NotebookLM's rise was the moment I realized how deep Google's AI rabbit hole actually goes. The majority of people still think Google's AI innovations end at Gemini, but in reality, Google has been quietly experimenting with powerful tools that deserve just as much attention.
All of these experiments happen in Google Labs, which is where the company ships products that are too experimental for the main brand (even when they're shockingly good). Some of these experiments end up getting widely released as standalone products (like NotebookLM did), while many just... vanish. I've been staying on top of these experiments, and three lesser-known ones now feel more powerful than NotebookLM itself.
Illuminate
NotebookLM’s Audio Overviews, but better
While NotebookLM's excellent RAG capabilities are what makes it so unique, it truly took off when Google added the Audio Overviews feature to it. If you aren't familiar with it, the Audio Overviews feature lets you convert all the sources you upload to your notebook (PDFs, Google Docs, Word documents, images, audio files, webpages, YouTube URLs, etc.) into conversational podcasts.
Rather than an AI robotic voice reading your text, these podcasts are in-depth and conversational, making them an excellent way to consume long or complex material without staring at a screen. Illuminate is currently a Labs experiment that focuses on NotebookLM's Audio Overviews feature.
In fact, it's practically the only ability it offers, and it's centered around letting users convert research papers into five-minute audio discussions (though you can theoretically upload any web content as a link). Illuminate also offers a selection of pre-built audio discussions based on research papers and famous books in its dedicated Explore tab.
However, what's impressive is that while Illuminate's Audio Overviews are essentially built using the same technology as NotebookLM, they're several steps ahead. Instead of just giving you a static MP3 file like NotebookLM does, Illuminate generates a real-time, interactive transcript alongside the real audio.
As the AI hosts speak, the text highlights. The transcript is also interactive, meaning you can click on any text in the transcript and the audio will instantly jump to that exact second. With NotebookLM, on the other hand, you need to manually scrub through the audio playback! Illuminate also makes customizing the Audio Overviews extremely simple. You can define the audio dialogue, which includes the style and tone of the audio discussion.
You can also hit the handraise icon when listening to an Audio Overview generated within Illuminate and type any question you might have. You'll then get a typed response grounded in the source you've provided. While NotebookLM offers a similar option in its interactive mode where you can join the podcast and directly ask a question, it's not always ideal. For instance, I'm always listening to Audio Overviews in public places, and the last thing I want to do is start talking to my phone. So, Illuminate’s typed Q&A feels much more natural.
Illuminate is also trained specifically on academic papers, meaning its scope is narrower. This means its ability to parse dense, technical charts and methodologies into conversational English often feels sharper. Finally, Illuminate also saves all the Audio Overviews you generate in a separate Library tab for 30 days after creation.
With NotebookLM, you need to access each notebook individually to find a podcast you had generated. With all your audio overviews accessible in one place, it’s much easier to revisit past discussions without digging through old notebooks.
Learn About
Better for studying than NotebookLM
From day one, my primary use case for NotebookLM has been studying. Learn About is a learning tool Google Labs is currently experimenting with, and it's strictly designed to be a learning companion. With NotebookLM, you need access to source materials before you can even begin using it.
If you don't have study material and are diving into a topic from scratch, it's no good. Learn About removes that friction entirely, and is designed to, well, help you learn about a topic from the ground up. When you head to the tool's website, all you need to do is let it know what you'd like to learn. You can also upload a PDF or enter a URL if you'd like.
Learn About is more like an AI chatbot, like ChatGPT or Gemini, but instead of just giving you a wall of text, the tool creates a structured, multimedia layout with images, videos, and learning cards.
For instance, I typed "arduino" and it created an Interactive List learning card that covered the key components of the Arduino platform.
The very first response included multiple other learning cards, like Why it Matters (which explained why Arduino matters) and Stop & Think cards, which included a question to force me to pause and verify I actually absorbed the information. Instead of including the answer to the question in the card, it hid it within a Tap to Reveal button.
Given that we now live in a world where we default to AI to answer the simplest of questions, we're increasingly losing our ability to just freely think. So, this simple card forces you to pause and actively engage with the material, rather than just passively scrolling and consuming information.
On the left of the main chat section, you'll also find the response broken down into clickable sub-sections, which are all neatly divided into headers. Clicking any of the sub-sections automatically leads to an expanded detailed response.
At the end of every response, three follow-up questions are also displayed, along with related content you can explore. So, while I do still prefer NotebookLM for learning when I have the sources on hand, Learn About is a completely frictionless way to dive into a topic from scratch.
These tools take NotebookLM to the next level
I've been using both these Google Labs experiments since the company announced them, and I absolutely love them. There's also another Labs experiment called Learn Your Way, which I think is incredible, but unfortunately, I'm still waiting to get off the waitlist. So, I can't work with my own documents within the tool just yet.