"Computers have already beaten the world’s best humans at classic games like chess and Go. Now one has beaten one of the world’s best players of the smash-hit video game “Dota 2.”
Shakespeare for that: “When sorrows come, they come not single spies, but in battalions.” Hamlet, Act 4, Sc. 5
“In a video ahead of the matchup, OpenAI CTO Greg Brockman explained that the company’s special bot was trained by playing a “thousand lifetimes” of matches against itself, with “coaching” from the company. Brockman also boasted that the bot had already bested several pro “Dota 2” players.”
Not until faster computer speeds allow playing a thousand lifetimes per second instead of per month do we need to worry about very general purpose AI (human like). So until then many easy to ID skills like gaming will be knocked off one by one.
“The idea of “self-playing” is key to the advances that OpenAI is pushing for, Brockman told Business Insider following the match. It’s a useful way for an AI system to learn even the most complex tasks. It doesn’t learn anything if it goes up against either a weaker player or a far stronger one. By playing itself, it always has a worthy opponent, he said. “You kind of see this AI go from total randomness” into the game-winning system we saw here, Brockman said.”
This is a tell that artificially intelligent robots may never be programmed for the “emotional need” to socialize but only the “utility” of socializing - with themselves especially.
Tags: Technoskeptic, AI, Robotics, Elon Musk, Gaming