Original video in Chinese.
Key Takeaway
- New NotebookLM features: added Deep Research (customized in-depth research), PPT export (structured Slides outline building), flashcards (automatic extraction of key concepts for testing), and multimodal support (YouTube/web/audio), evolving from document Q&A into a research/creation assistant.
- A transformation in learning logic: from linear “jigsaw puzzle” learning (suffer first, enjoy later) to precise “dictionary lookup” learning (instant access to knowledge points/full-text insights); from lonely hard study to Socratic dialogue (Audio Overview podcast debates, using hearing/social interaction to digest complex concepts); from input-oriented learning to output-driven learning (flashcards/PPT forcing testing/reorganization).
- Why I recommend it: the best learning tool in the AI era, overturning traditional methods; paired with Gemini, it only requires a $20 Google One subscription; the author has recommended it across five videos, opening the door to a new and highly efficient world.
Google has two products in the AI era that I absolutely love:
One is Gemini.
I’ve made many videos repeatedly recommending it. As early as the 2.0 stage, I said Gemini would lay the foundation for Google’s dominance. Later, it really did not disappoint me.
The other is NotebookLM.
I rarely use the word “miracle tool.” But NotebookLM is absolutely a “learning miracle tool.” Since I recommended it half a year ago, it has now added four categories of new features, making it even more comprehensive and powerful.
In this video, starting from NotebookLM’s new features, I’ll share why it is a learning miracle tool and exactly how it overturns the logic of traditional learning.
Hello everyone, welcome back to my channel. Humbly speaking, I’m one of the few creators in China who can clearly explain both the Why and the How of AI. What I provide is far more valuable than tutorials. Remember to follow me. If you want to connect with me, come join our newtype community. This community has been running for 600 days, and more than 1,900 friends have paid to join.
If you are in China, you can join through Knowledge Planet (Zhishi Xingqiu). If you are overseas, you can join through Substack. My first course, daily newsletter, and exclusive videos can all be seen inside the community.
Back to today’s topic: the miracle tool that overturns traditional learning logic — NotebookLM.
Over the past half year, NotebookLM has gone through a series of frequent and major updates, and has evolved from a simple “document Q&A tool” into a multimodal deep research and creation assistant.
In terms of core research capabilities, Deep Research has been added. The significance of this feature is that it allows users to provide custom guidance. AI conducts deep research according to your instructions, rather than speaking in generalities.
In terms of multimodal output, the recently added PPT feature is especially great! You can export your notes or organized content into Google Slides with one click. And it doesn’t just format the text nicely; it actually tries to build an outline and structure.
As a learning aid, it has added a flashcard feature. The system automatically extracts key definitions and concepts from your documents and generates study flashcards. These cards are especially suitable for quickly checking whether you’ve really learned the material.
As for supported data sources, it now supports YouTube video links, webpage links, and audio files. So you can feed those two- or three-hour podcasts on YouTube into NotebookLM. If you record a class or meeting, you can upload the audio file too.
So if you connect all these features and use them together, NotebookLM is the globally recognized learning miracle tool.
At the input stage, you can throw in PDFs, videos, recordings, and web pages.
At the digestion stage, you can use flashcards and quizzes to assist your learning; use Audio Overview to help you identify the key points.
At the output stage, you can use Deep Research to dig into specific viewpoints; and finally generate presentation materials directly with Slides.
So the question is: it’s useful, sure — but why do I say NotebookLM changes the logic of traditional learning?
I interpret this “new logic” from the following three dimensions.
First, from “linear decoding” to “on-demand retrieval.”
The old logic was that learning had to proceed step by step.
When we got a thick book or a long paper, we had to read from the first page to the last. We had to build context first before we could understand a certain concept in the middle. This is called “suffering first, enjoying later.”
But the new logic now is that AI instantly absorbs the context of the entire book. You can directly jump to the knowledge point you care about most through a question.
What changes is that learning is no longer a “jigsaw puzzle”; it becomes as precise as looking something up in a dictionary. And what you look up is not just a definition, but deep insight based on understanding the full text.
Second, from “lonely hard study” to “Socratic dialogue.”
The old logic was that reading was a solitary journey. When you ran into something obscure and difficult, you could only grind through it yourself, or go look for other materials. So your thinking would easily get stuck or drift.
But today’s new logic is that Audio Overview — the feature that generates podcasts — can turn dry text into a debate between two people. You can even hear them questioning a certain viewpoint.
This dialogic format makes use of our human auditory strengths and social instincts, making it easier for the brain to digest extremely complex concepts in a relaxed atmosphere.
So the overall learning experience and learning outcome naturally become better.
Third, from “input-oriented” to “output forcing input.”
Under the old logic, we spent 80% of our time on “input,” such as reading. But only 20% of our time on “output,” such as taking notes. That is very inefficient, because the Feynman technique tells us that output is the best way to learn.
But today’s new logic is that flashcards and quiz features force you directly into a “testing” state; the PPT generation feature forces you directly into a “presentation” state.
So the transformation is that it lets you skip the inefficient transcription stage and go straight into the higher-level stages of “review” and “reorganization.”
From “linear decoding” to “on-demand retrieval”; from “lonely hard study” to “Socratic dialogue”; from “input-oriented” to “output forcing input.” These three points are the learning logic and learning methods I’ve summed up for the AI era. This is also why I repeatedly and strongly recommend NotebookLM.
Just as I said in the community, over the past year I made five videos recommending Gemini and three videos recommending NotebookLM. And you can get both of them with just a $20 Google One subscription. I’m very certain that for most people, they are definitely the best solution.
OK, that’s it for this episode. If you want to understand AI, want to become a super individual, and want to find like-minded people, come join our newtype community. See you next episode!