Dwarkesh uses LLMs like Claude as active reading companions, transforming passive consumption into deep understanding by generating insights and spaced repetition flashcards. He applies this method to master complex topics—from ancient DNA to philosophy—before interviews, allowing him to engage experts at a higher level. His workflow emphasizes compounding intelligence: each new insight builds on prior knowledge, making future learning faster and more meaningful. By organizing ideas through concept clustering and long-form synthesis in tools like Apple Notes and Mochi, he constructs a coherent worldview over time. Early adoption of AI and spaced repetition has been crucial, enabling him to decode dense material, such as Nick Land’s writings or Wittgenstein’s philosophy, and retain what he learns. The discussion also demonstrates Claude in action, helping prepare for an interview with geneticist David Reich by analyzing ancient DNA research and human migration patterns. Ultimately, Dwarkesh treats learning as a lifelong project, using AI not just to gather information, but to become a more thoughtful, knowledgeable thinker.