Learn something from the game industry. God damn!

I wonder how much of that is simple state machine / prescripted actions. Does the robot learn how to do these rather abatrary games, or is it instructed how to.

I think the main thing to work on is to figure out what subsystems are involved in human thought... from small things like vision, spacial reasoning, and work from there to related things. like recognizing objects and deciding where in space they are. Then predicting movement of objects Then perhaps manipulating objects. then perhaps predicing objects if you manipulate them. etc...

enough subsystems would make a pretty beliveable human.

Artificial intelligence is easy if you've got the help from the creator of the universe, a nice link up to precisely what and where everything is in the world and the safety of knowing your virtual machines never need break down or fault in any way.

Artificial intelligence that works in the real world doesn't have that luxury.

Yep, that's actually a pretty big achievement.