

One thing that I find useful is that you can do --filesync with an existing zip, which means it only updates the files if the date/size don’t match. This saves a bunch of time on large archives that don’t need to be created from scratch every time.
One thing that I find useful is that you can do --filesync with an existing zip, which means it only updates the files if the date/size don’t match. This saves a bunch of time on large archives that don’t need to be created from scratch every time.
How does this compare to something like openwebui https://docs.openwebui.com/ ?
I have a self-hosted AI system that works pretty well. I can interact with it via my phone, the shell, my IRC server, and I can verbally talk to it.
But I want to get it to remember things, so I need to start working on RAG or something. Eventually I’d like to be able to have it draft emails for me, and schedule appointments.
There’s a fun movie about it: https://www.themoviedb.org/movie/125414-the-final-member
Wouldn’t that be getting rid of the libs?
I think this would be ‘owning’ the libs:
chown -R $(whoami) /usr/lib
If the responsibility is being shelled out to “the algorithm”, then doesn’t that mean they have less responsibility? Shouldn’t they be paid less if they have less responsibility?
Something tells me they won’t see the logic in that though.