The first time I was told that voice recognition is here and my keyboard is going away was about 20 years ago. Here I am today, still strangling to have Alexa understand simple commands.
Thought controlled input? Forget about it in this lifetime.
Also I remember when computer mice were new, and plenty of people said, "I can't imagine using that."
I think that's true of other types of user interface as well, from haptics to whatever comes next. Sometimes you have to envision what _could_ be, and discover from experience what reaches people... or a subset thereof.
It's a bit boring to think about, but given some of the research that has been able to reconstruct a (poor) image from brain activity, I could imagine using something like that to open applications by imagining their icon.
Continue down this line of thinking and it is entirely possible that we will begin to construct an entirely new symbolic language to interact with these interfaces. I mean if a picture is worth a thousand words, maybe it's more effective path to follow than obsessing over perfect speech recognition.
Voice has gotten dramatically better in the past few years. Another decade? There are also gesture technologies:
https://m.youtube.com/watch?v=0QNiZfSsPc0
Anyway, if you’ve got 3 or 4 decades of life, you still think we won’t have sufficient advances to replace the keyboard?
The radar video is impressive (as were Leap Motion videos, which didn't live up to the hype), and sure it makes sense for extending a watch or phone, but the problem with autocorrect and voice recognition and predictive keyboards and spell checkers is much of the time, I'm not writing clear, coherent, English sentences.
Sufficient advances to write foreign words and jargon and deliberate misspellings and usernames and hostnames and slang and quotes which are deliberately not spelled normally?
A lot of the claims about "you'll just think it" hide the reality - e.g. when I try to speak a password and I can't, because it's just a pattern of motion (and you can't escape this overall problem by saying you won't need passwords in the future).
Human hands are dextrous and sensitive, almost no other body parts are anything like as much - voice is laborious and prone to problems of pronunciation and tiredness and background noise and homonyms, anything else isn't going to come close until you can have brain surgery implants - and even then a) no thanks and b) I bet that still underestimates the complexity of reading clear intent.
Similarly, I think we will see some degree of brain reading, but at least the first generation or two won't replace other input methods.