I’d like to express a contrarian view. And that is that even though we’re rushing to make narrow AI that can understand vocal commands, that isn’t going to be the overwhelmingly accepted way that we communicate with our devices. I mean, if you think about it, there’s a lot of things you probably can do faster with an old keyboard than you can do with your voice. You’re looking at an email on your device, and you can hit “Forward” and start typing the person’s name and hit “Send” probably faster than you can say out loud, “Forward this message to so-and-so.” In addition, typing something is less error-prone than saying it. And for a long time to come, the vocal AI is going to ask for confirmation, “You want to forward this email to so-and-so, correct?” And you’ll have to say, “Yes.”
Audio is interestingly quite slow as an interface compared to typing for simple things. If you think about it, there are already programs that do a pretty good job of doing dictation, but most people still type things themselves. In addition, audio has problems with privacy. You say things out loud, and anyone within the vicinity can hear it. Whereas, if you type it, they can’t. Finally, audio doesn’t really work that well if you’re in a place with a number of people, like at work or in a room with other people. Having audio commands is a distraction to everyone and they will end up stepping on each other’s toes. So I think the keyboard, that 120-year-old device that Mark Twain wrote one of his novels on is actually going to be with us for a long time to come.