This turned out to be something on my end, but leads to a question about how the behavior of the app should be.

I did some digging, turned out this was due to some certs or something on my home router that didn’t like sh.itjust.works. The web interface shows a lemmy.world mirror of the image, but Boost shows the image from the original post on sh.itjust.works.

The question is, should the app be pulling from the original source? If that source instance goes down, Boost won’t be able to show it anymore. That seems against the philosophy of the Fediverse (and the actual Lemmy website shows the mirror).

Thanks for checking. I thought maybe it was my choice of theming or some weird state I’ve confused the app in, so I cleared the data in the app but the problem persists.

Yup, you’re right, good catch 🙂

You need to report two numbers for a classifier, though. I can create a classifier that catches all cases of autism just by saying that everybody has autism. You also need a false positive rate.

That is an incredibly overgeneralized view of the situation. It’s precisely this kind of polarization, and viewing everything in terms of black and white, that Putin and allies are reveling in.

Are you aware of anything like this for text messages? Google messages filters a lot, but a) it’s not perfect, and b) it’s app specific

I’m running Immich on an Odroid N2 and it’s great! https://immich.app/