At the time, it was giant news in the world of coding: In 2006, Netflix offered $1 million to anyone who could improve its movie-recommendation algorithm by 10%. In doing so, CEO Reed Hastings took the unprecedented step of making available the company's huge troves of data from customers. In a recounting of the competition at Thrillist, Dan Jackson likens Hastings to a "tech-age Willy Wonka" who had decided to let "any curious hacker into his digital Chocolate Factory"—but "instead of a chocolate river, he offered a gushing stream of data." Because this was a relatively low-tech contest, amateurs went head-to-head with veterans, and it became surprisingly intense. It would go on for nearly three years, but with a twist: Competitors would share insights with one another in the quest toward the elusive 10%, though they eventually formed into rival supergroups of sorts.
Eventually, one of those groups that included AT&T experts finally claimed the prize in 2009, though a rival group known as the Ensemble tied the mark, only to miss out on sharing the honor because its submission came in 20 minutes late. (The money eventually went to charity anyway.) The odd kicker: Netflix didn't actually use the winning formula because by the time the contest ended, it had shifted into a streaming company and the world was a much different place. Still, "even if the contest didn't play out like many might have first imagined it—one brilliant genius scoring a million dollar jackpot—it instead helped make large strides in the fields of artificial intelligence, machine learning, and recommender systems," writes Jackson. Click for the full story, which notes how outlier films such as Napoleon Dynamite gave the coders headaches.