The market for smart speakers boggles my mind. Are people really willing to sacrifice their privacy for a minor upgrade in usability? Is it really that hard to just tap play on your smartphone instead of yelling out "Alexa play Despacito"?
Smart speakers are not all that scary if you know how they work and that information, in detail, is all out there. They go to a low-power mode waiting for the keyword and only then do actually turn on the main CPU and do any processing.
They may the safest least privacy leaking Internet device you actually own. Your TV is probably taking screen shots of what you're watching. Your phone is leaking data to dozens of different entities all the time. Your computer is constantly sending stuff out. A smart speaker is tame by comparison.
I also have my phone on me, all the time, even while at home so really my privacy issues aren't increased by having that Alexa.
If you ask me, device owners should have unimpeachable control over what data, if any, is sent from their devices.
Standalone GPS navigators are perfectly capable of navigating without a network connection. The only thing you lose is real-time traffic info, which in a better parallel universe would be broadcast in the open by every DOT.
But the utility is too great to ignore. My smartphone is integral to how I communicate, travel and even run my business.
Smart speakers are just gimmicks
I also put more trust into Apple/Siri than I do with Alexa or Google. Their differential privacy and anonymized Siri requests have limited the speaker to few features which in my experience work well with the latest 13.x OS.
The one caveat, I do know someone who is blind who uses the speech-to-text on her Google device a lot. She has all the fonts set to their maximum size, high contrast colors, and various other accessibility features enabled. Speach to text helps her greatly.
Really? The Mycroft project claims to be just that, and it shows promise, but it still seems really clunky. And right now, as far as I can tell, you need to set up an account on their service so it can do things like voice recognition, which is a pretty critical component to this type of thing.