Kakeya Conjecture in Dimensions Four and Higher

Kakeya Conjecture in Dimensions Four and Higher

OpenAI's claim

Four-dimensional Kakeya sets: The manuscript claims that a set in $\mathbb R^4$ containing a segment of length one in every direction has Hausdorff dimension four. It permits arbitrary sets and arbitrary choices of witnessing segments.

Mathematical statement of the claimed result

Mathematical claim 1.

If $K\subset\mathbb R^4$ and for each unit direction $e\in S^3$ some $a_e\in\mathbb R^4$ satisfies $\{a_e+te:0\le t\le1\}\subset K$, then $\dim_H K=4$. No measurability assumption on $K$ or regularity assumption on the choice $e\mapsto a_e$ is imposed. This theorem treats dimension four, rather than all higher dimensions.

Outline of the argument

Four-dimensional Kakeya sets: A hypothetical dimension loss is converted into a weighted incidence model for lines. Nested boxes and quadratic fits describe persistent concentration; each fit must retain enough incidence mass to contribute. Scalar comparison and planar rigidity then contradict the simultaneous density and time-length requirements forced on substantial terminal chart systems.

Formal verification

Four-dimensional Kakeya sets: The pinned catalogue lists no Lean formalization for this result.

This site has not run the Lean development. These links describe OpenAI’s selected formal statements and should be read with their stated scope.

Status caveat

OpenAI reports that an internal model produced the manuscripts. Their arguments have not been independently verified as proofs by this site. Any portion outside the documented formal scope may contain mistakes, and the catalogue label reports the announced mathematical claim.

Sources

• OpenAI, Every Four-Dimensional Kakeya Set Has Full Hausdorff Dimension, 2026-09-24; citation key OAI:Every-four-dimensional-Kakeya-set-has-full-Hausdorff-dimension-September-24-2026. PDF (GitHub); PDF (direct); Manuscript page and citation; Latest version; Statement source: thm:main.
• OpenAI catalogue entry.

Kakeya Conjecture in Dimensions Four and Higher

Definitions and mathematical statement

For $s\ge0$, Hausdorff measure is obtained from covers by sets of diameter tending to zero, weighting each covering set by its diameter to power $s$. Hausdorff dimension is $\dim_H E=\inf\{s:\mathcal H^s(E)=0\}$. A compact Kakeya set satisfies \[ \forall v\in S^{n-1}\ \exists a_v\in{\mathbb{R}}^n:\quad \{a_v+tv:0\le t\le1\}\subseteq E. \] The selected assertion is $\dim_H E=n$ for every such $E$ and every integer $n\ge4$. Positive Lebesgue measure is not required, and replacing Hausdorff dimension by a different dimension would change the statement.

Formulation and concept references: [S1].

Short English statement

Must a compact set containing a unit segment in every direction have full Hausdorff dimension in every dimension at least four?

Research attempt

Current frontier.

Wang--Zahl prove the three-dimensional set conjecture; Guth's 2026 Bourbaki survey explains that proof. This is not the $n\ge4$ assertion selected here, nor is the set theorem the full Kakeya maximal-function estimate. [E1], [E2]. The formulation source proves Hausdorff dimension at least $3.25$ for sticky Kakeya sets in ${\mathbb{R}}^4$, and describes the distinct general-set bound obtained by the planebrush method. Stickiness cannot be inserted as a hypothesis of this entry. [S1]. The September 2026 paper on lifting curved sets concerns a transfer of geometric problems, not a theorem giving dimension $n$ for every linear Kakeya set in every dimension. Its abstract was checked. [S3].

Attack route.

Projection to three dimensions preserves some line segments but only gives a three-dimensional lower bound, not the missing fiber dimension. Pairwise tube intersections are directly computable but appear to lose too much. The chosen route combines a scale-uniform directional estimate with a weighted allocation of directions to scales. The point of the weighting is to reach Hausdorff dimension, where covers have unequal radii, rather than silently prove only a Minkowski statement.

