🎧 Is NotebookLM—Google’s Research Assistant—the Ultimate Tool for Thought?

We use it to find bestselling author Steven Berlin Johnson’s next project

Every illustration/Pulitzer Center.

TL;DR: Today we’re releasing a new episode of our podcast AI & I. I go in depth with Steven Berlin Johnson, editorial director of NotebookLM and Google Labs and bestselling author of 14 books. We dive into the new version of NotebookLM, Google’s AI-powered thought tool, that launches today. Watch on X or YouTube, or listen on Spotify or Apple Podcasts

I sat down with bestselling author Steven Johnson to see if we could come up with a concept for his next project—using AI. 

We loaded 200,000 words of interview transcripts that NASA conducted and all of Steven’s reading notes since 1999 into NotebookLM, Google Labs’s personalized research assistant. We wanted to see if it could help us explore the story of the fire that broke out on a preflight test of Apollo 1 in 1967. 

The model condensed the disparate transcripts into readable formats like FAQs and chronological timelines. It sifted through the material to identify the catalyst for the fire. To top it off, NotebookLM even went through Steven’s Readwise notes to find a relevant, and unexpected, story from history that we could use to explain the origins of the fire. It was a wild ride, and we screenshare through the process live on the show.

NotebookLM uses software to organize ideas—something that Steven Johnson has been toying with for over three decades, first as a college student who built an app to organize his class notes, and now as the editorial director of NotebookLM and Google Labs. Steven is also the bestselling author of 14 books, including his latest release, The Infernal Machine, which chronicles the rise of the modern detective. (As an author himself, Steven emphasized that any material uploaded to NotebookLM is only sent to the model’s context window or “short-term memory,” and is not used to train the AI.) 

This is a must-watch for anyone who is a fan of Steven Johnson’s work, or is interested in AI as a creative tool. Here’s a taste:

  • Embrace emerging connections. Steven collects his notes in one place and uses software to discover connections between them, believing that categorizing them himself will restrict his ability to make novel links—a principle that has influenced the design of NotebookLM. “I'm on the side of emergent chaos…don't organize it and just let things bubble up, and figure out tools that will let that bubbling up happen,” he says. 
  • Create an expert AI companion. A user would start by creating a new project in NotebookLM and uploading material that is the “source of truth” for the project, like journals, research documents, or quotes from books. “Everything inside of NotebookLM is grounded in the documents you provide…and once they’re uploaded the model becomes an expert in the information you’ve shared,” Steven explains. 
  • Build a unified workspace. The user can interact with all the material they have uploaded on one cohesive platform. “The whole interface is designed to let you load a lot of different documents, move back and forth between those documents, or potentially read those documents while you're working…we want to have a single integrated surface where you can do all that work,” Steven says.
  • Generate concise recaps of dense texts. A user can upload multiple sources for each project, and NotebookLM has a “source guide” feature that summarizes the information in each one. Steven thinks this is useful because “generally a source is on a single topic and you can get a high-level” overview of the content.
  • Customize source material for high-quality responses. When a user poses questions to NotebookLM, they can select the sources they want the model to prioritize when generating its response. “[Y]ou're able to shift the focus of the model to various different things really, really easily,” Steven says.
  • AI that can present unique perspectives. Users can ask NotebookLM a wide range of questions because the model is trained to recognize concepts like “interestingness” and “surprise” in the uploaded material. “I often say, ‘What's the most surprising idea here?’ because you think about that as an author, and the idea that that can be like effectively a search query is just totally bonkers,” he explains.
  • Get the lay of the land with the Notebook Guide. NotebookLM has a new feature called the “notebook guide,” which provides a bird’s-eye view of your research project, generating things like a table of contents, timeline, and cast of characters. Steven says it's useful when “somebody gives you a bunch of files [and] you're trying to make sense of it.”
  • Lean into the harmony of multiple intelligences. Steven reflects on our creative process in conceptualizing the opening scene to the documentary about the Apollo fire, noting that NotebookLM allows for a “fusion of so many different, separate intelligences”: the original authors of the source material, the user’s curatorial eye, and the AI model “synthesizing all these things and making connections possible.” 

Steven’s approach to curating valuable information

Steven believes that over time, software like NotebookLM will be able to understand one’s “general sensibility” enough to “scour the internet for things that could be useful” for them. Until then, this is how Steven curates information:

  • Specific points of research. When Steven conducts research for a specific project, like a book, he directs NotebookLM to sift through the sources that he has uploaded for the project. “I could very easily tell NotebookLM, ‘These are the key themes of the book, help me find passages that are relevant to those themes,’” he says.
  • Information that isn’t immediately relevant. While deep in the research process, Steven saves information that he finds interesting even if it isn’t directly relevant to his present query. “[Y]ou get a little glimmer of, ‘That could be something,’ and so you save those things, even though they don't have a slot to go in, or don't have a chapter to go in, or even a book to go in yet,” he explains.
  • Unfamiliar albeit thought-provoking content. Steven bookmarks information that he finds interesting even when he doesn’t understand the nitty-gritty of it. “[W]hen I read something I understand, I’m like, ‘Well I already know this on some level, so I don’t need to save it,’ whereas when I read something…that’s provocative, but I don’t really get it, I often try and save that,” Steven says.

You can check out the episode on X, Spotify, Apple Podcasts, or YouTube. Links and timestamps are below:

  1. Introduction: 00:00:53
  2. The roots of Steven’s obsession with organizing ideas: 00:04:47
  3. Steven’s belief in being receptive to organic connections: 00:08:16
  4. NotebookLM’s north star: 00:11:16
  5. How Google Labs’s note-taking app respects copyright law: 00:14:02 
  6. Using NotebookLM to analyze highlights from books that Steven has read: 00:16:17
  7. Steven’s personal approach to curating valuable information: 00:19:02
  8. Why NotebookLM shows restraint when it is asked to speculate: 00:21:51
  9. Using the model to co-create the beginnings of a documentary: 00:29:13
  10. NotebookLM generates the opening script: 00:52:54

What do you use ChatGPT for? Have you found any interesting or surprising use cases? We want to hear from you—and we might even interview you. Reply here to talk to me!

Miss an episode? Catch up on my recent conversations with LinkedIn cofounder Reid Hoffman, a16z Podcast host Steph Smith, economist Tyler Cowen, writer and entrepreneur David Perell, founder and newsletter operator Ben Tossell, and others, and learn how they use ChatGPT.

If you’re enjoying my work, here are a few things I recommend:

The transcript of this episode is for paying subscribers.

Thanks to Rhea Purohit for editorial support.

Dan Shipper is the cofounder and CEO of Every, where he writes the Chain of Thought column and hosts the podcast AI & I. You can follow him on X at @danshipper and on LinkedIn, and Every on X at @every and on LinkedIn.

Like this?
Become a subscriber.

Subscribe →

Or, learn more.

Read this next:

Chain of Thought

How Sora Works (and What It Means)

OpenAI's new text-to-video model heralds a new form of filmmaking

1 Feb 16, 2024 by Dan Shipper

Chain of Thought

How Hard Should I Push Myself?

What the science of stress tells us about peak performance

2 Oct 17, 2023 by Dan Shipper

Chain of Thought

Admitting What Is Obvious

I’m a writer—what are you?

26 Sep 8, 2023 by Dan Shipper

Thanks for rating this post—join the conversation by commenting below.


You need to login before you can comment.
Don't have an account? Sign up!

Every smart person you know is reading this newsletter

Get one actionable essay a day on AI, tech, and personal development


Already a subscriber? Login