Writing With Your Second Brain

Two new videos about the research and writing process—and the ways AI might transform it.

This installment of Adjacent Possible is short on text, and long on videos for you to enjoy. The first is a 90 minute conversation I recorded last month with David Perell, for his excellent How I Write YouTube/podcast series. If you haven’t followed this series and are at all interested in the writing/research process, I highly recommend it. David gets terrific authors as guests (present company excepted) and really goes deep into their craft and their workflow. In our conversation, we spent some time discussing NotebookLM and the whole idea of software as a kind of collaborator for the writing process (which dates back to my use of Devonthink in the early 2000s.) But we also discussed in some depth the thinking I went through to figure out the structure of books like Ghost Map and Good Ideas, along with a number of the techniques I’ve developed over the years to keep myself from getting bored with my own material (which is a serious problem when you’re writing a book over months or years.) There’s also a fun section talking about the role of surprise both on the level of the paragraph and the overall architecture of a book or essay.

Watching the interview after David posted it, it occurred to me that while there is no shortage of problems with the contemporary state of digital media, one truly amazing thing about the current landscape is how many deep dive conversations like this are now possible. There were basically no media venues where you could talk about craft for 90 minutes when I started out as a writer. (Maybe the old Paris Review “Art Of Fiction” interviews came the closest.) Now we have series like How I Write, not to mention a whole host of other podcasts in the Sam Harris mode, where you have the amazing luxury of time to really get to know an author and the way their mind works.

The other video I wanted to share is from Tiago Forte, author of the book Building a Second Brain, which explores in much more practical detail many of the techniques for working with research material and notes that I discussed in the “Creative Workflow” series here at Adjacent Possible. Tiago shares my obsession with the potential of software tools for thought, and he’s a wonderful teacher. I reached out to him early in the development of NotebookLM to show him what we were up to, and just last week, he released a detailed how-to video describing the best practices for using NotebookLM as a second brain tool. One of the things I love about the video is that he explores a series of use cases that go beyond the typical “chat with a PDF” scenario; there’s a great sequence where he takes the transcript of a meeting he’d had, and uses NotebookLM to turn the information from the conversation into a business proposal.

After watching Tiago’s video, I thought it would be helpful to create a doc that gives you a step-by-step guide for integrating outside research material (from books or web pages) into NotebookLM—most importantly, using ReadWise’s new “Export to Docs” features, which has been optimized for NotebookLM. If you are a Kindle reader and have been highlighting passages in books you read, you can use ReadWise to easily export all those quotes into NotebookLM, and effectively work with version of Google’s Gemini model that has become an expert on all the important ideas from your reading history—exactly the kind of “second brain” that Tiago describes in his book. I’ve shared a few of these examples before at Adjacent Possible, but it never ceases to amazing me that I can sit down and my computer and ask questions like this one below, and get answers with this level of sophistication in seconds.

(Speaking of sophistication—note how Gemini drops into the Russian in the first answer; that phrase apparently means “single organism.” We see this every now and again with Gemini; it switches to another language for a word or phrase, including different alphabets, the legacy of its multi-lingual training. I once asked it a question about John Searle’s famous “Chinese Room” thought experiment, and it dropped into Chinese briefly in the answer, which I thought was pretty hilarious. It’s considered a bug technically—I imagine it will be fixed at some point—but I prefer to think of it as Gemini just being slightly pretentious.)

If you’re interested in these research applications for NotebookLM, I posted the guide at Medium: “How To Use NotebookLM as a Research Tool.”