I actually want this feature, but I want to own the data. There are some OSS projects writing basically identical things but they aren’t too popular (https://github.com/jasonjmcghee/rem seems to be the most popular I could find, but I wasn’t able to get the cross-platform version running on my machine).

I also wrote the dumbest possible clone of this feature in bash, the basic data gathering steps are actually pretty easy to do. I’d build this into a real program but I’ve just been too busy lately with other projects: https://jackson.dev/post/cloning-windows-recall-in-30-lines-of-bash/

I highly doubt that they actually managed to do this, at least any time recently.

As another commenter noted, Android alerts you when an app is accessing the microphone in the background, and it would also absolutely destroy the phones battery life more than the FB app currently does. The only way that we have the “Hey Google/Siri” command prompts active all the time is with custom hardware not available to the apps, and certainly not without Android knowing about it.

Maybe they actively listen while the app is open, but even then I think recent Android/iOS would let you know about that.

That is not the same thing as listening in the background.

0? It’s anchored to the sea floor.

I actually prefer to just integrate the AI stuff into my usual chat apps [1]. I don’t see much point in building a whole new app (except to get more control over the settings, but for most people even that’s probably unnecessary)

[1] https://jackson.dev/post/chaz/

If you just run a personal private network, then yea pick anything because you can change it fairly easily. Companies should try to stick to things that they know won’t change under them just to avoid issues

Together.ai has a number of uncensored models too. I’ve found that those are so cheap that it’s not worth trying to self just models unless you really need more privacy.