Kakeya maximal conjecture in three dimensions and the Kakeya set conjecture in four
Two manuscripts claim that every set in four-dimensional space containing a unit segment in every direction has full Hausdorff dimension, and that the Kakeya maximal function estimate holds in three dimensions, extending Wang and Zahl's 2025 resolution of the three-dimensional set conjecture.
- Lab
- OpenAI
- Model
- Unreleased internal OpenAI model, not named in the release
- Field
- Mathematics
- Date
- 2026-10-06
Sources
Original work
Announcement
Media coverage
What was found
The four-dimensional paper proves that every containing a unit segment in every direction has , with no measurability assumed on the set or on the choice of segments. Its introduction gives the best earlier lower bound as , from Katz and Zahl's planebrush argument, against Wolff's from 1995; Rai Choudhuri's covers only sticky sets. The three-dimensional paper proves the stronger maximal form: for every the radius- tube maximal operator maps to with norm . It builds on the streamlined Guth, Wang and Zahl proof of the set estimate and addresses the gap Wang and Zahl state after their own theorem, replacing a density loss by . Scientific American's report names only the four-dimensional result.
Novelty check
Read both manuscripts' introductions at commit adc7f1241b42e322a6451854ab7e4b4c146bf78a on 2026-10-07. Wang and Zahl's 2025 preprint resolved the Hausdorff and Minkowski dimension Kakeya set conjecture in three dimensions; the maximal conjecture in three dimensions and the set conjecture in four were open, and both papers say so with citations. No earlier claim of either result was found in a web search or in the registry; 2025-11-03-alphaevolve-at-scale mentions Kakeya only for an explicit construction that turned out to match prior work, not for a dimension bound.
Caveats and known objections
No Lean formalization accompanies either paper, unlike much of the release, so the only check available is reading them, and both are long multiscale harmonic-analysis arguments of the kind that take specialists months: the source of the four-dimensional paper runs to about 640,000 characters and the three-dimensional one to about 340,000. Nobody outside OpenAI is on record as having read either. Autonomy is graded ai-led because the README places neither among the exceptions to its single-agent procedure, but that account rests on OpenAI alone and the model is unreleased; it is not graded autonomous because humans chose the problems and filtered the output. The three-dimensional result depends directly on the Guth, Wang and Zahl set estimate, which is itself recent. A later manuscript in the same collection, recorded at 2026-10-26-local-smoothing-all-dimensions, claims local smoothing in every dimension with a Lean formalization whose statements include the Kakeya maximal estimate and full Hausdorff dimension for Kakeya sets in every dimension, which would contain both results here; it is OpenAI's own and unaudited, and it does not check either of this entry's manuscripts.
The open problem
MathDBThe maximal Kakeya conjecture
367525
Nothing mechanically. It records what a problem says, what is known about it, and the standing of any claimed solution.
Also recorded at
vibemathedThe Kakeya maximal conjecture in three dimensions
kakeya-maximal-conjecture-r3
Nothing mechanically. It is a parallel listing of the same result, carrying its own verification label rather than an independent check.
Covers the three-dimensional maximal half of this entry only, recorded there as unreviewed; vibemathed keeps the four-dimensional set result as a separate record.
No independent check is recorded on this entry yet.
Reading the primary source closely enough to say whether it supports the claim counts as a check, and you are credited on the entry.
Or on GitHub: submit a check challenge the grade send a correction or send a pull request
Flag this for triage
Signals order the review queue and nothing else. They are never published, and they never move a grade: that takes a citation.
History of this entry
- CorrectedNoted that OpenAI's local smoothing release of 26 October 2026 claims both results here within a Lean-formalized result for every dimension. No grade moved. source ↗
- CorrectedRecorded the MathDB record for the maximal conjecture as a registration. No grade moved.
- CorrectedRecorded the vibemathed record for the three-dimensional maximal result as a registration. It carries the same claim and the same unreleased-model attribution this entry does. No grade moved. source ↗
- AddedEntered the registry graded Claimed and AI-led.
Entries are never deleted. A grade that does not hold up is downgraded on the record, with the reason beside it.
Community discussion
Graded claimed for verification and ai-led for autonomy. What these mean.
Cite this entry
whataifound.org. (2026). Kakeya maximal conjecture in three dimensions and the Kakeya set conjecture in four. whataifound.org: A Registry of AI Scientific and Mathematical Discoveries. https://whataifound.org/finding/2026-10-06-kakeya-3d-maximal-4d
BibTeX
@misc{whataifound-openai-2026-4d,
title = {Kakeya maximal conjecture in three dimensions and the Kakeya set conjecture in four},
author = {{whataifound.org}},
year = {2026},
howpublished = {whataifound.org: A Registry of AI Scientific and Mathematical Discoveries},
note = {Result by OpenAI. Verification: Claimed. Autonomy: AI-led.},
url = {https://whataifound.org/finding/2026-10-06-kakeya-3d-maximal-4d}
}