System's theorist Niklas Luhmann famously published 90 books in his lifetime. He created a type of pre-hypertext card sorting system for cataloguing his own ideas. Each index card held an atomic idea. After enough had accumulated, he could literally "surf" his own mind and connect ideas this way, much like a wiki. He had a second file system for bibliographic data. He called this system his Zettelkasten.
Starting in the summer of 2025, I started creating an automated version of the bibliographic Zettelkasten to help me write my PhD. Using AI, I process and ingested the entire corpus of literature I was reading. The books were converted into a vectorized database that was searchable by an LLM, thus greatly expanding the amount of text that the LLM could process. The vectorized database was then annotated with links back to the item in my zotero library so any annotations could be directly traced back to their source. A summary spine and list of concepts were produced and then fed back into the vectorized database so that any textual processing would then become searchable. Summaries and concepts were also rendered as markdown files for use in the Obsidian text editor environment.
After this work was finished, I could then start creating "synthesis" concepts that would pull concepts and excerpts across the entire corpus of literature to help me find examples of ideas as they iterated through multiple authors and books.
