Elon Musk's OpenAI bot beats human player in Dota 2 tournament

Share

The computer player was built, trained and optimized by OpenAI, Elon Musk's AI boffinry squad based in San Francisco, California.

As part of a special demonstration for the crown, the OpenAI bot defeated Dendi easily in the first of three rounds, with the victor being the first to two kills or destroying an in-game tower. "This guy is scary", a shocked Dendi told the huge crowd watching the battle at the event.

"Maybe if I played it a lot of times", Dendi ponders as he loses, perhaps questioning whether fighting back against the machines is futile or not.

NASA is sending a supercomputer to the space station next week
Atop SpaceX's Dragon spacecraft, which is slated to take off next Monday, the first supercomputer will be delivered to the ISS. A "smart" spacecraft, however, will require a considerably more powerful and robust computer.

Engineers from OpenAI, which is backed by Tesla and SpaceX CEO Elon Musk, explained that their bot earned a "lifetime" of experience after two weeks of real-time Dota 2 training against itself, eventually working up to go undefeated against other professional players, including Artour "Arteezy" Babaev and Syed "Suma1L" Hassan.

It's not only Google's AI that has the potential to defeat humans in their own game.

"League of Legends" (pic) and "Dota 2" have dominated recent PC gaming revenues, says DFC. Was he indirectly comparing the OpenAI bot with AlphaGo? The bot learned the game from scratch by self-play, and does not use imitation learning or tree search. "And now, we want to see how it stacks up against the best players in the world". It learns to predict its opponent's movements and pick which strategies are best in unfamiliar scenarios. From what Brockman said, it seems that the company is hoping that the bot keeps on learning as quickly as it now is, and improve upon its cooperative techniques to take on a full team of professionals before long. Elon Musk himself has expressed concerns about the growing threat of AI and the need to regulate it "before it's too late".

Share