and rests on the Morton-van der Veen cabling formula, the Habiro-Le formula for the figure-eight knot and a derivative estimate at the root of unity.

What the AI did

The paper's acknowledgment, in full: "OpenAI Codex assisted with developing and checking the argument and preparing the manuscript." The paper does not say which steps came from the model or which model version was used.

Verification

No independent mathematician has checked this yet. Read by this site on 19 November 2026 in the TeX source of arXiv:2611.09063v1 (single author, Haibin Hang, no affiliation given): the whole 5-page paper with its AI acknowledgment, and the volume conjecture in the source of Murakami and Murakami's arXiv:math/9905075, which is stated for any knot and matches. A sanity computation was written and run here, independent of the paper's estimates: the paper's cabling sum with the Habiro-Le formula, in high-precision fixed-point integer arithmetic close to the root of unity, after checking the formula on the unknot, on the trefoil as a cable and on the Jones polynomial of the figure-eight knot. For the (1,2)(1,2)-cable of E#EE\#E it gives R1001=4.186R_{1001}=4.186 and R1000=2.113R_{1000}=2.113, against 2πNlog⁡JN(E)=2.093\frac{2\pi}{N}\log J_N(E)=2.093 at both; for the (1,3)(1,3)-cable R601=4.255R_{601}=4.255 and R603=2.161R_{603}=2.161; the control (1,2)(1,2)-cable of EE gives 2.1092.109 at both parities. This agrees with the theorem at finite NN but proves nothing about the limit, and it takes the cabling formula and the JSJ volume from the paper and the literature. The proof was not refereed; a web search found no public expert reaction.

Sources

Submitted by MellowGander562 on

Changelog2 changes

Discussion