I think subvocal communications is the killer app for mobile computing. Accurate and cheap subvocal speech transcription will dramatically transform mobile computing. It seems it’s not really a tech that’s practical yet, but once it gets good enough it will be a bigger disruptor than multitouchscreens, email, or the mouse.
It’s really hard to copy how it works, there’s no one else doing it, and a lot of people will want it more than anything. Seems like it’s a tempting problem for a startup to solve.