Transcription
Notebook LM just got a huge suite of upgrades that make it more powerful than ever for learning, teaching, and working. Notebook LM generates custom podcasts, reports, mind maps, flashcards, and more based only on the information you choose. The assets you create in Notebook LM, like the quizzes and reports, podcasts and video overviews, are all grounded in the information that you choose, the information that you upload, which significantly increases the quality of the information that's being generated and reduces hallucinations with the AI.
Hi, my name is Callum, also known as Wonderloots, and welcome to today's video on the new notebook LM, the four major upgrades it got this week. Quizzes and flashcards make learning more effective by helping you recall information and test that recall. Custom chat types make your conversations more relevant to your use case. From business analysts to personal tutors, custom reports and AI suggested reports make distilling information significantly easier with reports you may never have thought of before. The four new audio overviews, AI podcasts, help you learn in new ways, demonstrating different perspectives. And the direct integration for teachers and educators makes customizing information for your students easier than ever.
In today's video, I'm going to walk through those four major upgrades while also showing you practical use cases and how you can take the basic features deeper so that you can learn, teach, and work more effectively. If you find this video helpful, please consider liking, subscribing, and hyping this video. I'm working on making YouTube my full-time career, so any support you can give me is very much appreciated. Now, let's dive into the new notebook LM.
So, the first thing you'll notice when you get into notebook LM is that there are options to create new notebooks or check out old ones. So, I'm going to show you one that I created previously called Fineman's favorite problems and problem solving. And the whole idea of the favorite problems framework by Fineman is that he's a Nobel laureate who had a bunch of different questions or problems in mind all the time. And then whenever he came across something new, he would have potentially an aha moment or an epiphany, some kind of significant insight that helped him dive deeper into his favorite questions. Well, you can see that we have three different panels here. We have the sources panel, the chat panel, and the studio panel. So, I'm just going to quickly walk through each of them so you have an idea on how this works overall. But honestly, the best way to understand how this is going to work for you is to just start trying it out. So, I'm going to try and give you a very quick overview today that walks through all the different features, including the new updates. But then I recommend just trying it. Use one of your favorite problems and go explore your curiosity or try and solve a problem that you've had and see how Notebook LM can actually help boost that learning and understanding.
Okay, so now let's take a look at the three parts of Notebook LM. So, we have the sources section on the side here. And you can see here that it includes PDFs, websites, YouTube videos, you can include markdown files. There's a whole bunch of different file types that you can upload here. And the key with Notebook LM is that you're able to upload your own sources. So, what that means is you can ground the answers that you're getting in the chat section that I'll get into in a moment with specific documents that you've curated yourself. This significantly increases the quality of the output of the AI of the LLM because it's grounding its answers in your own work. So you can connect this to Google Drive. You can link websites or YouTube videos. You can upload PDFs, markdown files, audio files. I'll go for a walk. I'll create a voice note on some idea I'm working on or learning about. And then I'll drop that audio file in here and I can have a conversation with it. And once you've done that, you'll have all these sources in the side here. It will generate a summary of all of the sources to give you a bit of an idea on how the AI is understanding the documents that you've included in there. So, that's a good way to get an at a glance view on everything that you've uploaded.
There's also this option to discover sources. And basically what that does is you can search specifically on Google and it will go through and rank the articles and websites that it pulls in for credibility. And then you have the option to import them into the side here. But I'm not going to get too deep into the discover sources component because I do have a dedicated video on how to use the discover sources feature to do deep research with Notebook LM. So it's one of my most popular videos and I recommend checking that out if you want to understand how you can use the discover sources component more effectively.
Next, we get to the chat section here in the middle. And this is where you're able to have a conversation with all of the sources that you've included on the side here. So for example, I could say, "What is Fineman's favorite problems technique?" and the AI will go through and pull in specifically the resources that I've included on the side here, which include some of my own YouTube videos, and it will give me a concise summary on what that is. And along the way, it includes citations, so you can see exactly which source it came from, and you can go and dive deeper into that source once you've used it to locate the specific part of that source that is helpful to answer your question. So, you can go through and you can customize this however you want. You can save this as a note and it will appear on the side because if you do refresh the page, you lose your current chat. So, anything that you want to save, you have to save as a note on the side into the studio, which again, I'll get into the studio in a moment. And honestly, this is just an incredibly powerful way to have a conversation with your notes to chat with your sources and your research and try to use it to help you learn or help you prepare resources that you could use for work or for school. And the key here is that it's all grounded in the research that you've included in here or the research that you've discovered. So you can always know exactly where the information is coming from.
