Artificial Intelligence Learns to Learn Entirely on Its Own
Quanta Magazine

A new version of AlphaGo needed no human instruction to figure out how to clobber the best Go player in the world — itself.
Spread the word