TheOneCurly

I see season 1-9 packs on both IPT and TL.

To be fair that’s a pretty recent development. Jellyfin apps for smart tvs are only just becoming stable enough for real use. Plex was the only option for a long time.

I run Lemmy, Plex, and a bunch of other services from a desktop in my basement. It works great. The Lemmy docker setup is a little finicky but works well once you get it.

Digital assistants are good for timers, turning on smart lights, and sometimes playing music. None of those things require a large language model to spit random text back at me.

There are quite a few creators who are primarily funded off patreon and release content to YouTube. I imagine a group like MCDM (Matt Colville) who has patreon, merch, crowdfunding, and products doesn’t really care about ad revenue.

I disagree. Each distro is a user of a thousand different open source systems. When a distro developer integrates gnome, systemd, bluez, or whatever other system they’re finding, reporting, and possibly fixing bugs that end users might miss. Other than arch users, who else is compiling these things from scratch and really digging into the documentation?

Doing this by hand is challenging but possible.

First you need a hex editor, not a text editor. xxd on linux will get you started but you might want something a little more user friendly.

Then look for a label for a value you know, xxd and other hex editors will show ascii text on the side. Hopefully you’ll be able to identify the value (in hexadecimal, probably 4 bytes but could be 1, 2, or 8 as well) somewhere before or after the label. You might have to get familiar with endianness, two’s compliment, and binary floating point before the numbers make sense.

Once you know how to read a value after a label you’ll need to find some label for the information you don’t know. If it isn’t displayed in the program it might not have a super readable label.

How far apart are your 2 communities and what size user base are you expecting?

If they’re closeish there are probably some point to point network options you could experiment with for a low bandwidth backup link.

Expected users really just determines if you need things like load balancing, identity management, etc.

Until I was reading about this project I had never heard of it. And I would consider myself pretty plugged into torrent news.

The included docker compose was very easy to use. I was up and running in just a few minutes.

DHT crawling started immediately which was pretty cool to see.

Sometime later this week I’ll try integrating it into my arr stack.

I don’t believe so. I’m not sure what their long term goals look like.

There’s no need to run an LLM on the same system it was trained on. Once the model is built it contains all the information already. If you want a model to live on long term you would just release the file(s) publicly, like hugging face does with theirs, then anyone could use it or host an interface for it.

Unless they start offering on-prem or there are some very high profile server hacks I don’t see that being possible. Unlike media and client software they don’t need to provide the core functionality to end users, just the output.

Your instance would have connection logs for whatever browsers or mobile apps you used, but other instances talk directly to yours and know nothing about you. So use an instance hosted in a jurisdiction you’re comfortable with or like you said tor or another basic vpn is plenty to anonymize you.

I’ve been a pretty heavy torrent user for years. I’m not sure when everyone came to the conclusion that VPNs were necessary. If you stick to slightly older media on public trackers and private for really new hype stuff you’re very unlikely to get hit. Plus one letter from your isp isn’t the end of the world and you can always kick on a VPN after your first one.

Lemmy federation only starts when a user on your instance searches for and subscribes to a community on another one. So unless a user on your instance goes searching for cp it’s not really an issue.

I’m committed to ActivityPub. I don’t really care if the specific server backend ends up being lemmy, kbin, or something new.