RE: Your Favorite Boardgame
May 5, 2010 at 2:04 pm
(This post was last modified: May 6, 2010 at 12:43 am by Violet.)
Evie Wrote:I don't know of it being the case yet. You get an AI to control a human character and the AI won't be able to control it as tactically as the better human players. I was talking about the present, not the future. I am talking about the present state of technology.How about that advanced chess program?

Even when it comes to current technological capacity... it is quite possible to create a synthetic intelligence with a greater capacity to think and learn than a current typical human. Cheap? No. Small? No. Blue? I might give you that one.

Quote:Tactics you learn off human players tend to be considerably more evolved and complex than ones just programmed into AI programs in any game I've played."more evolved"? "complex"? Such ambiguity followed by an argument from anecdote?
I highly doubt you have ever played a game against an "AI". You really should stop calling any computer you've played against more than a simple construct.

Quote:I mean simple enough for an AI to be skillful enough to beat a world-class chess player, a chessmaster.We're calling a thing simple because it can beat the best?
That seems ridiculous to me. What then do we call an "AI" that can't beat a chessmaster? "Nomnom-munchi-wakka-ja"?
Quote:Fair settings in the sense of the AI having an equally powerful character or characters (or 'units') as the human player and playing by exactly the same rules, without cheating. Humans can learn strategies that machines still aren't capable of doing yet as insofar as I know. Humans are still smarter than AI opponents. For AI to be more powerful and difficult, it has to be the old-fashioned single player campaign/story mode where on the higher difficulty levels the AI has all sorts of unfair or unbalanced advantages (not even playing by the same rules, or it 'cheating' or whatever).
I disagree with your idea of "fair" is. In my understanding, it does not mean "equal". If indeed (for sake of argument, and I do hope you realize I'm playing the devils advocate yet again) the human player is quite capable of doing more with less... then the computer needs to be buffed in some way (if the idea is to create a challenging fight for the player). To do otherwise is unfair to the computer.
Also, if a machine 'learns' your behavior (which they can currently, especially and all the more so with common SI in the future), it can come up with strategies specifically against what it anticipates to be your strategy. Further... (and I'm referring to common current technology here) some programs are demonstrably much more capable than the human player (In example: chess).
Please give me a home where cloud buffalo roam
Where the dear and the strangers can play
Where sometimes is heard a discouraging word
But the skies are not stormy all day