The Setup

Before starting, he loaded NotebookLM with everything he could find:

  • 6 textbooks on the subject
  • 15 research papers
  • All the course notes and syllabi he could locate

Then he did something most people don't do: he asked questions that forced synthesis, not summary.

The Questions That Actually Work

Most people ask AI to "summarize this" or "explain this topic." That generates a compressed version of what's already in the material. It's efficient, but it doesn't build understanding.

The Stanford student's questions were different:

"What are the 5 core mental models that all experts in this field agree on?"

This forces the AI to identify the foundational frameworks — the things that experts take for granted but beginners don't know to look for.

"What are the 3 questions that divide experts in this field? What does each side believe and why?"

Understanding where experts disagree reveals the actual frontiers of a field. It also surfaces the assumptions that are contested vs. settled.

"What's the most counterintuitive thing about this field that surprises newcomers?"

First-principles knowledge often runs against common sense. Knowing where intuition fails is as valuable as knowing the core frameworks.

"If I had to make a decision in this field tomorrow, what would I most need to know?"

This forces practical prioritization. Not "what's interesting" but "what's load-bearing."

Why This Works

The method isn't really about AI. It's about asking questions that surface structure rather than content.

Content is easy to find. Structure — the mental models, the debates, the counterintuitive findings, the decision-relevant knowledge — is harder to extract and usually buried in papers that take weeks to read.

AI with sufficient source material can extract structure quickly. The limiting factor is knowing what to ask.

Applying This to Your Work

You don't need 6 textbooks to use this method. You need:

  1. Enough source material to give the AI real signal (PDFs, articles, documentation)
  2. Questions that ask for synthesis, not summary
  3. A second pass asking the AI to test your understanding

For solo founders entering new markets or technical domains, this is worth incorporating. The 48-hour deep dive isn't a substitute for years of experience, but it's a dramatically faster way to get oriented than reading one thing at a time.