Three major problems I see: 1. When these systems go offline, they’re nearly useless. One WiFi glitch and the system gives me the audio equivalent of a blank

Three major problems I see: 1. When these systems go offline, they’re nearly useless. One WiFi glitch and the system gives me the audio equivalent of a blank stare. This directly reduces the chance I’ll casually use voice and instead prefer my more-reliable phone. 2. Most systems haven’t figured out reasonable responses and are basically chatty and full of sounds. Make it Unix-like (silence on success = golden)!!! If I say “turn on the light” and you turn it on, I CAN SEE THE LIGHT ON so I don’t also need to hear a loud chime and some voice confirmation like “sure, no problem”! The fact that I can silently do things quickly in other ways (e.g. phone) is another strike against voice. Yet this is something they could easily fix. 3. Voice systems are not 100% perfect at comprehension yet they tend to babble out long responses. This puts me in the situation of trying to shut them up for long enough to listen for my intended query. Maybe they can improve this by erring on the side of fewer words in replies, with more pauses? Not sure.