OpenAI, an artificial intelligence research firm which is (of course) funded by Elon Musk, created a team of AI Bots designed to jam some Dota 2, and then put said bots up against a group of semi-professional Dota players.
The team of bots, called OpenAI Five, took on players ranked in the 99.95th percentile in the world, and beat them 2-1 in a set of three games. The human players included former Dota 2 professionals and casters Merlini, Fogged, Cap, and Blitz.
Maybe we’re not the superior gamer species after all?
After winning the first two games without breaking out the cooling system, OpenAI let the audience pick five human players to compete in a sudden death match, which the AI lost.
The Dota 2-playing artificial intelligence isn’t the type to be contained by a single desktop PC. The AI team consists of five neural networks that train using 128,000 CPUs at once, playing 180 years’ worth of matches against themselves every day.
Through this reinforcement learning, the networks learn to cooperate with each other, build ‘team spirit’ and take down in-game opponents. OpenAI has also been tweaking the neural networks, altering the bot team’s in-game reaction time to near-human levels, and helping it learn additional strategies.
With that sort of commitment and teamwork actual humans would be able to achieve literally anything, but we’ve got Netflix and Home Assistants to do things for us now, so maybe tomorrow…
OpenAI Five will compete against the best pro team at The International, which is the largest annual Dota tournament in the world, if everything goes according to plan.
Great work by @OpenAI. Need the neural interface soon to enable human/AI symbiosis.
— Elon Musk (@elonmusk) August 6, 2018
Elon Musk congratulated the (robot) team on its victories, and mentions possible human/AI symbiosis using the neural interface being created to train OpenAI Five, something he’s been keen on for some time now. For someone who is afraid of the robot uprising, Mr Musk really needs to invest in less AI.