Scientists have long filled notebooks not just to record their thoughts, but to have them. The act of dragging a fuzzy intuition out of your head and into sentences forces you to confront what you actually mean, what you’re assuming, and what the real question is.
Large language models give us something notebooks never could: a notebook that talks back.
When you sit down to write a careful prompt — spelling out your half-formed theory, your stuck project, or your tangled dilemma — you’re performing the same cognitive labor as journaling.
The difference is that now, instead of a blank page, you get a response that reflects your ideas back to you, often cleaner, better organized, and slightly reframed. In many cases the real advance in understanding happens during the prompting itself, long before the model replies.
