A cellular automaton is a grid where every cell holds a state (alive or dead, a color, a number) and updates each tick by looking only at its neighbors. Conway’s Game of Life, from 1970, needs two rules: a live cell with two or three live neighbors survives, a dead cell with exactly three comes alive. From that, gliders crawl, oscillators pulse, and patterns grow and collapse with the drama of a petri dish, so much so that Life supports self-replicating machines and working computers built inside it.
Stephen Wolfram catalogued the one-dimensional versions, showing that even elementary rules split into stable, periodic, chaotic, and the interesting fourth class that hovers between, complex without repeating. That edge-of-chaos zone is where the visually compelling automata live.
On a GPU, automata are almost free: the update is a tiny shader reading a texture of the previous frame, so screen-sized grids run at full frame rate. Reactive use writes the music into the grid: hits seed new live regions, sustained energy tunes the rule thresholds, and the automaton digests the input into growth and decay. The cousin with continuous values instead of binary states is reaction-diffusion, and the family resemblance shows.