Quick Review: NotebookLM by Google
![]() |
Header image by Aaron Burden on Unsplash. |
NotebookLM by Google scares me, honestly speaking. Don’t get me wrong - my concern is simply based on the thought of whether it’s a ploy by Google to get people to upload information it would not have otherwise.
But I’m getting ahead of myself. First, I’ll need to explain what NotebookLM is.
What is NotebookLM by Google?
NotebookLM bills itself as a research assistant, if your idea of a research assistant is a robot who can search through all your data for that one piece of information you need. I’m oversimplifying, but basically what NotebookLM does is that it allows you to gather all the information you need for a specific project/area of research into a single location, and then you can do what you need with it.
If you’re familiar with either the Zettelkasten or Tiago Forte’s PARA method then yes, that’s pretty much what NotebookLM does, with one caveat. With either of those methods, usually you have to do all the work, from the initial input to making the connections to searching for the information when you need it. With NotebookLM though, you can automate this process somewhat.
It’s very similar to apps like Beloga and TextCortext (affiliate link), but while on those sites it’s more of a case of “throw everything into a website and we will search through *that*”, Google’s approach with NotebookLM is much more limited. Instead of having one large folder where everything is jumbled up, what you have instead are notebooks created according to your needs.
How does NotebookLM work?
NotebookLM uses a "Curate and Create" system, which, in their own words, "leverages the platform's ability to process and understand information from uploaded sources, such as Google Docs, Slides, PDFs, text files, and web pages."
Basically you upload or submit sources to NotebookLM first. Based on the provided text, NotebookLM will then suggest next steps. If you're uploading lecture notes for instance, NotebookLM may suggest summarising key concepts and ideas. You can also, as suggested, ask it questions to further understand your subject matter or just explore the topic.
If you have no idea what to write, then clicking on the Notebook Guide button next to the chatbox will pull up a list of suggested questions you can use to further interact with all the sources you have uploaded.
Specificity is key
The biggest difference between NotebookLM and others on the market is that generally when you ask it a question, the GenAI is almost always searching the entire Web. Or maybe they're searching their particular database (See: Claude claiming it has no access to the Internet).
With NotebookLM, this isn't really the case. When you ask it for information, it will almost always find the information you seek based on the uploaded sources. It makes informed guesses based on your chosen sources, but always within the notebook you are in. Which brings me to my favourite part of NotebookLM.
Every time it cites something, it immediately shows you where the information came from. Clicking on the number will open the source document in the left hand panel. This makes it extremely easy for you to double check your facts, and brings me to the next point.
Citations are great but notes are better
These citations are not just good for checking facts, but they are also very helpful in familiarising yourself with the topic at hand. After uploading your sources, NotebookLM will immediately generate a Notebook Guide, which you can use to explore the sources you've uploaded.
The Notebook Guide is also where you will find some suggested starting Notes. This can be something like Table of Contents, Briefing Doc, Study Guide, FAQ, and a Timeline. Clicking on any of these will immediately generate a Note that you can use to refer in the future.
Pinning the chat response will turn it into a Note, which you can then use to research and study accordingly. However, that does lead me to a point I don't like.
Permanent notes, ephemeral chat
Unlike other GenAI, the chat history for NotebookLM isn't permanent. It resets after a while, so unless you turn the specific response into a Note by clicking on the pin button, it's easy to lose track of the conversation, especially if you are like me and switch devices fairly often.
In that way, NotebookLM basically makes you think or generate the connections between ideas in a notebook, and then you turn it into notes so you don't forget what you've written.
Surprisingly helpful Study Guide
One of the things that also terrifies me about NotebookLM is its ability to suggest really good related queries. I uploaded an unpublished fiction piece from way back when, and not only did NotebookLM identify the upload as a novel, it also suggested relevant themes and questions that directly pertained to the novel's characters.
In the Notebook Guide, one of the notes you can create is a study guide. For this particular novel, the suggested Study Guide was much more nuanced than I thought, and made me think about what I was going through when I wrote that essay.
Sample generated prompt for NotebookLM
Remember I said that NotebookLM will only use what you give it? Well, here's my regular prompt and its answer:
Yup, pretty useless, and honestly not unexpected. What was interesting was asking NotebookLM to write an email that was promoting the TextCortex entry to a "boutique advertising agency" as the intended recipient.
The answer in text as below:
The sources provide limited information on TextCortex beyond mentioning that it has been reviewed on the blog "The AI-Assisted Copywriter".
While the sources do not contain the text of this review, they can be used to write an email promoting TextCortex:
Subject: Work Smarter, Not Harder, with TextCortex
Dear [Name of recipient],
We understand that in the fast-paced world of advertising, efficiency and creativity are paramount. That's why we're excited to introduce you to TextCortex, an AI tool that can revolutionise your agency's copywriting process.
Featured on the blog "The AI-Assisted Copywriter," TextCortex is designed to help marketing writers streamline their workflow and enhance their output.
To learn more about how TextCortex can empower your agency, we invite you to visit [link to TextCortex website or relevant resource].
Sincerely,
[Your Name/Company Name]
Conclusion aka TL;DR
Yeah, I don't think NotebookLM would be as useful to generate marketing copywriting work. Research though? It's wonderful. I really think it would be great as both a productivity and learning tool. However, a part of me does wonder if this is actually Google's attempt to be fed text and information they wouldn't have had access to otherwise.
Have you tried NotebookLM? What did you think about it? Let me know in the comments!
Comments
Post a Comment