What Are the NotebookLM Updates and How Do They Work?
NotebookLM Updates are a set of new features that turn NotebookLM from a simple note-taking tool into a full research assistant. In this episode, Julian Goldie shows how the tool can now find sources, build files, and even run code on its own.
The core change is a new AI brain. NotebookLM now runs on Gemini 3.5 plus a system called anti-gravity. That means it thinks deeper, reads your files better, and does smarter web research. It also shows you its thinking step by step, so you can follow along.
Google says the new model beats its old tests by a wide margin. For reading big documents, it has close to a 70% win rate. For deep web research, it scores over 78%. In plain terms, NotebookLM got a lot smarter.
Main Features: Cloud Computer, File Creation, and Auto-Sources
The biggest upgrade is not just the brain. Every notebook now comes with its own secure cloud computer. This little computer can write and run code all by itself. It can dig through data, make charts, build reports, and do math.
NotebookLM also comes with over 100 built-in skills ready to use. You do not need to be a programmer. You just type what you want, and it does the work.
- Create charts as image files
- Build documents like PDFs, Word docs, markdown, and plain text
- Generate pictures using a tool called nano banana
- Make spreadsheets as CSV or JSON
- Build full Excel files
- Create complete PowerPoint decks
Another key feature is auto-source finding. You can start with just an idea or a question. NotebookLM searches the web, finds relevant articles and guides, and suggests them to you. You approve the sources, and it builds your notebook from scratch.
How to Get Started with the New NotebookLM Features
Getting started is easy. You do not need to upload any files if you do not want to. Julian shows a simple process that starts with a prompt.
- Open NotebookLM and create a new notebook.
- Type a prompt that describes what you want, like research a topic or build a report.
- If you want NotebookLM to find sources, let it search the web and suggest sources to you.
- Approve the sources you like and reject the ones you do not.
- Let NotebookLM process the information and generate the output.
- Review the results. Make any edits if needed.
- Download the file if you created one, like a PDF or Excel sheet.
That is all. You can go from an idea to a finished report in minutes.
How to Use NotebookLM: Real Examples from the Episode
Julian gives three practical examples in the episode. They show how to use NotebookLM for research, documents, and plans.
First, for research, you type: 'Research the top five reasons people join AI automation communities and the top five reasons they leave, then build me a clean report.' NotebookLM finds the info, spots patterns, and hands you a finished report. No uploads needed.
Second, for a presentation, you type: 'Build me a clean professional slide deck that shows the huge value of the AI profit boardroom.' NotebookLM builds the whole deck with charts and slides. You just download it.
Third, for a content plan, you type: 'Build me a 30-day content plan in a spreadsheet that promotes the AI profit boardroom.' You get a full Excel sheet ready to go.
- Research: get a report without manual searching
- Documents: create a one-page PDF guide in seconds
- Spreadsheets: generate a 30-day content plan in Excel
Results and Comparisons Shown in the Episode
Julian shares specific test results from Google. The new model beats old tests by a large margin. For reading big documents, the win rate is close to 70%. For deep web research, it is over 78%.
He also compares the old NotebookLM to the new one. The old version only helped you understand documents. It could answer questions about files you uploaded. It could not go beyond that.
The new version does much more. It finds sources, builds files, runs code, and completes tasks. Julian says it feels like a research robot. He calls it an 'AI analyst in your pocket.'
He emphasizes that you can go from a random idea to charts, a report, and a finished slide deck without doing much work. The time saved is huge, possibly hundreds of hours per month.
Tips, Limits, and Mistakes to Avoid with NotebookLM
Julian gives three important tips. First, always check the work before you share it. The AI is good, but you are the boss. Give it a quick look first.
Second, be patient with the rollout. The shiny new features are rolling out to AI Ultra users and some Workspace folks first. You might not see every feature right away. Hang tight if it is not there yet.
Third, good sources still win. Feed it junk, and you get junk back. Feed it solid info, and you get gold. This means you should choose your sources carefully when you let NotebookLM find them.
Avoid the mistake of assuming the AI is perfect. It can make mistakes. Always review the output before using it in a professional setting.
Who Should Use NotebookLM and Why It Matters
NotebookLM is for busy business owners, marketers, and anyone who needs to do research or create documents quickly. It is especially useful for people who do not code or design.
Julian explains that the tool combines three things: the old NotebookLM that reads sources, deep research that hunts the web, and an AI agent that gets tasks done. This makes it a full analyst for your business.
It matters because it saves time. You can turn ideas into finished work in minutes. That used to take hours. For a business owner, this time can be used for other important tasks.
If Google keeps improving, NotebookLM could become one of the most powerful AI tools on the planet. For now, it is already a major step forward for anyone who works with information.