The Checkers Solution...
Recently, researchers at the University of Alberta have created a computer who can't lose at checkers. The computer, or 'Chinook', can't be played to a draw...but will never actually be beaten by a human counterpart.
While my first question was "They haven't done this yet?", feeling that codes and algorithms should have accomplished this already. Surely my first computer, (a brand spanking new Commodore 64) would have an unbeatable level to its checkers games. My first question was quickly followed by my second, (predictably) "Who cares?"
But then I stopped, thought, and realized how difficult this must have been. No matter how you play, no matter how a computer plays, the best you can do is a draw... it is mathematically impossible to beat this machine at a simple game. Regardless that there are a finite number of moves in checkers, it's a significant step in the Artificial Intelligence staircase.
In the same article, it references that computer that will go on to play in a poker championship against 2 human players.
What's next?
A computer program that can't be beat at Chess? That step is a scary one, as in essence the moves are finite, as in checkers, but strategic planning is much more complex.
Is it a comforting to think that computers could be eventually be unbeatable at everything?
Comments
So why don't they apply this technology to warfare and have it come up with a way to end them peacefully?
Now that would be comforting.
Ridiculous notions can also be infinite.
Posted by: Jenuine | July 23, 2007 05:01 AM
I think what is scary is that the computer can not be beat by a person, making Artificial Intelligence greater than human intelligence. I will not be terrified though, until a computer's emotional intelligence is proven greater than a human's.
Until then, let's distract these super computers by pitting them against one another while we enjoy food, sex, fresh air and all the other pleasures of human life.
Posted by: Heather Lynne Morrison | July 21, 2007 03:40 PM