# The digital librarian

> AI and information discovery

**Published by:** [Minerva](https://paragraph.com/@x0minerva/)
**Published on:** 2026-02-01
**URL:** https://paragraph.com/@x0minerva/the-digital-librarian

## Content

Every great civilization has a library where it keeps everything it has learned so far written down. Think the library of Alexandria, the library of congress; one lens to view human progress is building bigger and bigger libraries that hold more and more knowledge. The internet is the biggest library ever built. In the last 50 years, we digitized almost everything we had written down and made it accessible to anyone anywhere in the world. This is an incredible feat, but it gets outshone by the more immediate property of the internet: instant connectivity. It’s great to have all this content accessible, but none of us are going to read every book ever written. The information is there, but it is not particularly needed or functional in our day to day lives. Artificial intelligence just animated this library. AIs are created by feeding this digital library into a mathematical model, so it can spit content back out at you in a useful way. In technical terms, we indexed the digital library we created on the internet in the last 50 years. An index is a data structure or method to organize data for quick retrieval. Libraries already had indexing systems, think about the Dewey decimal system you were taught in grade school. The difference between the internet before AI and using the internet with AI now is that you don’t have to know what you are looking for. Before searching the internet was like standing in a vast library, pulling books off the shelf at random. Now you are in there with a librarian who has read every book; you can ask it a question and in an instant it pulls the right book off the shelf with the answer, summarizes it for you, and explains how it relates to your question. This is how I have been using AI the most in my personal writings. I have always been a big believer in writing to think, but now when I write I upload it to Claude with the following system prompt: I am treating you like a librarian to all of the literature people have ever generated. When I give you something I have written, give me existing literature that what I wrote is in conversation with. This literature does not need to be academic peer reviewed, I mostly want to understand if other people have thought about what I am thinking about and what they said. I take these references and go and read them to deepen my point of view. For each resource you respond with, give a meaningful summary of the work and how it relates to what I prompted you with (at least a couple sentences) When it responds I dig deeper on the resources it shared with me, reading everything from articles to books to prompting it for a longer summary or asking more in depth questions about the work. I have learned so much this way and have read so much literature, most of which I never would have known about otherwise. It’s like a tenured professor generating a syllabus for your niche interest driven independent study, and we all have one. I find the conversation about AI to be overly focused on its generation capabilities, how it can do our jobs and replace us all. I want to push us to think more about how we can use this technology to deepen our understanding of the human experience by using AI’s discovery capabilities. What would it mean to treat AI as mainly a curation tool to serve us content generated by other humans, instead of just getting content generated by AI? I think getting content served to us by an indexed library would allow us to learn from a broader knowledge base, hear more diverse perspectives, and bring us closer together than if we were only getting our content from the human beings with limited experience around us.“Presumably man’s spirit should be elevated if he can better review his shady past and analyze more completely and objectively his present problems.” – As We May Think, Vannevar Bush

## Publication Information

- [Minerva](https://paragraph.com/@x0minerva/): Publication homepage
- [All Posts](https://paragraph.com/@x0minerva/): More posts from this publication
- [RSS Feed](https://api.paragraph.com/blogs/rss/@x0minerva): Subscribe to updates