Google's NotebookLM is way more than notetaking, writing, or organizational tool. It's an AI collaborator, grounded in your data with your unique view of the world. In this tutorial, I'll give you a tour of the main functionalities, demonstrate concrete use cases such as making sense of meeting notes or writing an article, and cover the current limitations.
TECHNICAL NOTES
[
08:37]: NotebookLM isn’t actually “trained” on your data. The model is pre-trained, and the software just shuttles your inputs into its context window temporarily so it can answer factually based on that information. Once you end your session, the information you entered is wiped from the model's memory so your data is secure.
[
13:30]: NotebookLM can answer specific questions, but holistic questions (such as "Give me an outline of the entire book, in order") are harder, because the model can’t take in the entire document at once; it can only see the most relevant passages for a given query.
CHAPTERS
00:00 -
01:20 Intro
01:21 -
09:04 Getting started
0
9:06 -
14:52 Use cases for understanding
14:53 -
21:12 Use cases for writing
21:13 -
23:12 Using Readwise with NotebookLM
23:13 -
24:23 Limitations
24:24 -
25:23 Conclusion
25:24 -
25:47 Blooper
