By the 1900s, mathematicians had broken mapmaking down to a simpler set of rules, and eventually were able to reduce the sheer number of possible maps to a manageable set of types that could be classified and colored one by one. Of course, manageable is in the eye of the beholder—that set still contained 9,000 maps. By the 1970s, computers were becoming more accessible, so mathematicians were able to use algorithms to narrow that set down even more. In 1976, mathematicians Kenneth Appel and Wolfgang Haken reduced the number to 1,936, then tested every single one to make sure they could all be filled in with four colors. They checked and re-checked their findings on different computers with different algorithms, and their results were the same. Finally, they had a proof that showed that you can't have a map with a number of regions that would require five colors.

*Related: Road Trip! Why Mathematicians Are Vexed By The Traveling Salesman Problem*

This was the first time a computer had been used to solve a math problem, and it was understandably controversial. According to the University of Cambridge, "Many mathematicians and philosophers claimed that the proof was not legitimate. Some said that proofs should only be 'proved' by people, not machines, while others, of a more practical mind questioned the reliability of both the algorithms and the ability of the machines to carry them out without error." Of course, the four-color theorem had been "proved" by people twice already, and both proofs had been faulty. Since then, mathematicians have embraced computers and computer algorithms, but that doesn't mean the argument is obsolete. It has echoes in the debates over autonomous vehicles today: people are nervous about the risks of letting a computer drive, even though human error is the reason for 94 percent of car crashes. Distrust of technology is a tale as old as technology itself.