Damn thats insane. I remember hearing about the study and it being a big step in medical field. How did nobody peer reviewing it look at the images and compare them. Shows a broken system
There is Mycroft.ai which can do homeassistant skills but its still centralised as the processing power required for Stt and context processing is still very expensive.
He’s not wrong but I think we are running out of ways to not use advertising to keep the lights on.