Garry Kasparov had to deny it. A team calling itself ZackS had just won an online chess tournament where players could use as many computers as they liked, and people wanted to know which grandmaster was hiding behind the name. Kasparov said it wasn’t him.

It turned out to be Steven Cramton and Zackary Stephen, two friends who played chess in their spare time. Neither held a title. Stephen was rated 1398, a respectable club player and nothing more.

Grandmasters entered that 2005 event with strong programs and decades of study. Cramton and Stephen brought three ordinary PCs. In the final they beat a grandmaster who had a highly rated partner working alongside him.

In a tough position, the two of them would agree on three or four moves that looked promising. Then they ran those lines through different programs to see which ones held up. Asked about their edge, they credited their “knowledge of each chess engine used.”

Kasparov chewed on that result for years. In 2010 he boiled it down to a formula, writing that “weak human + machine + better process” beat a strong computer alone, and also beat a strong human with a machine and a worse process.

That formula is about to matter far beyond chess.

Before long, machines may turn out new fields of math and science faster than any person could visit them, a thousand at a time. The menu will run longer than a lifetime, and it will be written in a language nobody fully reads yet. Answers will be cheap. And choosing which path deserves your years will be the hard part.

Kasparov gave the credit to process. But process comes second, though. Before any engine ran, two hobbyists looked at a board and felt which few moves deserved a closer look.

“This power belongs to everyone of us, although it manifests only in the degree that we are able to respond to it,” wrote Yogi Ramacharaka of intuition. “It grows by faith and confidence, and closes itself up, and withdraws into its recesses when we doubt it and would question its veracity and reality.”

A machine will have an answer ready before your own thought finishes forming. If you always let the machine answer first, you slowly stop trusting your gut. Do that for ten years, and when you finally face that endless menu, you won’t know how to pick.

So guess first. Before you ask the search bar or the chatbot, write down what you think, and then check. You’ll be wrong often, and every check trains the hunch a little more.