This brings me to the first new feature, the customized chat types. Google mentioned that they're releasing something called the learning guide, which is a way for you to configure the chat. And actually, if we take a look at my other account that does have access to this feature already because it's being rolled out, you can see that there's this configure notebook button in the chat section. So we now have the ability to define the conversational style. So you can go with the default chat analyst guide or create a custom style where you tell the AI how you want it to respond. You can also choose the response length to be default longer or shorter. So you now have a lot more customization ability so that you can use Notebook LM as a virtual research assistant, a personal tutor, or as a shared knowledge base or help center. For example, with the guide, you're able to upload your sources, perhaps all the docs for your company's website or for your software or whatever your services that you're providing, and you can create an open version of this notebook that you can then send people to so they can ask questions about your documentation. So, it kind of just adds a chatbot element to whatever the documents are that you're uploading.
And that actually brings me to another feature. Notebook LM has partnered with OpenStacks to create some public notebooks that include, for example, a textbook on psychology. So you can see here how this is a public notebook that's read only, but people can go in and they can still have a chat with the uploaded sources or use any of the generated sources on the side here. So while it's read-only in the sense that I can't add more sources or generate more elements in the studio, I can have a conversation with the documents here. So that's what they mean by use it as a shared knowledge base or help center because you can make your notebook public and then have anyone come in and have a chat and just learn about it. So that's just a cool way that you can start to use this more collaboratively and use it as an actual knowledge base to share with people. And they are building in the learning guide component which is going to turn the chat into a specific tutor. And the idea with the learning guide is that they're going to encourage participation with probing open-ended questions. So instead of just answering your questions, it's going to actually help you break down the problems so that you can learn much more effectively. So that's a pretty cool feature and I'm excited to explore that more. Once that feature comes out, I'm going to make a more detailed tutorial on how I would use Notebook LM as a tutor along with other features, including how I would use it for work, for knowledge work, and as a teacher as well. So, if you're interested in keeping up with the videos that I'm making on these new features as I explore them more, then please consider subscribing to my channel, which supports me a lot.
All right, now let's take a look at the studio. So, the studio has a few different generative elements. You can generate an AI podcast. You can generate a video overview. You can generate a mind map, which I'll just generate one right now. You can generate reports and newly custom reports that I'll get into in a minute. And we can get into flashcards and quizzes. So, the idea with the studio is that when you click on any of these buttons, it will automatically generate an asset for you based on all of the sources that you've included in here and based on the notes that you've saved or converted into sources. So, for example, I just clicked mind map. Now I can open this mind map here and it generates a full interactive mind map that breaks down all of the sources into these key topics with the ability to expand into subtopics. So this just gives you a really excellent overview of all of the information that you have in here. It helps reduce information overload so that you can use your notes and your research more effectively without getting overwhelmed by all the information. And if you find something that you're interested in, for example, let's say I want to go into the neuroscience of flow states. I can click on any button and it will go ask the chat based on the topic of that mind map. So I'm not going to go too much more into mind maps because I do have a dedicated video on how to use Notebook LM for visual learning, how you can expand with the mind map. So I recommend checking that out if you want to go deeper into that topic. But let's keep exploring the other features of Studio.
So we now have these two new options here, flashcards and quiz. And both of them are customizable. That's what these little pencil icons are. So you can just click the button and generate a flash card which will generate something like this here and it will automatically create a series. You can see there's 55 cards that were generated here. So each one has the option to be flipped over to give you the answer. So this is a really cool way where you could upload, for example, your whole textbook or your notes for your class or everything that you want to try and memorize and you can just automatically generate flash cards that will help you learn significantly more effectively because you don't have to go through and make the cards yourself. You can just start quizzing yourself right away. And that brings me to the quiz component where we can do the exact same thing. Once you've gone through and you've memorized your material based on the flashcards, you can generate quizzes based on it. And the quiz will generate a series of questions based on all the sources that you've included in there. Or if you go through and toggle off some of the sources on the side, it will change the number of sources that are brought into the quiz. So, for example, if you've uploaded your entire textbook, but you only want to get quizzed on chapter 1, you could go through and you could uncheck the chapters that you're not interested in being quizzed on, so that you can generate very targeted quizzes, kind of like what they've done on the side here with this open textbook. So, you can go through and you can click on questions and it will give you the option if you get it wrong to explain the answer. And clicking the explain button brings you over to the chat where it actually pre-prompts for you and says, "Here's what happened. How can I learn this better? Please explain." So the key here is that the flashcards can help you learn, help you memorize the information, but the quizzes can actually test your recall. Did you retain the information? Do you understand it? And if you don't, you can immediately get a breakdown with a deep explanation on why you didn't understand that and how you can get to that level of understanding. So this really helps you fill in any knowledge gaps you might have with your material. Honestly, I think this is going to be an incredibly powerful tool for anyone who's trying to learn anything. It could be preparing for a presentation at work. It could be preparing for a midterm or a final exam. It could be trying to expand your understanding of a paper that you're writing. There's a lot of opportunities here to use the flashcards in the quiz to boost your knowledge and then test it. It's very easy to think that you understand something when you're using an AI tool because the AI can feed you the answers very easily. But introducing a feature like flashcards and quizzes helps you to actually test what you know versus what you think you know. And that explain option I think is also just a great way to help you to go deeper into what you're learning based specifically on the components that you're struggling with. It's kind of like a personalized tutor.
