The possibilities there are huge. They could certainly do object recognition - a game could not only model and mimic you, but could know when you pick up a cup or soda can. At a most basic level, in social networking environments you could have an avatar truly do what you're doing including munching on chips exactly as you do or taking a sip of beer or whatever. At a more advanced level, you could build object recognition into games - literally take a sip of water or whatever you have around (or mime it with an empty cup) to drink a health potion. With that kind of notion as a starting point, imagine what they could do with recognizing other objects. Imagine the real-world and marketing implications of recognizing specific products. Get a bigger energy boost from drinking a Mountain Dew!I'm a believer. Even if the gameplay implications end up being Wii-grade—amusing, joyful, occasionally engrossing but typically lacking any verisimilitude of real behavior—the ability to to control media playback with simple gestures and voice controls is going to be a big deal. The first time someone flips through a list of Netflix streaming with their hand all the neighbors will be buying Xboxes.