Elon Musk’s OpenAI Proven Superior Over Mark Zuckerberg’s Facebook, Slaughtered In StarCraft

Tesla CEO Elon Musk might be the loudest voice of caution in the development of artificial intelligence right now, but he is still proving to be a revolutionary in helping to create advanced technology. Back in August, Musk’s non-profit OpenAI created an AI that beat a professional DOTA 2 player. In contrast, Facebook recently entered its own AI in a StarCraft match and was soundly defeated.

Back in July, a brief exchange of barbs between Musk and Facebook Founder Mark Zuckerberg occurred when the social media figure essentially called the Tesla boss “irresponsible” for saying that companies should be careful when developing AI. Musk fired back and said that Zuckerberg’s understanding of AI is “limited.”

Based on the performance of Facebook’s AI in a StarCraft match where all the participants were machines, it would seem that this “limited” understanding extends to Zuckerberg’s employees, as well. This was proven when the social network’s AI research lab quietly entered a bot called CherryPi, which apparently made some poor decisions during the match and was slaughtered, WIRED reports.

The match was part of an annual competition to show off how far AI technology has come. It was meant to be Facebook’s opportunity to finally show something of substance from its AI research division, which is reportedly made up of 80 researchers and led by NYU professor Yann LeCun. The lab has been producing numerous papers on machine intelligence, but data can only get them so far.

What’s even more humiliating about the defeat in the StarCraft match was the fact that the winners were made by lone programmers who, by all accounts, were simply hobbyists. In an effort to save face, Facebook tried to pass off the loss as a means of creating a “baseline” for future projects. On the other hand, now that corporations are starting to get in on the action, future StarCraft competitions might just end up getting bagged by major tech firms.