Transcription
So, this channel is all about sharing with you the news stories and tools that are grabbing my attention, swinging it around, and refusing to let go. And right now, Notebook LM is really front of mind for me. In fact, here's a quick anecdote that illustrates why I think this tool deserves more attention.
The other day, I was messing around with it on my computer and my son comes in and asks, "Hey, Ank, what the Sigma are you 67ing right now?" And I said, "Well, my dear boy, this is notebook LM. What you want to know?" We had been talking about the planet Saturn recently and wanted to learn more. So, I plugged it into Notebook LM and we proceeded to spend the next couple of hours learning about Saturn through a custom-made podcast and accompanying video. And this is just one of the reasons why this tool is so powerful.
So, today I want to show you everything you can do with it from top to bottom and all for free. Starting with a quick primer on what it is and why it's actually useful. For starters, you can find it over at notebooklm.google.com.
Now, Notebook LM does use Google's Gemini AI models under the hood, but it uses it in sort of a different way. It's much more fine-tuned to focus on knowledge that you give it. So, see all of these articles on my left side of my screen? These are all articles that I either searched for and uploaded or copied and pasted directly into Notebook LM. Now, I can actually chat with it and it uses this information to respond to my chat.
Now, if I ask it a question like, "How many moons does Saturn have?" You can see it gives us the answer 274 confirmed moons. But more importantly, it actually gives us the source of that information. So, if I hover over these little numbers here, it will actually tell you which of these documents and where in the documents it found that information. So, if you think it might be hallucinating, you can always double-check the source that it found that information from. So, while hallucinations are still a thing in Notebook LM, they happen way less often because it's grounded in the information that you gave it and because it does its best to cite its sources, making it a lot easier for you to double-check what it gave you, ideally preventing hallucinations.
One particle of untitanium has a nuclear reaction with a flux capacitor. Carry the two, changing its atomic isoton into a radioactive spider. you science.
So for this video, let's go ahead and create a new notebook and find out what we can really, really do with this thing. So inside of notebook LM, I'll click new notebook. As soon as you create a new notebook, you'll get this add sources screen where you can start to add sources. If you accidentally close this, it's real easy to bring it back up. Just click the little add button under your sources column here and it will bring it back up where you can add more sources.
We can upload almost anything. You can upload PDF files, text down files, markdown files, even MP3 audio files. You can connect it to Google Drive and have it read various documents and spreadsheets. You can give it links to websites or YouTube videos and have it actually watch or read the transcript from the YouTube videos. Or you can simply copy and paste text.
If you don't have specific PDFs or articles on the topic you want to learn more about, no big deal. You click on discover sources, tell it what topic you want to learn more about. In this case, let's say I want to learn more about humanoid robotics. I'll click submit, and it's basically going to like do a Google search for us, find a bunch of resources about that topic, and then we can select which of those resources we want to pull in as our sources of information. A complete guide to humanoid robots, history of humanoid robots, we have a PDF here, the mechanics, training a whole body, ethics. All right, so this looks like some good sources. I'll go ahead and import these. and we'll notice it populates our sources here in the sidebar. It even skims our sources to figure out what this notebook's about and accurately said the landscape of humanoid robotics.
Every once in a while, you'll have an article that you can't actually inject to Notebook LM. For example, I'm a subscriber to the information, but it's a paywalled website. So, if I was to take the URL to this article, toss it inside of Notebook LM as a source here. We can see that the source is red. And if I hover over this, it says the source is behind a paywall. So, it won't let me inject that. So, the easy fix, click add, click copied text, and I can simply copy and paste this entire article from the information in, and it will be one of the sources that I can use inside of Notebook LM. So, we'll just paste this here. And now we have the whole article inside with no issues. I'll remove this as a source. And that's a nice little workaround. If you ever see a source that for whatever reason Notebook LM doesn't want to read, just copy and paste it.
I'm going to give it a prompt. What are the bottlenecks for development of humanoid robots? And we get our response here that breaks down what it found from within this information. However, if I was to leave Notebook LM and come back in, this entire response that Notebook LM just gave me will be gone. It just removes it. I would have to ask it again and get a new response. If I really like what it spit out here and I want to make sure that I can read this again later, I will have to click on this button that says save to note. If I click save to note, you can see over here it saved this response as a note. So now I can access it again in the future if I want to.
If we want to take this one step further, we can even take this output and turn this output into a source. If I click on these three dots, I can convert to source and it will now save this response as one of the sources over on my left-hand side.
Notebook LM can handle up to 50 sources. So you can upload or copy and paste or give URLs to up to 50 different sources to use. And each source can be up to 500,000 words each. So that's 25 million words among all of the documents that you load into Notebook LM. One little tip, if you do use all 50 of your sources and you need to add even more content, well, you can actually merge multiple documents into one source if you want. So, I can take this article and this article, create a new document, paste both documents into one source, and now those two articles just became one source, freeing up more space for me to add more sources. As long as I stay below that 500,000 words per source, you should be good.