Okay. Now let's take a look at some of the other generations. Specifically, let's take a look at the new report types and the customization with autosuggested reports. If we look on the side here, there's this report button. And I can actually show you what it used to look like because it hasn't rolled out to my other account yet. But effectively, we used to be able to generate these four types of reports: a briefing doc, a study guide, FAQ, and a timeline. But now, if we take a look at the updated version, and I just click on the report, it has a list of four default ones: the create your own, briefing doc, study guide, and blog post. But it also comes up with a suggested format based on the information that's in here. So, it's generating customized reports based specifically on the sources that I've uploaded. For example, a personal growth plan, research proposal. If I go over and take a look at my molecular zettocasten system, which just for context, zettocasten is a nonlinear note-taking system that helps you understand how information goes from an atomic concept to a connected concept. So, I have my own version called molecular zettocasten and I have this notebook to help me explain it to people. So, if I wanted to generate a report on this, I could click on the reports button, and you can see how it's loading suggestions based specifically on the newsletters and resources that I've uploaded here. So, you can see here it's creating a white paper, which to be completely honest, that's exactly what I would need because in my Obsidian and personal knowledge management playlist, I actually have videos on what is zettocasten and how I've converted it into something I call molecular zettocasten. I'm not sure I would have thought about generating a white paper right off the bat. That's an interesting report type that it's suggesting here. And what's cool is within each of these report types, you can see exactly what the prompt was to generate that type of report. So this also gives you some insights into how you can prompt the AI if you want to generate and create your own custom style. So for example, I can create a custom report here. So maybe I'll call it "Create a tutorial outline explaining how people can learn about the power of zettocasten for learning more effectively with Obsidian," my note-taking tool that I use to incorporate all of my research and learning. And what's nice about Obsidian is you can add markdown files. So you can just drag and drop your notes directly into Notebook LM, which I do actually have a dedicated video on if you're interested in learning how I combine Notebook LM with my note-taking to make it much more practical. Great.
Okay, so the white paper just finished. Let's take a look. "A Framework for Cultivating Growth in an Era of Constant Change." So this is in large part because I've dropped in a lot of my newsletters which was all about the concept of self-actualization of being the best version of yourself and how we can maintain our humanity in the age of AI. So that incorporated into my note-taking system where I'm trying to use AI to help me learn but also maintain the benefits of taking notes myself because that's a core part of how we learn with something called the generation effect. So, this is actually a really interesting summary and it gives me insights on my own notes that I've already included in here. So, you could see how this might be helpful for taking in a bunch of research notes, help you outline an essay that you can then go through and write yourself. So, there's just a lot of options here for the different types of reports that you can create.
Now, let's take a look at the final two here. We have the audio overview and the video overview. So, this brings me to feature four, the four types of audio generation. So now in the audio overview section, if we click on the pencil button, we get brought up to this customized audio overview panel and this shows that we have four different types of formats. Now we have the deep dive, a brief critique and a debate. So the deep dive is a conversation between two AI hosts. It's customizable and also you have the option to make it interactive which you can see this little handwaving button down here. So you can actually as you're listening to the podcast ask questions to the hosts and they'll modify the podcast to answer your question. So that's pretty cool. I'll often download an audio overview on something I'm trying to learn or something I'm trying to figure out and then I'll go for a walk and then have that interactive conversation without having to sit at the computer. And this can be done through the notebook mobile app which I have a dedicated video on if you're interested in learning more how to use the mobile app. Then we have the brief, which is a 1 to 2 minute overview on one particular concept or a few small concepts. We have the critique, which will actually do an audit or a review of the sources that you've uploaded. So, for example, you could drop in a presentation for work and get some feedback on if you have any holes in your arguments. You could drop in your essay draft and ask the AI to analyze it and see if there's any way that you could make your arguments more impactful. There's a lot of cool ways that you could use the critique component to target specific sources and try and improve the quality of them. And then we have a debate. And the idea with the debate is that they take a particular issue and then they have each host argue for either side of the debate to try and illuminate the different perspectives on the sources that you've uploaded. And as always, you have the option to change the length and change how the AI host should focus. So you could upload a whole bunch of sources, ask it to focus on one specific concept or one particular chapter. There's literally no end to how you can do this. And then you can just hit.
