Why no LLM could replace Leo Strauss, at least not yet

I was talking to a buddy yesterday, and he made the claim that the University of Chicago probably still has a strong philosophy program largely because of Leo Strauss, even though he was there like 75 years ago.

At first, that sounded like a stretch. But then he started pointing out how these great teachers tend to create intellectual lineages that persist for generations. People like Allan Bloom or Harvey Mansfield trace back to Strauss, and then their students trace back to them.

And it’s not just philosophy. You see the same thing in music, people tracing their style or training back to composers like Beethoven or Mozart. He even mentioned that Alex Karp studied under Jürgen Habermas, who comes out of the Frankfurt School of thought (created first by people like Adorno and Horkheimer).

But this struck me as odd at first because all of their ideas are… available. You can read the same books Strauss read. You can listen to the same Beethoven pieces. You can literally ask AI to explain any of it to you, step by step. Information has never been freer. And with AI, you can probably learn anything you want.

This begs the question: why isn’t this sufficient? What is so special about being in Stauss’ classroom?

I think the difference is something like… judgment? The value of having Strauss teach you philosophy is not so you can better understand the text or transfer knowledge; it’s to discover his approach to figuring out what questions to even ask about the text. What can you notice about how and why what you are reading was written? What seems convincing and what seems shallow? And the key thing here is that this looks different for everyone.

The observations you make about the ‘hidden’ meanings in a text are personal, and there is no right answer (though there are probably wrong ones). An LLM can certainly teach you the Straussian school of thought. It can explain the methods, highlight patterns, and teach you what others have written. That said, I don’t think it can fully teach you to notice things the way Strauss would. Specifically, those things that are shaped by your own curiosity and judgment. It can’t perceive what excites you or draw out what makes your perspective unique. The Straussian method is a frame for looking at things; it’s up to you to chase the ideas and insights that spark your curiosity. Maybe a prompting god could push an AI in that direction, I’m not sure. What I am sure of is that with lazy prompting, I don’t think any modern LLM is close to achieving this.

This is true not just for philosophy, but for music, science, or even leadership. The lessons of great teachers are embedded in the way their students notice the world, ask questions they care about, and draw out their uniqueness. It is notably NOT just about knowledge transfer or training a skill.

Maybe one day AI will nudge us toward our own curiosities, but for now, the spark that a great teacher lights in us remains irreplaceable.