The Internet has many virtues (and no small number of vices), but its most surprising effect has been the way it has made research both easy and addictive. While you have to be critical of what you read, of course, at no other time in our history has it been easier to scarf down information like a big bowl of knowledge-flavored ramen.
But this is mainly useful for certain types of knowledge--mainly intellectual, abstract data. For example, when I was in high school I decided to learn how to play the harmonica, which is not a skillset that you can really pick up from written description (although I certainly spent enough time on the HARP-L list, just in case). Likewise, I may have mentioned my recent interest in breakdancing--you can watch a lot of videos and read a lot of forum posts, but I think that's a relatively ineffective way to learn. I don't mean to say that online communities for these activities are useless, because they have value in other ways. But for concrete tasks, you can't beat physical instruction.
So anyway, I'm kind of intrigued by Kinect (and, to a lesser extent, the Playstation Move/Eye or the Wii remote/balance board combinations). We have been working for a while now toward a world where we can query the Internet's store of information based on a macro-level location in space and time, via smartphones. Inventions like Google's local search, and to a lesser degree Foursquare or Yelp, add geographic location to human input. Kinect and its brethren, on the other hand, are attempts to turn the perspective around: interaction based on the topology of the user's body itself.
These early attempts are primitive. They'll be used in crude ways, for gaming and parlor tricks, and they'll have limitations like Kinect's inability to handle prone positions and relatively low resolution. But think of the potential here one that's only hinted at in Harmonix's Dance Central. Among other things, real motion interfaces are a first step toward extending the tremendous communication and educational value of the Internet out into the realm of physical movement. Imagine an educational program for athletic skills that could see your movements, compare them to a model, and tell you how to correct them--or a video chat session with a teacher who could walk "around" to critique your technique in 3D space. Even if it were non-interactive, this could have real advantages--I'd love to have a clean motion-capture of Vic Wooten's slap bass technique to study in slow motion. And surely there are commercial applications, like virtual dressing rooms or telepresence tourism.
Thanks to some literal handwaving, the vision of motion control since Minority Report has been to provide a fancy, grand gestural control mechanism for data manipulation--because there's a problem we've all had, right? In much the same way, the current focus on camera-view augmented reality ignores its real, current applications in relatively dull location-sensitive mapping, probably because most critics are more interested in the human-machine interface than the way these new technologies shape our culture. But surely we should have learned by now: in the age of networked communcation, it's the mundane social uses--chatting, teaching, and sharing--where innovation will get really interesting.