You know that feeling when you're trying to wrap your head around something really complex, maybe a new concept at work or I don't know, a challenging personal project >> and you just hit a wall. You can read it again and again, but it just doesn't quite click >> right. We're swimming in information these days, aren't we? >> Totally. But >> so that gives you a bit of an idea on how it'll work. You can then download that and listen to it anywhere. But I found it's a great way to get an overview of a ton of information, break it down into more digestible format, like audio, and then I can go through once I've listened to the audio, take some notes, and then focus specifically on the sources that were most interesting to me. I just put out a video that goes into all four of those options a couple days ago because it was released last week. So, if you want to see a practical workflow on how you can tie all four of these audio overviews together in a more effective way, then I recommend checking out that video.
And then finally, we get to the video overview, which is similar to the audio overview. It specifically creates an explainer video that's more of a presentation. So, the form is less of a podcast and more of one person talking to you to explain a particular concept. Here's an example. >> Have you ever wondered how some people seem to just bridge that gap between really intense analytical thinking and those sudden, almost magical flashes of insight? You know, it's a path from hard, deliberate work to the state of effortless mastery. All the overthinking just gets out of the way. >> So, you can see how it goes through. It creates custom slides for you. So, this could be a cool way to understand information with a little bit more of a visual component rather than just the audio. I have a video that goes more into depth on the video overview component and how I would incorporate it into my workflow. So, I recommend checking that out if you want to go deeper.
Again, I know this has been a lot of information, but I just wanted to show you all of the various updates that we have here so that you can figure out which one's going to be the most helpful for you. And then hopefully one of my other videos can help you go deeper into that with more practical use cases. And now we get to the final part of the tutorial, which is the bonus feature for teachers, for educators, for anyone who's trying to help others learn. Not only is Notebook LM partnering with Open Stacks to create peer-reviewed free textbooks for people like I showed you with that psychology one here, but Notebook LM is also giving teachers and students expanded access. So, they're integrating Notebook LM into the Google Classroom in the upcoming weeks. And they're also going to incorporate it, I believe, into Workspace for Education. So you'll be able to actually for example as a teacher generate a notebook like the Fineman technique or Fineman's favorite problems or AI knowledge graphs, quantum computing or intro to physics, whatever the topic is and you'll be able to share this notebook directly with your students so that they can come in check the sources that you've uploaded, which you can curate to make sure they're credible, and then they can have conversations or they can take a look at the predetermined resources that you've generated for them. You could generate quizzes and flashcards and mind maps and podcasts, anything to help the students learn a little bit more effectively.
I'm really impressed with how clean Notebook LM is getting. I've been using it for almost a year now, and I just keep finding more and more ways to use it. So, if you're going through and you have any questions about this, please feel free to ask any questions in the comments below. I'm happy to answer questions and help point you to particular tools or prompts or styles on how you could perhaps use this tool more effectively for your particular use case. I use this in my business all the time. I use this to understand concepts for myself. I use it to help write articles and plan YouTube videos and learn complex topics. I use it to help me code. There's a lot of different ways that you can use the power of customized sources. It's called retrieval augmented generation where you're grounding all of your answers and all of the generated reports. All of this information is being pulled directly from the sources. It's important not to underestimate how significant the quality improves when you're curating the sources that the AI is pulling from. It's much better than just pulling from the internet at large. I hope that this video shows you how you can take these features of Notebook LM, including the upgrades, and tie them together into workflows that would be more effective for studying, learning, teaching, and working. The ability to upload or select your own custom sources and then tie it to artificial intelligence with Notebook LM makes it an incredible tool for any use case because you're grounding the answers and the generated AI in the information that you've selected. If you want to go deeper into any of the topics or features that I've talked about today, I have a whole Notebook LM playlist that goes through each feature in depth along with some practical workflows and how I use it with various other tools like Obsidian. I hope that you find it helpful. Thanks again for watching and I will see you in the next video.