Recently had to lay someone off because they just weren’t producing the work that needed to be done. Even the simplest of tasks.

I would be like we need to remove/delete these things. That’s it. It took some time because you had to just do some comparison and research, but it was a super difficult task for them.

I would then give them something more technical, like write this script and it was mostly ok, but much better work than the simple tasks I would give.

Then I would get AI slop and I would ask WTF are you thinking here. Why are you doing this? They couldn’t give a good answer because they didn’t actually do the work. They would just have LLMs do all their work for them and if it requires them to do any sort of thinking, they would fail miserably.

Even in simple PR reviews, I would leave at least 10 comments just going back and forth. Got to the point where it was just easier if I would have done it myself. I tried to mentor them and guide them along, but it just wasn’t getting through to them.

I don’t mind the use of LLMs, but use it as a tool, not a crutch. You should be able to produce the thing you are giving the llm to produce for you.

This is really interesting. I currently have a z800 running as a backup for my nas, but it’s overkill and uses a lot of power. I also had to do some weird stuff to get set up properly. This setup might actually work for me. I have a bunch of mini pcs I could probably use. I just need to be able to add a 10gbe nic and I’ll be good to go.

I don’t even know if that’s being maintained. I couldnt find anything on docker hub. It does look like this one is maintained though https://blog.pixelfed.de/2020/05/29/pixelfed-in-docker/

Maybe give that one a try. It seems like you are already using his image anyway. Also it seems like the command to run for the worker is different in his image.

If you have issues with that setup, feel free to post what issues you having.

Technically you don’t really need to setup a bunch of networks unless you are trying to host this or something, which I wouldn’t recommend docker compose anyway. So, if you wanted to keep it simple, you could just the default network.

Another note, while zknt/pixelfed does exist on docker hub. It was moved to quay, so continue using that image.

Not sure what pixelfed is, but you yelling I need help doesn’t really help anyone. What problems are you having exactly. What image are using? What commands are running? What have you tried?