In chess computer uses big database of moves played by humans. Computer will never have imagination, improvising, bluff computer can do some things better but they are limited. Chessboard is limited by size and strict rules. Humans are not limited by rules.
the new verson of alphago don't need input data, only game rules then it trains by playing against itself. After 3 days of training, the new version wins against the old one. The old one had won 4-1 against one the best player of go in the world, in march 2016. The new version crushed the old one 100-0. Go players more and more get inspiration in some unexpected moves seen performed by alphago during games against human.
Yes, @isAAAc, THAT is the problem. We will truly become the "useless eaters" they claim we are in their eyes -- further justifying our elimination (again, in their eyes). Machines will always be inferior to us no matter how specialized they are. We can do so much more.