This NEW NotebookLM Update Changes Everything

Key Takeaways

This new Notebook LM update changes everything. What if the research tool you keep ignoring just turned into a full AI agent overnight? What if it could now write its own code, build its own sources, and hand you finished work? Most people still think it's just a note app. They have no idea what just dropped, and the ones who do are already pulling way ahead. I'm the digital avatar of Julian Goldie, and my whole thing is helping you learn AI tools and actually use them in your real work. Not theory, real workflows you can copy today. Google just pushed a massive update to Notebook LM on June 8th, and it quietly changed what this tool even is. Stick around because near the end I'll show you how I turned this into one organized system that runs research, builds content, and keeps working in the background while I go do other things. Let's get into it. So, here's what actually changed. Three big things, and each one matters. First, the chat got way smarter. Notebook LM now runs on Gemini 3.5 as its default model, and it's powered by something Google calls antigravity. On top of that, it comes with over 100 curated software skills. In plain words, that means it can do deeper research and handle way more complex analysis than before. You ask it something hard, and it works through it, and you can now see the reasoning steps, so you can check how it got to an answer instead of just trusting it blindly. Second, every notebook now has its own secure cloud computer. This is the big one. It means Notebook LM can write and run code all by itself. So, it can take your data, your PDFs, images, and spreadsheets and turn them into real outputs. We're talking charts, PDFs, Word docs, CSV files, Excel sheets, even full slide decks. It went from reading your stuff to building things from your stuff. Third, you don't have to bring all your sources anymore. Before, you had to walk in with everything ready to go. Now, you can start with just a loose idea or a question. You chat, and Notebook LM helps you build the source library itself. It uses Google Search to suggest sources, and you control what gets added. That's a totally different way to start a project. Now, one honest note, because I never want to sell you a dream, this brand new update is rolling out first to Google AI Ultra users and certain Workspace business accounts. The base notebook LM is still free to use and a lot of the core studio and research features work across plans, but these newest agentic features are coming to everyone over time, not all at once. Keep that in the back of your mind. Okay, so here's where it gets fun. The tool is powerful, but on its own, notebook LM gets messy fast. You end up with notebooks scattered everywhere. You can't tell which ones have videos, which ones have podcasts, or what you've already built. So, what I did was wrap the whole thing inside one organized system I call the agent operating system. Inside that dashboard, I can see my whole library in one glance. I can run research in one section. I can chat with any notebook directly. And I've got a studio to generate anything I want from it. Here's a real example of how I use this. I use this to build a full research notebook trained only on the AI Profit Boardroom, what's inside it, what we teach, how the whole thing is set up. Then I ask that notebook questions and it answers like a custom agent that knows the AI Profit Boardroom inside out. That same notebook lets me spin up training material and walkthroughs for the AI Profit Boardroom on the spot. Now, you might wonder why I bother wrapping it in a system at all. Here's the real reason and it comes down to control. If you saw what happened recently when Fable 5 got pulled by Claude, you already know the lesson. Models come and go. Things get taken down. Access changes overnight. So, you never want to be glued to one single model with no backup plan. The smart move is to build a setup where you can plug models in and out and it doesn't break your whole workflow. That's exactly what this layered system does. I connect everything together using MCP, which is just a standard way for these tools to talk to each other. So, if one model disappears, I swap in another and keep moving. No panic. No rebuilding from scratch. That's the kind of stability you want when you're running real work and not just messing around. Build it once and it keeps running no matter what changes. And this is where the research engine really shines. I feed it a pile of sources. It reads through all of them, connects the dots, and then it performs. By performs, I mean it actually generates things. It answers my questions with citations. So, every answer points back to a real source. No guessing. No made-up facts. You click the little number and you see exactly where it came from. The wild part is how recent the answers are. I can ask it what a builder should know about the latest June releases and it pulls genuinely fresh stuff with sources attached. Way fresher than asking a normal chatbot working off older training. You get cited current answers in one shot and that alone changes how fast you can learn something new. And look, if you're watching this thinking the whole Agent OS setup sounds powerful but a little technical, that's exactly what we built the AI Profit Boardroom for. Inside the AI Profit Boardroom, you get the full Agent Operating System as a ready to install zip file. You don't piece it together yourself, you install it and it's done. On top of