However, you left out one important category---adaption...
Ahh, real AI! Yes, that would be nice, wouldn't it? I guess I was focusing more on "simulated" AI. Well, the main problem with this aspect is usually the memory. You have to store that learning somewhere, and it takes a lot to become meaningful! But yes, that would be a good ultimate goal! I'm looking forward toward seeing your game. I am (slowly) developing a game myself, similar in genre to the mario games. But right now I'm stuck on the mathematical physics of it (not my strong point...). Hopefully I'll be done with the first level in a few months...here's his profile:


Very dorky, but my neices and nephews love him! They've been bugging me to get him to do more than just walk and jump, of course...

Anyway, just let me know if you get stuck on anything. Good luck!