NotebookLM | a writer’s workshop

NotebookLM* is one of the most fascinating (and potentially game-changing) uses of AI I’ve yet seen.

Practically, it’s an AI-powered assistant-with-a-scrapbook for any digital pieces of content (“Sources”) you throw at it.

In my first test with the tool, I gave it the eBooks to all three works by Neil Postman I’ve been working at for my second pathway and asked it some preliminary questions about their content, unifying factors, and trajectory if seen in light of a transcript for a long-form podcast conversation about AI and media today I uploaded additionally.

Firstly, it was able to read, interact with, and “understand” the primary thrust and trajectory of Postman’s arguments over the course of all three books and their ten-year unfolding through his later life. As I’d read each of these books in concert with each other numerous times and had been writing about their content and interconnectivity for the past six months, I felt confident checking its work.

It ended up checking mine.

Ruling out confirmation bias because of its thorough citation of passages in each book, NotebookLM deftly corroborated my theories and reading of Postman’s primary motifs to affirm the direction my analysis had taken. More than that though, it gave me deep reasoning about the citations it provided to outline his reasons for shifting perspectives on our media crisis across his literary canon.

To read and pull helpful quotes in my research would have been enough. Exceeding my expectations, this “writer’s workshop” performed an analysis I didn’t know I needed.

I’ve written before about AI as a tool to adjust the aperture of the mind- never before has this resonated so clearly.

For my followup training run, I tested its Mind Map feature.

It generates 25-30 topics, themes, keywords, and ideas from the sources and allows you to click in and drill down into them. As a learning or deep-dive tool, this is revolutionary.

Imagine the ability to converse with a friend over coffee about some minutia within some trifling subject matter that you’ve accrued expert status in over the years. This friend can pull as many (and certainly many more) citations as you can, and it can engage in a deep, thoughtful analysis of the work in front of you both. As you question it, throw around concepts, and hop across the breadth of a field or niche only you typically know the hooks for, it creates connections you’ve never been able to come across and picks out those insights that you only stumble upon in the true depth of dialogue.

There are absolutely things lost in the adoption of this tool.

Is serendipitous learning stolen by the LM’s thinking? Maybe (I apologize for anthropomorphizing so much, but that’s where we are in 2025).

Are there communities and book clubs and workshops that won’t be created around shared ideas and a desire to learn and grow? Probably (but social media and the internet have already done much to disband common places of unity over the past two decades).

But to level the playing field to those without access to these communities of thought and to provide the background and catalysts for those who have never been shown how to collaborate, learn, and ideate around deep work like this is invaluable.

And I can’t help ending with the refrain- we can’t go back…

*The most recent update I’m utilizing that expands the token count to larger than you can imagine, adding Audio Overview with interactive mode, alongside a streamlined 3-panel UI takes home the prize for best LLM-interface to date, in my book.