The GUI's may be good enough, but we still have work to do on the speech recognition and visual recognition fronts.

Ultimately GUIs will be thought as, as something that is just there, the primary form of input will be a hybrid of Touch, Speech, and the computer understanding facial expressions.

That is why MSR has so many projects open for alternate human to computer interfaces, and why speech recognition improvements what been increasing in Windows for the last few versions. Multi-touch going in for Windows 7. It wouldn't surprise me to find some of the facial recognition stuff making it in after that.