Steven Johnson has written 13 books, on subjects starting from a London cholera outbreak to the value of video games. He’s been a television presenter and a podcast host. He’s a keynote speaker who doesn’t must name himself that in his LinkedIn profile. And for over a yr now, he’s been a full-time worker of Google, a standing that’s clear when he badges me into the search big’s Chelsea places of work in New York to point out me what his workforce has been creating.
It’s known as NotebookLM, and the best manner to think about it’s as an AI collaborator with entry to all of your supplies that sits in your metaphorical shoulder to information you thru your mission. NotebookLM was soft-launched to a choose group earlier this yr however is now accessible to all as an “experiment”—that’s Google’s low-risk solution to see how the app behaves and the way we behave with the app.
Johnson discovered his solution to Google by means of a lifelong obsession with software program as a “a dynamic thought associate,” a instrument to hurry up and improve the inventive course of. When he was in school he grew to become obsessive about HyperCard, Apple’s software program that broke information into chunks and allowed you to navigate an information-space by way of hyperlinks. It anticipated net navigation earlier than the online existed. “I fought mightily to show HyperCard into that dream instrument, but it surely wasn’t fairly prepared,” he says. He ultimately grew to become an fanatic of Scrivener, a mix phrase processor and mission organizer widespread with e-book authors. (I’m a fan too.)
When Johnson bought entry to OpenAI’s GPT-3 text generator in 2021, he recognized that AI may stage up a brand new era of thought instruments. Oh, wait, he mentioned to himself, this factor that has at all times been behind my thoughts is now going to be doable. Situations unthinkable even a yr earlier than had been immediately on the desk. Johnson didn’t but know that Google not solely had related massive language fashions, however was already engaged on a mission very a lot consistent with his considering. In Could 2022, a small workforce within the experimental Google Labs division cold-emailed Johnson. They arrange a gathering by way of Starline, a Google Labs mission that permits for eerily intimate in-person conferences. “I principally had a dialog with a hologram who mentioned, ‘, this factor you have been chasing your entire life? We will lastly construct it,’” Johnson says. He grew to become a part-time adviser to the small workforce, at first sharing the workflow of knowledgeable author. “Here is 4 or 5 engineers, this is an precise creator, let’s simply watch him,” is how Google Labs head Josh Woodward sums up the method. Ultimately Johnson bought concerned within the improvement of the product itself and was sucked in to the purpose of accepting a full-time gig. His title at Google Labs is editorial director.
NotebookLM, initially known as Undertaking Tailwind, begins by creating a knowledge set of your supply materials, which you drag into the instrument from Google Docs or the clipboard. After the app has digested all of it, you may then ask NotebookLM questions on your materials, due to Google’s massive language mannequin know-how—partly powered by its just-released upgrade Gemini. The solutions mirror not solely what’s in your supply materials but in addition the broader basic understanding of the world that Gemini has. A vital characteristic is that each reply to your queries comes with a set of citations reporting the place precisely the data got here from, so customers can test the accuracy of its output.
Google will not be the one firm envisioning merchandise that permit folks create customized information units to discover with LLMs. At OpenAI’s developer day final month, the corporate launched customized mini-GPTs that may be tuned to a particular process. Woodward acknowledges a “core similarity.” However he argues that NotebookLM focuses extra on enhancing a workflow, and is geared to offer superior accuracy in its outputs. Additionally, he says that the OpenAI merchandise have extra of a character, whereas NotebookLM is designed to don’t have any such pretensions.