Cheery Chimp

The idea starts before the game. In the 1940s John von Neumann was chasing a serious question — could a machine build a copy of itself? — and, on a suggestion from Stanisław Ulam, moved it off the workbench and onto a grid of cells that changed state according to fixed rules. It worked, but his automaton was ungainly, with 29 states per cell.

John Conway, a mathematician at Cambridge, spent much of the late 1960s hunting for the simplest set of rules that could still do something interesting. He and his students tested candidates by hand on a Go board, discarding any rule where patterns obviously exploded forever or obviously died out. What survived was a knife's edge between the two: born with three neighbors, survive with two or three.

It reached the world in October 1970, when Martin Gardner described it in his Scientific American column. The response was enormous. Gardner had passed along Conway's standing offer of $50 to anyone who could prove a pattern grew without limit, and within weeks Bill Gosper and a group at MIT claimed it with the glider gun — a structure that fires off a new glider every thirty generations, forever. Through the 1970s, Life is said to have quietly consumed a remarkable share of the computing time available on research machines.

It turned out to be far more than a curiosity. Life is Turing-complete: anything a computer can calculate, some arrangement of cells on this board can calculate too. People have built working clocks inside it, and even a copy of Life running within itself.

Conway, who died in 2020, had a complicated relationship with his most famous creation. He was a formidable mathematician — surreal numbers, group theory, knot theory — and was known to be a little rueful that of everything he did, the one thing everyone knew him for was a game played on a grid.