We talk a lot in The Brief about creative uses of AI, but this one blew our minds, so we wanted to not just tell you about it, but also show you what it can do.

Google has launched a new product called Notebook LM, which they’re calling “your personalized AI research assistant.” The basic idea is that you can load it up with a bunch of information like notes, studies, data sheets, or entire textbooks and it will organize, summarize, and even guide you through the materials.

For those of us whose job it is to collect, analyze, and interpret research, that’s a potentially game changing use case.

ChatGPT, and tools like it, have been able to do something like this for a while. What makes Notebook LM different is how they’ve built a suite of research assistant tools on top of it that would never have been possible before AI.

For example, when you load in your materials, the assistant instantly creates a summary report along with table of contents, study guide, FAQs, and a briefing doc. The summaries also include citations linked to the exact quotes from the sources used.

It also trains itself on your data, which means that you can ask it questions or give it tasks that relate only to the information you’ve shared.

Notebook LM claims to keep all of your data private, allowing you to upload reports that you may not want to share with other open models. *Be skeptical about this one. Yes, it’s Google, but there’s always a risk that sensitive data could get out.

Finally, one feature that really shocked us is what they’re calling “Deep Dive Conversation” where Notebook LM creates a podcast-style conversation between two hosts about whatever you’ve trained it on.

We tried it out on this week’s issue of The Brief, including this section, which created an incredibly meta situation where the robots are talking about themselves.

We’ve exported the conversation so you can listen here.

As you’ll hear, what’s most remarkable is just how human-sounding the banter is. Not only do the hosts present the information, but they discuss, debate, and even sometimes make jokes about the content, all with very human-sounding umms, ahs, and interjections.

Beyond Notebook LM, we see this as an example of a broader trend where we’re going to shift from all-purpose AI tools to ones that are specifically built to help with with all sorts of different tasks.

We can choose to be afraid of AI, or we can get curious about it. The reality is that, just like the internet and other digital tools, people who learn to use these AI tools will be able to do much more in their work. In the near-term, AI will take some jobs, but those jobs will go to people who are skilled at using AI tools.