Skip to content
All apps

Kakeya needles: tubes in every direction (3D)

About this app

How tightly can needles pointing in many different directions be pushed together? The Kahane construction as a bundle of 16, 256 or 4096 tubes: pushed together, the volume of the union shrinks, but only logarithmically slowly, more slowly than any power of the thickness δ. Glowing needle through all directions, chart 2D versus 3D, planar Besicovitch set. Result from openai/math (family 074): Kakeya maximal conjecture in 3D, full dimension in 4D. Exhibit-ready.

The Kakeya problem

In 1917 Sōichi Kakeya asked how small a region can be in which a needle can be turned all the way round. Besicovitch showed that a set containing a needle in every direction can have area 0. The Kakeya conjecture asks about the dimension of such sets; it is closely tied to Fourier analysis, wave equations and number theory. In the plane the dimension is 2 (Davies 1971), and in space Wang and Zahl showed dimension 3 in 2025. An AI-generated preprint in github.com/openai/math (family 074) claims the stronger maximal conjecture in 3D, and a second one full dimension 4 in ℝ⁴. According to the catalogue there is no Lean formalisation, and the preprints are not peer-reviewed.

What the app shows

The app shows Kahane’s construction from 1969: tubes run from a Cantor dust at the bottom to one half the size at the top, in bundles of 16, 256 or 4096 tubes. Squeeze them together and they share a lot of space, so the volume of their union falls, but only logarithmically, slower than any power of the thickness δ. A glowing needle sweeps through all directions, a chart compares 2D with 3D, and a planar Besicovitch set is included. This is a known construction, not the one from the preprint, and it proves nothing.

What to look for

Squeeze a bundle together and watch the volume: it drops with each level, but only a little. That is what the maximal conjecture predicts: tubes in separated directions cannot overlap arbitrarily, and the volume of their union stays close to the sum of their volumes, up to factors of size δ^ε. The picture is simplified in a few ways: square cross-sections, directions only within a cone, and the volume at intermediate positions estimated on a grid. The app can run as an exhibit.

Frequently asked questions

What is the Kakeya needle problem?
The Kakeya needle problem asks for the smallest area of a region in which a needle of length 1 can be rotated by 180°. Besicovitch showed that this area can be made arbitrarily small. This led to the question of the dimension of sets containing a segment in every direction.
What does the Kakeya conjecture say?
The Kakeya conjecture states that a set in ℝⁿ containing a unit segment in every direction has full Hausdorff dimension n, even though its volume can be zero. In the plane this has been known since 1971. The maximal conjecture is a stronger, quantitative version about tubes of finite thickness.
Has the Kakeya conjecture been solved in 3D and 4D?
For three dimensions, Wang and Zahl showed full dimension 3 in 2025. An AI-generated preprint in the catalogue github.com/openai/math (family 074) additionally claims the maximal conjecture in 3D, and a second preprint full dimension 4 for Kakeya sets in ℝ⁴. According to the catalogue these preprints are not formalised in Lean and have not yet been peer-reviewed.
What can I try out in this app?
You squeeze bundles of 16, 256 or 4096 tubes together and watch the volume of their union shrink. A glowing needle shows that every direction occurs. A chart compares 2D and 3D, and a planar Besicovitch set shows the original problem.

Subject: Open problems, solved by AI

More from Open problems, solved by AI

A VisuApp by heyprof: interactive, free in your browser, no sign-up. What is a VisuApp? · All apps