In the first episode of Thinking with AI, I ask a deceptively simple question: when AI explains something clearly, does it actually deepen understanding, or does it sometimes only create the feeling of understanding?

Using a structured dialogue with a language model, this episode explores the difference between fluency and comprehension, between assisted performance and durable understanding, and between AI as a substitute for thinking and AI as a scaffold for it.

The conversation moves from principle to practice: what would have to be true for AI to genuinely deepen understanding, and why might real-world use drift instead toward convenience, outsourcing, and the appearance of competence?

This episode features structured dialogue between me and a language model. The AI voice is synthetic, and its contributions may be based on multiple prompted drafts edited for clarity, rigor, and length. The framing, interpretation, and conclusions are my own.