Perhaps I'm missing something, but there is nothing you can currently do with one of these multi-touch surfaces that you couldn't do just as easily with a multi-buttoned mouse - and with better precision.
That video spoke of the benefits ("increased bandwidth") of using five fingers instead of one pointer, but didn't show much use of this alleged benefit.
One good example though is at 6:35 where they mention using two hands - one to drag a window while using the other to drag the application space. But the demonstration in the video wasn't too impressive, as they didn't really do both things at once.
This does look like a case of "We have this new technology, what can we do with it?" rather than "How can we improve the way we work, and what new technology should we invent to make it happen?" Not that that's a bad thing in itself, it's just not as ground-breaking as it sounds at first.