Attempt.

For $0<\delta<1$, let $T_\delta(a,v)$ be the radius-$\delta$ cylinder around $\{a+tv:0\le t\le1\}$ and define \[ K_\delta f(v)=\sup_{a\in{\mathbb{R}}^n}\frac1{|T_\delta(a,v)|} \int_{T_\delta(a,v)}|f(x)|\,dx. \] The following unproved hypothesis is deliberately stronger than the set statement: \begin{equation}\label{wsref-0-72-33-38-3a-6d-61-78-69-6d-61-6c} \|K_\delta f\|_{L^n(S^{n-1})} \le C_{n,\eta}\delta^{-\eta}\|f\|_{L^n({\mathbb{R}}^n)} \quad\text{for every }\eta>0\text{ and every }\delta,f. \end{equation} It is a standard maximal-function target in the Kakeya framework discussed in [E2]. What follows is an explicit deduction from it, not a proof of it.

Lemma 38.1 (unequal-scale transfer). For a fixed $n$, (r38:maximal) implies the exact Hausdorff-dimension conclusion of this entry in that dimension.

Proof. Fix $s<n$ and choose $\eta>0$ with $n-n\eta>s$. Cover a compact Kakeya set $E$ by balls $B_i$ with radii $r_i\le2^{-J}$, grouping them so that $2^{-j-1}<r_i\le2^{-j}=\delta_j$. Write $N_j$ for the number in group $j$, $U_j$ for their union, and $U'_j$ for the union with each ball replaced by the concentric ball of radius $3\delta_j$. Covers with infinite $s$-cost need no argument; otherwise every $N_j$ is finite.

For each direction $v$, select only for this argument one unit segment $\ell_v\subset E$. Since the cover contains every point of $\ell_v$, \[ \sum_{j\ge J}\mathcal H^1(\ell_v\cap U_j)\ge1. \] Choose fixed positive weights $w_j=c/(j+1)^2$, $j\ge0$, with $\sum_{j\ge0}w_j=1/2$. Some $j\ge J$ has $\mathcal H^1(\ell_v\cap U_j)\ge w_j$. Thickening the corresponding points of the line in perpendicular disks of radius $\delta_j$ remains inside $U'_j$, so $K_{\delta_j}1_{U'_j}(v)\ge w_j$. This uses Fubini in cylindrical coordinates; endpoints have zero one-dimensional measure. Define the measurable set $D_j=\{v:K_{\delta_j}1_{U'_j}(v)\ge w_j\}$. No measurable choice of $\ell_v$ is needed: these maximal-function level sets contain all the chosen directions. Measurability follows by taking the supremum over a countable dense set of translations; integrals of bounded indicators over translated cylinders are continuous in the translation and direction.

The $D_j$ cover the sphere. Chebyshev's inequality and (r38:maximal) yield \[ |S^{n-1}|\le\sum_{j\ge J}|D_j| \le C\sum_{j\ge J}(j+1)^{2n}\delta_j^{-n\eta}|U'_j| \le C'\sum_{j\ge J}(j+1)^{2n}N_j\delta_j^{n-n\eta}. \] Set $\gamma=n-n\eta-s>0$. Since $\sup_{j\ge J}(j+1)^{2n}2^{-\gamma j}\to0$, this inequality gives a positive lower bound, independent of the cover, on $\sum_jN_j\delta_j^s$. Within a group $\delta_j<2r_i$, so the same holds for the original Hausdorff cost up to a fixed factor. In particular $\mathcal H^s(E)>0$, indeed the bound increases as $J\to\infty$. Thus $\dim_HE\ge s$ for all $s<n$, proving $\dim_HE=n$. $\square$

