How does this work with:
1. Touch interfaces like tablets, that do not have Alt, or any, keys?
2. Our "muscle memory" that acquires reflexive knowledge for the right spots to tap/scroll/click for frequently used apps and functions? That is always a big issue for ANY kind of interface change.
I don't really need to "search" for items I use on a regular basis - like my car's controls, my hands/fingers know just where to go for what I want without looking (much). Now if it could read my mind for my "intent", that would be cool (and creepy...).