During this year's The International World Championships for Dota 2 they held a special one-on-one demonstration, featuring one of the top players in the world going against Elon Musk's world-renown OpenAI artificial intelligence, and it did not go well for the professional.
There's a 24 minute video featuring top player Dendi going against the OpenAI bot, and it's featured on the official OpenAI YouTube channel.
Before getting into the match, the OpenAI team actually tested the AI bot against various other pro players in a closed environment. The other players didn't stand a chance -- we didn't get to see how the bot performed in Dota 2, but we did learn that the bot operates on actual standard playing procedures based on some light tutoring from the OpenAI team.
The actual match in the popular MOBA title gets underway several minutes into the video. The OpenAI bot starts on the right side of the map while Dendi starts on the left side. The match starts kind of slow, with Dendi taking it easy. Both Dendi and the OpenAI bot feel out the process, spending nearly seven minutes just slowly baiting and burning creeps.
After the first match -- where Dendi loses to the OpenAI bot after he gets a bit impatient and rushes the bot -- the OpenAI team explains that the bot managed to get good by playing against itself. They set some really basic parameters for the bot but left it to its own devices in order to understand and excel in playing Dota 2.
After spending a "lifetime" of training against itself, the bot learned how to man the lanes, block the minions, and bait opponents.
It essentially found winning strategies and only seemed to increase its effectiveness from there.
After losing the first round, Dendi tried a slightly different tactic for the second round but lost even quicker yet.
A lot of people in the comment section weren't entirely impressed. Some people noted that in a controlled one-on-one scenario there's a greater advantage for the bot to win, but how well would it have performed in a 5-vs-5 match up?
The OpenAI bot has been in the news for two reasons recently: first, because it defeated a renown Dota 2 player. Secondly, because Elon Musk has been warning people about the dangers of AI.
It may seem like an oxymoron to parade around the OpenAI intelligence system at one of the largest eSports events and showcase how superior it is to hardcore, professional players, while later saying that it could end humanity, but according to Futurism, Musk is more-so using the platform as a way to help "ethically" guide developers on how to construct ethically viable artificial intelligence systems.