In "I wrote an AI textbook — how long until AI can do it better?", Nathan Lambert uses his own experience as a technical author to ask a pointed question: when will AI be able to write a better book than the expert who wrote it? The essay treats writing not as a purely creative act, but as a skill that improves alongside model capability.
Lambert's reflection is grounded in the practical work of producing a textbook, which requires structuring complex ideas, anticipating reader confusion, and making judgment calls about what to include. He then turns that effort into a benchmark for AI, asking how long before the models can do the same work at a higher level.
Because this is a single-source reflection, there are no competing claims to reconcile. The piece offers one author's perspective on a fast-moving question, and its significance lies in the source: someone who has done the work is now wondering whether machines will soon do it better.