Earlier quoted context omitted.
Because the computational requirements of good speech rec exceed what you would build into a $100 device, and because doing it in the cloud makes training and improving it easier. The “wake word” is local, which is why it is constrained to a few choices. I suspect we are only a few years away from being able to do a slightly inferior version using local processing, or a private cloud (already possible today when ther…
It's the reason everyone says Siri sucks - they do all the processing for her on the device, not the cloud.
But Siri sucks for MANY reasons, and not just technical ones. It does worse processing but also does worse cloud integration, and now vs. alexa, has a far worse ecosystem around it (alexa "skills" are pretty awesome, and trivial to create)