The photo recognition and search would be nice, but creating a 3D environment from pictures of various location wouldn't look quite so beautiful and seamless. Different people use different cameras, different quality, different lighting conditions. Obviously the pics used for the demo were taken with one camera all at once in a fixed location.
Interesting stuff though. Photo, audio and video recognition and search ability is probably going to be the next big step in web surfing. I know Google is working on this, but I don't have any sources on info to share.