1) Doesn't always understand what I'm saying (voice recognition)
2) If it does understand (voice recognition), it actually changes it it something else (super dumb, no, idiotic "AI", changes someone I call almost daily into someone I never call)
3) If it asks me what I want to do with a contact, it doesn't understand "call"
4) There's no way to 'correct' what has been said.
5) I can't open "podcasts", because it will open Deezer (I don't use Deezer anymore, but it's still on my phone).
It's really just really really a dumb command line interface. And if it would just be that without the 'smart' things, it would be a lot more helpful. There are only a few things I use it for:
1) Call XXX
2) Alarm at XXX
3) Open google maps. Sometimes I say "route to xxx using google maps", and that works
Even these three common use cases fail about 10-20% of the time.
When I'm in the car and want to change route or add a poi, want to open an app, or text something, I try, but I get so frustrated which is more distracting than just typing it in your phone. So that's what I do.
Google's voice recognition is a lot better and faster. Also their search understands more context.
Voice assistants are like a skinner experiment. Solution: make it very rigid in terms of operations. People are quick to adapt this. Somehow the ai-crowd doesn't understand that a UI works best when the things you operate with are at the same location and always respond and work in the same way.
I'd compare the voice assistant experiment as a GUI where the buttons, their function, and labels always change.
1) Doesn't always understand what I'm saying (voice recognition)
2) If it does understand (voice recognition), it actually changes it it something else (super dumb, no, idiotic "AI", changes someone I call almost daily into someone I never call)
3) If it asks me what I want to do with a contact, it doesn't understand "call"
4) There's no way to 'correct' what has been said.
5) I can't open "podcasts", because it will open Deezer (I don't use Deezer anymore, but it's still on my phone).
It's really just really really a dumb command line interface. And if it would just be that without the 'smart' things, it would be a lot more helpful. There are only a few things I use it for:
1) Call XXX
2) Alarm at XXX
3) Open google maps. Sometimes I say "route to xxx using google maps", and that works
Even these three common use cases fail about 10-20% of the time.
When I'm in the car and want to change route or add a poi, want to open an app, or text something, I try, but I get so frustrated which is more distracting than just typing it in your phone. So that's what I do.
Google's voice recognition is a lot better and faster. Also their search understands more context.
Voice assistants are like a skinner experiment. Solution: make it very rigid in terms of operations. People are quick to adapt this. Somehow the ai-crowd doesn't understand that a UI works best when the things you operate with are at the same location and always respond and work in the same way.
I'd compare the voice assistant experiment as a GUI where the buttons, their function, and labels always change.