There used to be second or two between finishing a command and it getting executed. We easily use Alexa 20x a day at my house. Announcements, kitchen timers, lights & music.
I think the voice recognition is async now. It's streaming the data to a model. Before it would wait until the command was finished then send the .wav file off to a model.
I think the voice recognition is async now. It's streaming the data to a model. Before it would wait until the command was finished then send the .wav file off to a model.