Just curious whether anyone’s tried pointing Cowork at your Logseq directory. It seems like the kind of thing where it might do a lot of good, or it might completely destroy everything. I guess test it on a clone of your directory first lol.
1 Like
I would think that you need to begin with your use case before connecting Logseq to any LLM, of course begin in a Sandbox graph, backup first and bla bla bla.
Then you will need to explain the LLM what are their boundaries so it does not take attribution it does not deserve. To me, that’s an idea worth exploring, and in general I am enthusiastic of building a Personal AI system, where Logseq is the UI of all the MDs cocreated with well-behaved and limited AI assistant as a co-thinker. Always waiting for my approval.