I'm convinced that multitouch trackpads are a good idea, but I'm not convinced they're a big selling point. The ones in the MacBooks and MacBook Pro have a sort of multitouch, and if that were merely enhanced slightly we'd have the things like rotation and zooming -- the things that computer users can find fairly useful, I think, but generally don't use daily unless they're pro.
The future of computing, I am convinced, is going to stick with the keyboard and trackpad for a long time. The touchscreen, for most applications, is not much of an improvement.
I envision the distant future being a bit like Star Trek: The Next Generation -- they do simple things through voice communications. Database queries ("Computer, are there any inhabited planets in the Analingus system?"), media control ("Computer, play 'Echoes' by Pink Floyd. Louder. Louder. A little softer."), and similar functions are all performed through vocal commands.
Real work, though, is still done at a terminal. To be sure, they have futuristic equivalents of Wacom tablets, but when any sophisticated task is required, Data and La Forge are there typing, typing, typing. The keyboards are specialized, admittedly, but still keyboards. Touchscreen technology (direct manual manipulation of the object) is used, but not for real work -- more for presentations, demonstrations, hypothetical screwing around.
This is good tech, to be sure, but what Apple should be working on more is unifying applications at a low level -- making their functions available to the OS and other applications in an easier way. AppleScript, then Automator have been great leaps in this direction, although still relatively opaque and difficult to configure on-the-fly.
I imagine a day when I'll walk up to my computer and say "Computer, what books do I own that involve psychological disorders?" and the computer will respond "twenty-three." And I can say "How many of those have been made into movies which are in my library?" and the computer will respond "three." These are simple database queries that almost any n00b with a copy of iTunes and a suitable Library app or LibraryThing can accomplish, but they are largely out of the reach of vocal commands.
We have gotten so caught up in the visual capabilities of computers that we tend to forget they can communicate (and rather well) in audio. There is an enormous capacity for progress in that area, despite the large and ever-present problems with vocal recognition, text-to-speech, etc.