I’m sorry but you’ll need the Intel cpu if you want jellyfin. You can’t control what devices your users will play the media from, and eventually transcoding will be needed. I think the amd field isn’t doing great just yet on this regard. Any old Intel cpu past the 8xxx series will have enough transcoding ooomph to handle a bunch of simultaneous transcodes, once set. But yeah you need to make sure you have acceleration for the transcodes.

Also having a GTR4 with GB, same idea. Still hate having to send them data by having to use their app on an initial pairing to get the damn key for the first time. Also, now that I changed phone, and using the pairing key on the new phone worked to get it to connect to my GTR4, it seems a bit less persistent connection than when it was paired from their app on the same device. Not sure what else changed, because it’s in both cases pixel phones, running GrapheneOS.

What’s your consideration for choosing this one? I would have thought ViT-B-16-SigLIP2__webli to be slightly more accurate, with faster response and all that while keeping a slightly less RAM consumption (1.4GB less I think).

It’s…kinda like a VM? But without the VM part. It runs in a container, AFAIK, so it’s using a lot less resources than a full blown VM. It works for a lot of apps. And the ARM emulation plugin helps a lot, too. But then again, I usually stick to mostly FOSS apps, and refuse to install the gapps suite. So, no Play Store. I can still install apps via Aurora, but there’s a problem there between Waydroid and Aurora, which leads to frequent crashes (of Aurora) when trying to install an app. But once the app is installed (you can download it by other means, and just install it into waydroid by running something like ‘waydroid app install myfile.apk’.

You mean…waydroid? It’s literally a translation layer running on a container, AFAIK. Then you can add an additional ARM emulation plugin for specific apps that don’t have x86 versions.