Naughts and Crosses

Tic-tac-toe games have been a common computer science exercise since the 50's, so I thought, why not?

This is an associative learning tic-tac-toe AI, based on positive feedback (check back later for a complete version with negative feedback as well). After a few thousand plays, the AI should be smart enough to beat you most of the time...

http://www.box.net/public/dql7x7633e