That's right, the current price is ridiculous. But I think given the tremendous advantages to such a technology (my keyboards hardly ever displayed the "right" labels out of the box), we can expect it to become a commodity in the near future (10-15 years?).

Now, I think there's a little problem with this keyboard... looking up at your keys all the time is not the most productive way to use a keyboard. They market it to geeks, yet I think consumers would find it useful the most.

I agree that it's not very useful for normal typing, but I'd love to have had it when I played World of Warcraft!

I prefer to think in 10-15 years we won't use keyboards (but I guess I wouldn't bet on it - they've been surprisingly hard to replace).

What would we be using instead?

hand tracking, eye tracking and speech recognition are coming on strong with all the power todays computers have.

I don't think the keyboard will ever go away, but a lot of things will be way faster just with hand signals, and subvocalization dictation.

I think I disagree. I can't imagine writing code or a book with voice recognition, even assuming 100% accuracy. The initial dictation isn't the hard part, it's editing that's hard. What commands will we use? "Change that last foo to bar?" That still seems pretty tedious.

I could do without the mouse, though. I would rather press a link with my finger than move my hand all the way over to the mouse, then point, then click.

That's a pretty limited view of the capacity of computers. I'd bet a good 60% of your editing commands are cursor motion. So one pretty obvious improvement is to just look at foo, then do your edit. heck, it's still called cut and paste, how about pushing words around on the screen? not with the mouse, but with your hand on the screen.

The perfect ui is literally, read my mind and do what I want. I doubt we'll see fmri-at-a-distance in my lifetime, but I can indicate my wishes many many ways.

Also, there's some evidence that moving your hands to the mouse is faster, you just notice it, because it's effortless, whereas working out the key combination to move the cursor chews up clock cycles. (http://www.asktog.com/TOI/toi06KeyboardVMouse1.html) And even if you are faster with the keyboard, it probably took you years to get as fast as you are. If programmers can provide a ui to get users to some sizeable fraction of your editing speed in days instead of years... it's worth it. The only real risk is if the fastest of the new ui category can't edit as quickly as the fastest of the traditional. Really, there's no way your hands can compete with looking at what you want.

clearly we need some sort of voice versions of vim and emacs

Even then, it's still faster to press the keys for moving around. But maybe your editor could be in command mode by default, and you could import words by speaking. That might be cool.