VibeMathedMath problems solved by AI

Chafai-Dadoun-Youssef Questions on Logarithmic Energy Monotonicity

Chafai, Dadoun and Youssef asked whether the quadratically penalised logarithmic energy of mean empirical spectral distributions is monotone in the dimension, for Wigner matrices and for matrices with i.i.d. entries. Neither holds: a finite-energy Wigner counterexample and a one-parameter family of Gaussian-regularised Bernoulli entry laws answer both questions negatively.

Result
Disproved
Status
Resolved
AI contribution
AI co-developed
Method
Construction
Field
Random matrix theory
Posed by
Djalil Chafai, Benjamin Dadoun, Pierre Youssef
Year posed
Years open
Solved
2026-07-27
Model
ChatGPT Pro 5.5, ChatGPT 5.6 Sol Ultra
Vendor
OpenAI
Collaborators
Theodoros Assiotis
Verification
Unreviewed
Publication
Preprint
Significance
10 / 100
Disclosed cost
Wikipedia
No dedicated article

What the AI did

The disclosure says the tools were used for ideation, technical help, editing and checking the proofs, and puts the contribution at the level of a co-author, while the author retains responsibility for mistakes. It does not separate which of the two counterexamples came from where.

Verification

Single-author arXiv preprint; not yet peer-reviewed.

Source

arXiv:2607.24170 - On non-monotonicity of logarithmic energy for random matrices

Discussion