I have an excellent voice assistant on my Android phone. I never use it.
I could be dictating this blog post using the same voice assistant. I don’t do that except when I want to prove that I can.
We have had Alexa and Google Home in our home. We shut them off and sent them away.
But we use the Siri voice assistant on our AppleTV all the time.
Why?
Well for one, searching for video content does not have to be exact. Just close enough. So when you say “Allen Iverson crossover Michael Jordan” into your AppleTV remote, even if Siri doesn’t translate that perfectly, YouTube understands it and delivers up one of my favorite basketball moments reliably.
Second, the keypad entry on AppleTV is horrible. I spent some time yesterday setting up apps, entering passwords, etc and it is about the most frustrating experience I’ve had on a computer in a long time.
Siri on the AppleTV is so much better than the alternative and reliably good enough that it has become the way we interact with our AppleTV.
Another example is my car. I have a Jeep and it has this awful smart car UI called UConnect in it. It’s the worst. Except I can say “call Joanne Wilson” while I am driving and it does that pretty reliably. I have called a person we know named Jan Wilson a few times by accident, but that is way better than another kind of a accident in my Jeep.
So while voice imput has not taken hold in our life where text input works reliably and conveniently, it has taken hold where text input is not reliable, convenient, or safe.
What this tells me is the path forward for voice input technology, which has gotten very good, is in applications that are not mainstream yet but can get mainstream by solving the data input problem.
And, in fact, that is what is already happening.