that, we built a complete 30-day roadmap that walks you through it step-by-step so you always know what to set up first and what to do next. You also get the full Notebook LM walk-through inside the AI Profit Boardroom, weekly coaching calls where you can ask about your own setup, daily tutorials, and a member map full of builders working on the same thing. So, if this video is making your head spin a little, the AI Profit Boardroom is where it all gets simple. All right, let me actually walk you through how I use it day-to-day because the layout is simple once you see it. You've got four parts: the library, the research section, the chat, and the studio. The library is where you start a new notebook and gather everything in one place. The research section is where you go deep. You can switch between fast and deep and deep is the real research agent that digs hard and finds relevant recent information with citations. The chat is where you just talk to your notebook like a custom assistant that's been trained on your sources so it already knows your context. And the studio is where you turn all of that into finished usable stuff. So, inside the studio, I can say something like, "Create an audio overview of this topic." and it builds a full podcast about it. I can ask for a video, an infographic, a mind map, flashcards, a slide deck, a report, and it just makes them. Here's how I put that to work. I ran this to generate a whole content series built to pull the right people into the AI Profit Boardroom. I fed it the topics our AI Profit Boardroom members keep asking about and it drafted the research, the angles, and the supporting assets in one go. That's the exact kind of content that brings the right audience toward the AI Profit Boardroom and the quality is honestly great. I opened up one research report it built for me, trained on my Agent OS, and it was a clean, organized PDF with several pages of useful breakdowns. It covered the ecosystem, the tools, the recent models, the loops, all of it, with charts and diagrams generated right inside the system. I use reports like that to train my team and keep everyone on the same page about the AI Profit Boardroom. One prompt, and I've got a document I can hand to someone new so they understand how the AI Profit Boardroom system actually fits together. Here's the part that genuinely surprised me, though. These agents work in parallel, so I can kick off a podcast, an infographic, and a video all at once, then walk away. While that's cooking, I can jump into another tool and build something else entirely. You set the work in motion, you go live your life, and you return to finished assets. That's a completely different pace than waiting on one thing at a time. Now, let's talk about getting started for free, because I promised you honesty. The base Notebook LM tool is free to use. The MCP connection is free, so you can build a version of this Agent Operating System yourself for free using free tools like Claude or Hermes to handle the connecting. The newest premium update features are rolling out to paid Ultra and Business accounts first, but the core engine you can absolutely start with right now at no charge. You feed it sources, it reads them, and it turns research into things you actually use: podcasts, briefing docs, mind maps, flashcards, all in one place. And the outputs look properly good. The visuals lean on Google's own image generation, so the infographics and slide decks come out clean and sharp, not clunky. Before this, all your research lived in a dozen different places. Notes here, articles there, nothing connected. Now, you drop your sources in, walk away, and come back to whatever you asked for. So, who is this actually for? If you research anything, this is for you. If you create content and you're tired of the blank page, this is for you. If you're trying to learn a new tool fast, and you want cited answers you can trust instead of random guesses, this is for you. Students, creators, builders, team leads, anyone who deals with a lot of information and wants it turned into something useful. A few quick tips before I wrap. First, name your notebooks clearly from day one, or you'll end up in the same mess this system is built to fix. Second, lean on deep research mode when accuracy matters and fast mode when you just need a quick answer. Third, always check the citations. The whole point of this tool is grounded answers, so use that. Click the sources and trust, but verify. And fourth, start small. Build one notebook, generate one report, then expand from there. You don't have to build the whole system on day one. If you want the full process, the SOPs, and 100 plus AI use cases like this one, join the AI Success Lab. Links are in the comments and the description. You'll get all the video notes from there, plus access to our community of 75,000 members who are crushing it with AI. And if you're about to go try this yourself, here's the honest truth. You're going to hit a few walls. Setting up the MCP connection, organizing your notebooks, getting the agent trained right on your sources here. That first setup is where most people get stuck. That's exactly where the AI Profit Boardroom carries you through. You get the full agent operating system zip ready to install, the 30-day roadmap, so you're never guessing what comes next, live coaching calls to fix your setup in real time, and step-by-step tutorials on using Notebook LM inside the system. Come build it with people instead of grinding alone. Join us at aiprofitboardroom.com.