A Bot From Elon Musk’s OpenAI Has Thrashed The World’s Top Players At Dota 2

Most of Elon Musk’s usually center around some grand plan to save humanity. Tesla creates electric cars to help humans tide over the energy crisis; SpaceX sends rockets to space to turn them into a multi-planetary civilization. But every once in a while, Musk’s startups do something that’s a lot more relatable — like play Dota 2.

A bot from Elon Musk’s OpenAI initiative has just thrashed the world’s best players at Dota 2. In a one-on-one exhibition match at Valve’s annual Dota 2 tournament, OpenAI’s bot defeated Danylo “Dendi” Ishutin, a professional Dota 2 player who’s among the best in the world. Ishutin is a crowd favourite and pretty good at what he does — his career earnings from playing Dota 2 tournaments is nearing a million dollars.

But he appeared to be all at sea against OpenAI’s bot. The bot beat him in 10 minutes during the first match, and Ishutin resigned from the second. He declined to play a third. “Please, stop bullying me,” Ishutin said during the match, and also said the bot was “scary.”

Before the tournament, OpenAI engineers at claimed that their bot was better than professional players. They’d trained the bot over just two weeks by playing a “thousand lifetimes” of matches against itself, with “coaching” from the company. Musk said the achievement was the first time artificial intelligence has beaten the best human players in competitive e-sports.

OpenAI first ever to defeat world's best players in competitive eSports. Vastly more complex than traditional board games like chess & Go.

— Elon Musk (@elonmusk) August 12, 2017

Other companies too have been working on their AI efforts. Computers now consistently beat humans at chess, and just this year, a bot made by Google beat the best human player at Go. But Dota 2, which has 113 different characters with different abilities, might have be more complex than either.

And Musk’s successes with his AI efforts will only lend credence on his arguments on the dangers of AI. Musk has been telling anyone who’ll listen that AI could pose a threat to humanity, and needs to be curtailed. Just last month, he’d sparred with Facebook founder Mark Zuckerberg, who had a more rosy view of how AI could impact human civilization. Back then, Musk had quipped that Zuckerberg’s understanding of AI was “limited.” Given how he’s now the co-founder of the company that’s created the first AI that’s beaten competitive players at e-sports, the world would do well to listen.