"Imagine a game machine where you can just pick up a bat and swing it or the tennis racket," said Gates, to which Mossberg noted that one exists, referencing Wii Sports.
"No, no, that's not it," said Gates, "you can't pick up your tennis racket and swing it, ['Oh, I see what you mean,' said Mossberg] you can't sit there with your friends and do those natural things. That's a 3D positional device, this is video recognition, this is a camera seeing what's going on ... the camera will be ubiquitous."
Speech as an input is something we're starting to see more often (e.g. Tom Clancy's EndWar), and touch has been prevalent with Microsoft Surface, university research and the "money-printing" Nintendo DS (though that might fall under his "ink" category, given the use of a stylus). Video recognition is also something that has worked with mixed results, though games like the upcoming Eye of Judgment show a clever use of the technology.
But, returning to Gates example, is swinging an actual tennis racket for a virtual environment the best choice? Later in the interview, from Steve Jobs, talking about the post-PC market: "They don't want to drive with a joystick, they like the steering wheel. And so, you have to ... as Bill was saying, in some cases you have to augment what exists there and in some cases you can replace things." Perhaps in this case, the use of a small, plastic wand as a replacement for a racket is more ideal in an indoor environment. It certainly lowers the risk of breaking anything -- or maybe not.