Wind the counting numbers into a spiral: 1 at the center, 2 beside it, around and around forever. Now light only the primes: the numbers that refuse to be divided. Nothing about that recipe should produce a picture. And yet: highways. Diagonal roads of stars, running clean across a sky that ought to be lawless static. Drag to drift, scroll to climb. Every star is a number that cannot be broken; hover one and ask its name.
For the curious kid.
In 1963 the mathematician Stanisław Ulam was stuck in a boring meeting, doodling. He wound the numbers into a spiral and circled the primes, and stopped. The primes were lining up. They are not supposed to line up. A prime is simply a number that can't be split into equal smaller piles: 7 stubbornly won't, 9 quietly will. Whether a number is prime looks like pure accident, and no formula anyone has ever found can tell you where the next one falls. Yet here are boulevards of them. The reason is half-known: each diagonal of the spiral collects the values of one little formula, and some formulas are simply luckier than others. Press Euler's road to see the luckiest of all: a three-term formula Euler found in 1772 whose first forty outputs are every one of them prime. Why some roads run so rich, nobody has proved. You are looking at an open problem the way you'd look at weather.
Deeper.
Each diagonal here is a quadratic 4n² + bn + c, so Ulam's highways are the visible form of an old conjecture: Hardy and Littlewood's Conjecture F (1923) predicts precisely how prime-rich each quadratic should be, and the spiral's bright roads match its rankings, but the conjecture remains unproved, along with almost everything else worth wanting here. Euler's polynomial n² + n + 41 is prime for n = 0…39 (this page checks, live: it is), and 86 of its first hundred values are prime. The blue toggle lights the twin primes: pairs like 41,43 separated by two. Whether the twins ever run out is unknown; the best result of a generation, from an unknown lecturer named Yitang Zhang in 2013, is that some gap below seventy million recurs forever (since beaten down to 246). Meanwhile the crowd is lawful even though no member is: at fifteen, Gauss guessed that the count of primes up to x runs like the integral li(x), and the meter above compares his guess to the true count of every star on your screen (never fitted, off by a fraction of a percent). How far li can drift from the truth is exactly the Riemann Hypothesis, mathematics' most famous open question and a million-dollar prize. Museum policy: the sieve is checked against π(10⁶) = 78,498, the spiral against a million round-trips, the twins against the known 1,224 pairs below 10⁵.
Ask any number if it is prime, and it answers at once. Ask where the next prime lives, and all of mathematics goes quiet.