To me, an A.I. defeating the best player in a chess game cannot conclude anything about A.I. being better than human brain. (At least not at the current technology level)
Human brain can create ideas with the surprise factor. If you ask an A.I. program to surprise you, it will just generate some RNG BS. Chess game is well-understood for thousand years, pretty much no new move or new strategy can be created anymore.
Putting all knowledge and experience about the chess game into an A.I. program, which would not make any mistake on the game, and perform more calculation and predictions in a second than any human can do; pretty sure it shall always win-or-draw the top player unless it has bug.
I don’t understand why those top players would accept this kind of ‘challenge’, it is both pointless and not fitting the philosophy of chess game. We all know that copper and semiconductor transmit electrical signal much faster than a lump of fat.
If an A.I. program can create a new viable and efficient build in Grim Dawn, then I will give it a darn.
Having said these, I wouldn’t say A.I. can never create ideas tho. We don’t even know how human brain/mind can come up with surprise factor. Maybe when we do, we can simulate the mechanics/procedures in an A.I. program and A.I. program can then truely create ideas.
Generally, I agree with the sentiments about AI. However, if you’re referring to recent news, the computer AI Google’s AlphaGO, beat two of the world’s best GO players. Not chess. GO has a practically infinite number of moves, unlike chess. The AI doesn’t process the entirety of all possible moves to win - its “intelligence” comes from narrowing down its choices to the few best possible combinations after learning a bit about the competitor’s play style.
But, I agree AI has a ways to go to understanding the subtleties of the human mind.
GO (now I know the English name) is named as a type of chess in my language:rolleyes:
But simpler task doesn’t mean A.I. are better than human tho.
If we ask A.I. to clean a glass with a piece of cloth using a robot arm and visual recognition to identify dirt on the glass, the A.I. could clean the glass well. But if we replace the glass with a colored glass with pattern, the A.I. will probably rubbing the glass forever…
I’m pretty sure I could do this, given an API into the skills/items/enemies. But having to write that API myself is basically the reason why I haven’t bothered.
With mamba here. GD is a far simpler numbers game than chess is.
I guess it depends on how you’re dividing it. You’d need to model a whole bunch of stuff and create a lot of functions to parse the game mechanics. Where does the human input end and the AI input begin? I think it will take a while for a human to prepare such an AI, even if the AI’s part in it is simple.
This is an area where AI doesn’t quite pass muster: generalized abstraction. Computers don’t do abstract concepts well. To know what something ‘is,’ a computer has to be told what something is first and every possible variant and variation on it. As opposed to a person who, once given a generalized abstraction of a concept, can easily extend it to all manner of applicable subjects and materials.