Google NotebookLM aims to be the ultimate writing assistant

Steven Johnson has written 13 books, on topics ranging from a Cholera outbreak in London to The value of video games. It was a TV presenter And a Podcast host. He’s a keynote speaker and doesn’t have to call himself that on his LinkedIn profile. And for more than a year now, he’s been a full-time employee at Google, a case in point when he invited me to the New York search giant’s Chelsea offices to show me what his team was creating.

it’s called LM notebookThe easiest way to think about it is to be a collaborator with the AI ​​and have access to all the materials on your metaphorical shoulder to guide you through your project. NotebookLM was initially launched to a select group earlier this year but is now available to everyone as a “trial” – Google’s low-risk way to learn how the app works and how we behave with it.

Johnson found his way to Google through his lifelong obsession with software as a “dynamic thought partner,” a tool to accelerate and enhance the creative process. When he was in college he became obsessed with HyperCard, Apple’s software that divides knowledge into chunks and allows you to navigate the information space through links. He predicted web navigation before the web existed. “I fought hard to turn HyperCard into that dream gadget, but it wasn’t quite ready,” he says. He eventually became an enthusiast of Scrivener, a built-in word processor and project organizer popular with book authors. (I’m a fan too.)

When Johnson gained access to OpenAI’s GPT-3 text generator in 2021, I confess That artificial intelligence can raise the level of a new generation of thinking tools. Oh wait, He said to himself, This thing that has always been in the back of my mind will now become possible. Scenarios that were unimaginable even a year ago are suddenly on the table. Johnson did not yet know that Google not only had similar large language models, but was already working on a project very much in line with his thinking. In May 2022, a small team in Google Labs’ experimental division emailed Johnson. They set up a cross meeting Starline, a Google Labs project that allows for eerily intimate in-person meetings. “I had a conversation with a hologram that said, ‘You know, this thing you’ve been chasing your whole life?’ We can finally build it,” Johnson says. He became a part-time consultant to the small team, initially sharing the workflow of a professional writer. “Here are four or five engineers, and here is an actual author, let’s just keep an eye on him,” Josh Woodward, head of Google Labs, summed up the process. Eventually, Johnson got involved in developing the product itself and was absorbed to the point of accepting full-time work. His title at Google Labs He is the managing editor.

NotebookLM, originally called Project Tailwind, starts by creating a dataset from the source material, which you drag into the tool from Google Docs or the Clipboard. After the app digests everything, you can then ask NotebookLM questions about your materials, thanks to Google’s Big Language Model technology — supported in part by the just-released Gemini upgrade. The answers reflect not only what is in your source, but also the broader general understanding of the world that Gemini has. An important feature is that each answer to your queries comes with a set of citations that indicate exactly where the information came from, so that users can verify the accuracy of its output.

Google isn’t the only company envisioning products that let people create custom datasets to explore with LLMs. At OpenAI Developer Day last month, the company introduced small, custom GPT machines that can be tuned for a specific task. Woodward admits there is a “fundamental similarity.” But he says NotebookLM is more focused on workflow optimization, and is geared toward providing superior accuracy in its output. He also says that OpenAI products have more personality, while NotebookLM is designed to have no such pretensions.