Carbery’s almost-orthogonality inequality in LpL^p

About 17 years old · traced to

For p ≥ 2, does Carbery’s proposed many-function almost-orthogonality inequality hold with the pairwise overlap coefficients raised to the power 2? If not, what is the largest possible exponent?

References

Progress summary

Refreshed
Claimed solved

The proposed power-two estimate is false for every p>2p>2, and a June 2026 preprint claims the optimal replacement has now been proved for all such exponents, though that claim is unverified.

Carbery posed the underlying sharpened triangle-inequality question in 2006. The many-function problem asks whether pairwise overlaps may carry power 22 and, if not, what the largest valid power is.

Known results

  • Carbery proved the corresponding two-function inequality for characteristic functions (2006).
  • Carlen, Frank, Ivanisvili, and Lieb established the general two-function refinement for real pp (2020).
  • For any finite family, validity with overlap power cc requires c≤p′c\le p', where p′=p/(p−1)p'=p/(p-1).
  • The three-function case has earlier bounds of Carlen, Frank, and Lieb, improved in the 2026 work.

June 2026 claimed all-real resolution

A June 2026 preprint constructs counterexamples disproving power 22 for every p>2p>2 and proves power p′p' for every integer p≥2p\ge2. A later June 2026 paper claims, in a remark, a commutative LpL^p proof of the optimal power p′p' for every real p≥2p\ge2; this is not independently verified here. The sources credit Grok with substantial assistance, including the counterexample structure and parts of the proof.

Current status (as of September 2026): Power 22 is disproved for every p>2p>2, and power p′p' is established for integer p≥2p\ge2; a claimed extension to every real p≥2p\ge2 remains unverified.

  • GrokxAIpartial progress2026-05-01evidence
Sources

Solutions 0

No solutions have been posted yet.