BEFORE YOU PRESS PLAY
What is discussed
AskMonty AI editorial noteSource consulted:
During a workshop on AI agents, Montemagno wants a visible example of assistants working while he continues presenting. He asks his system to create an ebook explaining how to use agents, what to do and what mistakes to avoid. He reports that an initial draft arrives in about an hour and that the material is eventually condensed into roughly 200 pages. Cover changes, discussion with participants and requests for additional content follow. The headline therefore concerns the main generation stage within a longer process, rather than establishing that every aspect of producing and reviewing the book took only sixty minutes. The request does not begin without context. The system already has access to thousands of his video source materials, previous books, workshop slides and accumulated guidance about his communication style. These materials shape the subject matter, structure and voice. Montemagno believes the result captures his jokes and phrasing and compares favourably with his earlier business books. That is his own assessment, tied to the kind of writing he produces. His practical emphasis is on providing personal guidelines: without them, the AI will organise and express the material according to its default preferences rather than the author's distinctive intentions. He says he used Claude Code, partly because it already held useful context about his work. He distinguishes running the tool on a personal computer from running it on a VPS, or remote server. He prefers the latter because the work can continue independently of his laptop. The episode presents Hostinger as a partner and sponsor and describes a simplified installation process. Model access brings an additional cost beyond the server. He also notes trade-offs: remote operation is convenient but reduces direct control over the environment, raises questions about sensitive data and requires attention to security when systems are exposed online. In Montemagno's account, the system launches agents with distinct responsibilities: gathering information, checking claims, selecting material, writing, proofreading and offering critical counter-readings. It also produces title and cover suggestions. This resembles a small editorial team organised around a shared outcome. Assigning an agent to fact-checking describes the intended workflow; it does not by itself establish that the finished book contains no errors. Montemagno also involves a human colleague with writing expertise, asking that person to read the ebook and give feedback. The example ultimately combines automated production with human decisions and review rather than removing those elements altogether. The broader argument concerns the difference between originating a work and physically writing every sentence. Companies and public figures already rely on collaborators, he observes, and he openly describes using people to turn his instructions and voice notes into previous books. For practical business nonfiction, where ideas are organised and illustrated with examples, he sees agents as an extension of that collaboration. The author's contribution can remain the direction, substance, recognisable style and final judgement. Delegating execution does not necessarily mean those contributions disappear, although it changes the process through which the finished text comes into existence. Montemagno limits the comparison by distinguishing a practical guide from a literary work. He acknowledges forms of personal and artistic writing that should not be treated simply as information assembly. His prediction mainly concerns authors becoming responsible for a coherent direction while working with both human and artificial collaborators. He ends by encouraging direct experimentation with one's own material. In his account, production becomes dramatically faster, but the prepared context, choice of subject and editorial review remain essential to understanding what the experiment actually achieved.
Key points
- A workshop experiment: the speaker describes a roughly 200-page ebook about AI agents, with an initial draft generated in about an hour and further work afterwards.
- Substantial context already existed: thousands of source materials, previous books, slides and voice guidelines informed the writing.
- Claude Code on a VPS: the described setup runs agents on a remote server and involves both model-access and infrastructure costs.
- Security and separate responsibilities: the episode discusses exposure online and agent roles for research, fact-checking, writing and editing.
- Authorship as direction: for business books, Montemagno compares agents with the editorial collaborators authors already use.
- Scope and human review: he distinguishes practical nonfiction from literature and asks a human team member to read the completed ebook.
