I also started with Lemmy on a small VPS where I run a lot of other services for me and my family. Every update of Lemmy somehow needed more and more resources to run, so many in fact that it would overload my small VPS. I think it is because Lemmy is designed to be run for å huge amount of users, so it has quite a overhead at first just to be able to handle very many users later.
And because I didn't want to upgrade my server and make it more expensive, I decided to try PieFed which has from the start been concentrating on having a small footprint on the server.
And that played off emensly, not only was I able to keep running on a small VPS for cheap, but I also found an amazingly open developer community around PieFed on Matrix.
So the suggestion is the people who voted for this government have no idea what they were doing and now the workers need to go against their own government to save America? If that would be so then democracy doesn't work in America.
This is probably only a problem with the online version. In contrast to google and openAI they, like meta, let you download the model and run it offline, where they can't access any of this data I presume.
About the things we can't explain yet, we should be honest and the only moral thing is to say we just don't know. Lying and pretending otherwise is immoral and wrong.
Asking a LLM about itself leads to a lot of lies. Don't do that mistake like I did.
I asked llama if it sends data back to meta and it said yes it does. I thought that's big news and wrote a blog post about it, because it was supposed to be offline, etc.:
In China this led to something they call 'lying flat', in Korea it's called N-po generation.