All right, so that's the basic overview. That's pretty much setting the groundwork so you know exactly what Notebook LM is, what it does, the limitations, and the sort of basic main use case. You give it a ton of sources. You can chat with those sources. It's not really searching the whole web or trying to respond based on like what's trained into a bigger large language model. It's trying to respond based on what it finds in the sources that you gave it. But this again, just scratching the surface. Let's dive into the really cool stuff you can do with Notebook LM.
A couple videos ago, I tested Sora 2 compared to VO 3.1 to see which is the better AI video model. And to be honest, both of them still had a few issues. But apparently I spoke too soon because there's actually a third new AI video model that just dropped today. This one is by Litrix and their new LTX2 video model can apparently generate native 4K videos at 50 frames per second for up to 15 seconds. So let's test it out.
So here's a prompt that I have a feeling both models might have trouble with. Feel free to go ahead and pause the video to read the whole prompt. I'm worried about that mirror reflection part because that's an area that often can go wrong. Here's what Sora 2 made. And yeah, that gets wonky pretty quick. And now here's what LTX2 made. Honestly, this looks super realistic. And it followed my prompt to a T. And this isn't just upscaled footage pretending to be 4K. LTX2 is actually generating native 4K at 50 frames per second. So, if you zoom in, you can really see just how detailed this video is. This is a great tool for creators or studios. But if you want to go deeper into the technical side, the API access can be requested through the LTX2 site and is being rolled out gradually to early partners and teams. So if you want to see the difference for yourself, check it all out at the link in the description. Thanks so much to Lightrix for sponsoring this portion of today's video. Now, let's get back to it.
One of the most basic use cases is for sort of focused information retrieval. One thing I've seen people do, and I've started doing this myself, is to upload various user manuals for products you have, like I have nice camera equipment, nice microphone equipment, I've got a stream deck, I've got a mixing board sitting next to me, I've got all of this cool tech equipment, and well, I can throw all of these manuals into Notebook LM and then ask questions when I run into problems. So, for example, my camera keeps focusing on my background instead of me. What should I do? My camera is a Sony FX3. So, it is most likely reading this document to learn about how to adjust things on my camera. And sure enough, it gave me a whole bunch of things to try.
It's also a great place to consolidate all your documents on a specific project. So, let's say I was making a video about quantum computing and I did some interviews with people based on quantum computing. I can pull in those transcripts and those conversations. I could pull in articles I found about quantum computing. I could pull in my own notes that I wrote up about quantum computing. Any sort of resources that are linked to this project or this video that I'm creating, I can throw into a notebook inside of Notebook LM and query that information however I want, as well as do a bunch of cool repurposing and content creation directly inside of Notebook LM, which is something else we'll get to in a few minutes here.
If you run a business, you can upload all of your training documents, all of your SOPs, all of your support documentation for your product, put it all into one place, and then you can actually share access to that notebook to team members that work in your company. And now team members no longer need to go to you to ask questions. They could go to the notebook that has everything they could possibly need in there. And it will respond based on your SOPs, your product documents, and anything else you put in there. It's a great resource to help onboard new team members into your company because they can chat with pretty much the collective knowledge of your company.
It can create timelines for you. So, using my quantum computing example, create a timeline of all the important events in quantum computing. And it gave us a nice little breakdown of everything that's happened in quantum computing. And here's another cool little tip. This is something I totally stole from my friend Tina Hong, but I thought this was a cool idea, so I'm going to share it here as well. But if you were to copy the response from this whole note and jump into something like Google Gemini, you could turn this timeline into an interactive visual timeline. Create an interactive visual timeline using the dates below. And then I will paste in our entire timeline that we just grabbed from our other response here. Ship that off to Google. And here's what it created. Not my favorite interactive timeline, but you know, it's a start. A few more prompts back and forth and we would have something pretty solid.
I imagine it can understand the contents of images as well. So here's this infographic from the problogger website and you can see it's a very information-dense infographic. If I was to take this entire image here, let's go ahead and save it. We'll create a new notebook here. Now we can take that infographic and unfortunately we can't upload images directly, but we can convert that image into a PDF and have it read it based on the PDF. So I'll just open this image here on my computer. Export as a PDF. We'll toss it into Notebook LM here. And we can see it knows exactly what that image is all about. I can give it a prompt, give me a step-by-step guide on how to go viral. And based on our infographic here, it's given me a step-by-step guide. No need to squint and try to understand what's going on on that graphic.
Not only is it great for productivity and knowledge management, but it's also amazing for students who really want to learn anything. As we've already kind of touched on with our quantum computing and Saturn, but let's go even deeper. If we really, really want to understand a topic inside of Notebook LM, we can create things like flashcards and quizzes to really, really lock in the information.
If I click into reports here, I can also generate study guides, technology primers, concept explainers, or even click create your own and make like an FAQ. So, let's generate flashcards. Let's generate a quiz. Let's generate a study guide and let's generate a mind map. Our flashcards are ready. So I can click on this and we get a new box with some flashcards. What is the fundamental purpose of quantum computing error correction? I don't exactly know how to respond to this to reduce errors. I would suppose to protect quantum states from unwanted environmental interactions, decoherence, and other forms of noise. I don't know much about quantum computing. So yeah, in QEC a quantum error correcting code is a subspace within a larger mathematical space known as a I don't know Qbit. Nope, not even close. But yeah, you can do flashcards.
