By Mike G Robinson
I wouldn’t call myself a religious man, but when I sat down to write about Google’s NotebookLM, I was drawn to the Book of Gemini which said:
“In the beginning God created the heavens and the earth. Now the earth was formless and empty, darkness was over the surface of the deep, and the Spirit of God was hovering over the waters.”
And then God said, “Let there be light,” and there was light.
This story has nothing to do with religion—but you may experience something akin to a ‘religious experience’ when you realise the power behind this tool.
What is NotebookLM?
In my own words, NotebookLM is a large language model (LLM), powered by Google’s Gemini. It’s not necessarily tied to a specific Gemini variant like 1.5 Pro or 2.5 Flash—think of it as Gemini in its free-roaming, general-purpose form.
Like many of Google’s innovations, NotebookLM started life in Google Labs. If you haven’t explored Labs, I suggest you do—it’s where Google experiments before deciding which products to unleash upon the masses.
And unlike many tools in the AI space, NotebookLM is actually free. Not “7-day trial” free, or “tiny free tier with token rations” free—free free:
- Up to 50 notebooks per user
- 50 sources per notebook (PDFs, websites, Google Docs & Slides, YouTube URLs and more)
- One-click Text Summary (click Add Note)
- One-click Audio Overview (AI-generated podcast-style summary)
- Auto-generated Study Guides, Timelines, Briefing Docs, and FAQs
- Discover function to surface 10 relevant web sources from a prompt
NotebookLM’s Superpowers
The real magic lies in how it handles those 50 sources. In the AI game, context is everything.
What is context? It’s the prioritized content the LLM keeps in its short-term memory—like the ongoing thread in your favorite chat window. With NotebookLM, the context window is gigantic: it actively pulls and processes information from every single one of your uploaded sources.
Here’s how powerful that is:
- A complete legal Act in PDF = 1 source
- A 10-hour YouTube video? Scraped, transcribed, and summarised – 1 source
- A university course in PowerPoint or Slide Decks? – 1 source
- A full-length book? 1 source
- 46 more sources left
And then there’s Discovery, a web agent that returns 10 handpicked sources from the internet based on your prompt. It’s like having an AI researcher on staff, minus the HR headaches.
Some Real Life Use Cases
Here’s how I’ve used NotebookLM—and how you can too.
1. NotebookLM – The Lawyer
I like to think of the 50-source limit in bands of 10. Let’s say you’re working through a legal issue like I was recently. Nothing serious, a chance to dispute a government decision rooted in legislation….AI is the perfect partner in law.
Here’s what I did. I went about filling those 50 sources to the brim:
- 10 sources on legislation – Acts, precedents, whitepapers
- 10 sources on legal interpretation – Case law, expert commentary, legal blogs
- 10 sources on how to write legal opinions – Legal writing guides, textbooks, training slides
- 10 “Discovered” sources – Let NotebookLM find niche insights or obscure rulings
- 10 personal/working sources – Meeting transcripts, handwritten notes, strategy docs
Once NotebookLM could take no more, I jumped in the cockpit and started driving. These LLM’s are very capable these days but are still a little unpredictable. Give them an inch and they’ll take a mile. You have to wrangle any LLM like a shepherd works his flock and all biblical references aside, this is a very apt example. Let me demonstrate.
Steer an LLM in a direction, off it goes, diligently spitting out beautifully arranged words that at least sounds like a good legal argument. But like sheep, leave them on that track and they’ll just keep going so we throw a dog in the mix and the sheep steer back on the path. Throw another morsel at them, a certain section of the Act or a case precedent and off they race, bleating away, blinkers on until… woof, back this way.
You have to give the LLM a whiff of the story you want to tell, enough backbone for it to see the connecting dots, the gold that you’ve devised over many sleepless nights. If you get this right, AI will shine bright like the guiding star.
AI was made for this stuff, it’s a machine. AI eats legal opinions for breakfast, lawyers for lunch and judges for that after dinner treat.
2. NotebookLM – The Shepherd of Startups
Building a new business or product? NotebookLM turns chaos into clarity.
- 10 sources on competitors – Landing pages, pitch decks, case studies
- 10 market research sources – PDFs, investor reports, thought leader blogs
- 10 customer insight sources – Survey results, reviews, transcripts, Reddit threads
- 10 startup methodology guides – Lean Canvas, Y Combinator handbooks, founder blogs
- 10 internal sources – Meeting notes, idea dumps, investor Q&A
Then tell NotebookLM to:
- Generate a SWOT
- Summarize market gaps
- Draft a pitch deck outline
- Turn user interviews into personas or product features
3. NotebookLM – The Personal University
Forget passive consumption—build a living, breathing curriculum.
- 10 core learning sources – Books, lecture slides, expert blog posts
- 10 university-level sources – OpenCourseWare PDFs, YouTube lectures, transcripts
- 10 adjacent topics – Ethics, regulatory impacts, application use cases
- 10 case studies – Real-world implementations, startup docs, product pages
- 10 personal notes – Your own highlights, mind maps, summaries
Ask NotebookLM to:
- Quiz you
- Summarize each section
- Build timelines and concept maps
- Help draft a thesis or opinion piece
The NotebookLM Commandment
In today’s mad world, this Google product is a revelation. It’s use of sources as context with a 1m token window, creates raw thinking power seamlessly integrated with the now familiar chat interface.
We are pilgrims on the path to productivity. NotebookLM is our divine intervention delivering productivity of truly biblical proportions.
Let this become your gospel.
Leave a Reply