Can a simple overlap estimate prove the missing hypothesis? Take one unit $\delta$-tube in each direction of a $\delta$-separated net of unoriented directions, with $M\asymp\delta^{-(n-1)}$ tubes, and set $F=\sum_T1_T$. Elementary cylinder geometry gives \[ |T\cap T'|\le C_n\frac{\delta^n}{\theta(T,T')} \quad(\theta\ge\delta). \] Here $\theta$ is the acute angle between unoriented axes. For one tube, the number of directions at angular distance at most $\theta$ is $O((\theta/\delta)^{n-1})$. Summing over dyadic angular shells therefore gives, for $n>2$, \[ \sum_{T'}|T\cap T'|\le C_n\delta, \qquad \int F^2\le C_nM\delta\asymp\delta^{-(n-2)}. \] The diagonal term $|T|\asymp\delta^{n-1}$ is harmless. Since $\int F\asymp1$, Cauchy--Schwarz gives only \[ \left|\bigcup T\right|\gtrsim\delta^{n-2}. \] This basic calculation reaches a two-dimensional scale exponent, not $n$. It is not a new best lower bound; it diagnoses the loss in the particular pairwise route. Full ambient dimension would require bounds arbitrarily close to $\delta^0$, with a way to control heavily shaded subsets across scales. The lost information concerns simultaneous configurations of many tubes, exactly the type of structure not recorded by pairwise angles alone.

Stress test.

A lower bound for $|E_\delta|$ at equal scales need not establish Hausdorff dimension. Lemma 38.1 explicitly handles arbitrary countable ball covers. The constant in (r38:maximal) must be independent of tube locations, $\delta$, and the chosen cover; making it depend on the set at a shrinking scale defeats the proof. A family of tubes restricted to a hyperplane is not a counterexample to the target because it lacks most directions. Conversely, projecting a genuine Kakeya set to a three-plane provides no positive lower bound on the dimensions of its fibers. Such a bound cannot be inferred by subtracting dimensions without an appropriate slicing theorem and a verified fiber hypothesis.

The strong maximal estimate is not claimed as a necessary condition for the set theorem, so failing to prove it does not disprove Kakeya. The elementary overlap argument provides neither a candidate exceptional compact set nor the multiscale high-order estimate needed to replace (r38:maximal).

\Needspace{8\baselineskip}

Outcome.

Outcome: Conditional reduction.

The all-scale maximal hypothesis is shown to imply the exact Hausdorff target, with the unequal-radius and direction-allocation details supplied. This is a worked-out version of a classical implication, not a novelty claim. The attempt also locates the exponent loss of its elementary two-tube strategy. No higher-dimensional maximal estimate or counterexample set is produced; that remains the substantive unresolved step.

Sources

• S1 Mukul Rai Choudhuri, Revista Matematica Iberoamericana 42 $2026$, 1227-1256. https://ems.press/journals/rmi/articles/14299708. Catalogue formulation source.
• S2 Fourier analytic properties of Kakeya sets in finite fields. https://londmathsoc.onlinelibrary.wiley.com/doi/10.1112/blms.70367. Catalogue research/status reference; not a proof endorsement.
• S3 Lifting curved Kakeya sets to linear Kakeya sets. https://arxiv.org/abs/2609.06299. Catalogue research/status reference; not a proof endorsement.
• S4 A Unified Geometric Algebra Framework for the Kakeya Conjecture in All Dimensions and Links to the Riemann Zeta Function. https://www.preprints.org/manuscript/202602.1025. Catalogue research/status reference; not a proof endorsement.
• E1 Hong Wang and Joshua Zahl, Volume estimates for unions of convex sets, and the Kakeya set conjecture in three dimensions, arXiv:2502.17655, main theorem. https://arxiv.org/abs/2502.17655.
• E2 Larry Guth, The Kakeya conjecture, after Wang and Zahl, arXiv:2604.03416, Bourbaki survey, introduction and discussion of dimensional and maximal formulations. https://arxiv.org/abs/2604.03416. References consulted 27 September 2026; no extension from dimension three to the present range is assumed.