SuperPac said:How would it know when you're asking it something versus when you're just talking? That's the big trick. It would be nice, 'cuz having to hit any buttons or interact with an interface is one of the big things holding back voice commands. But it could lead to a lot of pocket-dialing instances where you send something to someone you didn't mean to. Unless the tech is just. that. smart.
Humans usually do it using intonation and context*, the software could do the same, but I don't think we are quite there yet, but it is definitely not out side of the realm of what is possible - especially in the future.
*depending on the language