is working on a new
type of speech recognition. Rather
than analyzing sounds, it detects nerve impulses in the throat.
The user does not have to speak out loud.
Instead, the user merely thinks about speaking, much as one
does when reading silently.
This sounds pretty neat, although the accuracy is not very good yet.
I could imagine this being used for things like cursor
control, or any number of other things, if the accuracy can be improved.