Gaussian moat — no path to infinity
About this app
Gaussian primes as stepping stones: with a fixed step size D the plane falls apart into islands. A walker runs from the origin to the moat, and a periodic sieve shows the proof idea. Result from openai/math (family 028, formalised in Lean): for every D all islands are uniformly bounded.
The Gaussian moat problem
Gaussian primes are the primes among lattice points a + bi, numbers that cannot be factored further in ℤ[i]. They sit in the plane like stepping stones, thinning out towards the edge. In 1962 Basil Gordon asked whether you can walk to infinity on Gaussian primes using steps of bounded length D. A big gap along one line is no obstacle, because in the plane you can walk around it; what is needed is a moat that closes all the way round. An AI-generated preprint in github.com/openai/math (family 028) answers no: for every D the network breaks into islands of bounded size. The catalogue lists the main theorem as formalised in Lean; the preprint has not been peer-reviewed.
What the app shows
The app computes the Gaussian primes in the window |Re|, |Im| ≤ 1200 and links every pair at most D apart. A walker sets off from the origin until its island ends; the panel shows the number of stones and the farthest distance reached. Other islands can be coloured, and a histogram shows island sizes. For large D the moat lies outside the window, and the app says so plainly: only the proof covers that case.
The proof idea as a game
In the “proof idea” panel you choose sieve primes p ≡ 1 (mod 4). All multiples of their Gaussian factors π and π̄ are removed, leaving a periodic pattern with period Q that contains every Gaussian prime except the sieve primes themselves. If this pattern breaks into finite islands at step length D, so do the primes. Find out which sieve is enough, and watch for the message when the pattern still allows an infinite path.
Frequently asked questions
- What are Gaussian primes?
- Gaussian integers are complex numbers a + bi with integer a and b. A Gaussian prime cannot be factored further within this number system. Ordinary primes of the form 4k + 1, such as 5 = (2 + i)(2 − i), split there, while primes of the form 4k + 3 stay prime.
- What is the Gaussian moat problem?
- Basil Gordon’s question from 1962 asks whether you can walk arbitrarily far from the origin on Gaussian primes using steps no longer than D. If every D had a closed moat free of primes, the answer would be no. For a long time only computations for individual step lengths were known.
- Has the Gaussian moat problem been solved?
- An AI-generated preprint in the catalogue github.com/openai/math (family 028) states that for every D the network of Gaussian primes splits into islands whose size is bounded by some B_D. The bound is not explicit. According to the catalogue the main theorem is formalised in Lean; it has not yet been peer-reviewed.
- What can I try out in this app?
- You set the step length D, send a walker out from the origin and see where its island ends. Clicking a stone shows its island, and a histogram shows island sizes. In the “proof idea” panel you can play through the periodic sieve from the proof on small examples.
Subject: Open problems, solved by AI
More from Open problems, solved by AI
- Anderson localisation: 2D vs. 3D (3D)
- Barker sequences and circulant Hadamard matrices
- Barnette's conjecture: a round trip over every polyhedron (3D)
- Birkhoff billiards: only the ellipse has no gaps
- Cardy's formula — percolation in a random tiling
- The plane is not five-colourable (Hadwiger–Nelson)
- Double-dimer loops and CLE₄
- Triangular billiards — irrational angles mix (3D)
A VisuApp by heyprof: interactive, free in your browser, no sign-up. What is a VisuApp? · All apps