OpenAI’s super bot has taken this year’s Dota 2 Championships by storm, challenging and defeating its top players at the world stage to test the limits of machine learning and what the future holds for AI in video games.
Super AI Crushes the World’s Top Dota 2 Players
Dota 2 is known for its notoriously competitive player base, spawning massive eSports events that pit the best of the best from around the world. It’s a game that tests the individual and collective mettle of players in both team and 1v1 matches. There is certainly a lot of depth in terms of Dota 2’s game mechanics. You need to adapt to the ever-changing situations to mislead, attack and read your opponents constantly, not to mention the sheer amount of calculated teamwork and coordination it demands amid the chaos of every group clash.
And just recently, OpenAI’s super bot has succeeded in destroying some of the most renowned professional Dota 2 players 1v1. Among those defeated by the AI were Dendi, Arteezy and even solo king SumaiL. To ready itself for the event, the AI was subjected to a crash course of just two weeks to learn the game. Engineers from OpenAI, however, reveal that the super AI actually built up “lifetimes” of experience condensed in such a small human time period. This paints an astounding picture of how the possibilities of machine learning far outpace that of humans in the current age, especially in a video game as demanding of skill and practice as Dota 2.
Asked of his thoughts on taking on OpenAI’s monster of a bot, professional Dota 2 player Dendi recounts that it “feels a little like human, but a little like something else”. He was bested by the bot twice before throwing in the towel, but not without admiring the AI’s formidable skill against an actual world-class player. Check-out the video below for some of the highlights of the matches.
Source: OpenAI (Youtube)
The super AI’s amazing grasp of the complex MOBA title is explained by OpenAI in their blog, which reads:
“Success in Dota requires players to develop intuitions about their opponents and plan accordingly. In the above video you can see that our bot has learned — entirely via self-play — to predict where other players will move, to improvise in response to unfamiliar situations, and how to influence the other player’s allied units to help it succeed.”
OpenAI was founded by tech visionary Elon Musk, known for his firm stand on regulating machine learning and AI before an actual takeover becomes a reality. He is joined by likes of Bill Gates and Stephen Hawking who warn of the eventual collapse of mankind should AI sentience be realized in the not-so-distant future. For now (and with a sigh of relief), they’re just owning us in video games.
Leave a Reply