We've also got a quiz which I will do equally as bad on. What fundamental principle of quantum mechanics allows a qubit to exist in a combination of both zero and one states simultaneously? I think that's superposition. Yeah, I got one right. but it generated 15 questions for our quiz so that I can actually test my knowledge on quantum computing. If you're in school right now, Notebook LM is probably going to be your favorite tool/study guide ever. Here's the study guide that it generated for us. Now, this is more just for reading. It gives us a short quiz with an answer key here. It gives us some essay questions, glossary of terms, and pretty much what we would need to know to kind of understand the basics of quantum computing.
It also generates mind maps for us. So we got our quantum computing and error correction mind map here. And this is really handy to dive deeper and deeper and sort of understand subsets of topics within the quantum computing niche. So foundational concepts and components. I can click this little arrow and expand on this topic. And now we expanded to qubits, quantum phenomena exploited, quantum information processing. And I can dive deeper into this one specifically and then deeper into this one. And you guys know how mind maps work, but you can go deeper and deeper on a topic in the form of a mind map if that's something that helps you understand and grasp the topic even better.
Now, I'm also a big journaler as well. I journal almost daily on just random thoughts, and just get things out of my brain. That's where some of my ideas come from, video ideas, problem-solving ideas, things like that. Most of it comes from journaling for me. And I actually really love putting all of my journal entries into Notebook LM and then having conversations with my journal or generating audio video overviews or mind maps or things like that about my journal entries. I find it really therapeutic to have like an audio overview that's discussing my thoughts or to look at a mind map. It's almost like a mapping of my brain and the way I think, right? My mind map really kind of shows where my thoughts lie. Business and brand strategy, personal habits and self-management, delegation and saying no, events and travel, creative insights and philosophies. And I can just dive deeper and deeper, right? New venture projects, newsletter growth, future tool strategy, YouTube channel improvement, current focuses, and it's like my brain and thoughts in like a visual mapped-out mind map. And it's just kind of crazy to dive deeper and deeper into this. To me, like I find it really valuable. It might help highlight some blind spots that you didn't even realize you had.
But another thing I like doing is because this does actually accept audio files, MP3 files, sometimes I'll just turn on my microphone on my phone and just record audio into the notes app and just brain dump my thoughts in audio form and then you can upload them into here and gain insights from your thoughts and have discussions with your thoughts. To me, that's really valuable. I found it to be really helpful personally.
Now, this Notebook LM account that I've been playing with so far is actually on my business Google account. If I was to switch over to a personal Google account, like one that's connected to my normal Gmail, there's actually another resource for education, which is public notebooks that other people have shared. Now, again, you only see this if you're on a personal Google account, not a business account. But I can click in here, click see all, and see some of the recent featured shared notebooks that expert curators have created. So, for example, parenting advice for the digital age. If I click in here, they curated a whole bunch of sources around this topic and already pre-made a podcast, an FAQ, a mind map, and gave some sample questions that you can ask that they saved the results on. So, not only can you create your own notebook and search it and understand topics on what you're interested in, but you can go and find notebooks that other people have created and sort of continue down the rabbit hole that they started.
But now I want to talk to content creators because there is so much opportunity with Notebook LM if you're a content creator because it can do things like audio and video overviews. These are my two favorite features and I haven't even talked about them yet. An audio overview essentially creates a podcast with two hosts talking to each other about the topic that you gave information on. So if I click into audio overview here we have a few options. We can have a deep dive podcast created, a brief podcast, a critique or a debate about this topic. And you can even steer it a little bit. So I can tell it to really focus in on just the rings of Saturn and how those were formed and get it to make a podcast specifically dialed in on that topic. We can also choose a length of shorter, default, or longer. And you can choose a language that you want the podcast outputted in. So I already created a deep dive here about Saturn. This is one that I created with my son a couple days ago and it sounds just like a podcast. Check this out.
>> Welcome to the deep dive. We cut through the noise to get you the clearest science insights. Today, uh, we're strapping in for a trip to Saturn.
>> Yeah, Saturn. Visually, it's just stunning, isn't it? Those rings are well iconic.
>> Totally. The biggest, brightest rings out there. but uh scientifically.
And one really cool thing that they've recently added to these audio overviews is the ability to interject in the conversation and ask additional questions. So, we can see here that as I press play, there is this little interactive button here. Let's click on that. Now, we can play our podcast and listen to it normally, but there's also a join button. So, check this out.
>> Welcome to the deep dive. We cut through the noise to get you the clearest science.
>> Oh, wait. Someone wants to join. Hey, go for it.
>> What are Saturn's rings actually made out of?
>> Oh, that is a fantastic question.
>> Welcome in. We love it when you chime in.
>> That's actually the perfect starting point for our discussion today about the ring's age paradox.
So, you can see that I can click this join button and as they're having their podcast, I can jump in and ask a clarifying question or maybe a question that they didn't cover. It's like this interactive podcast that I could join in on, which I think is really, really cool. But I can also download the audio overview that it created and listen to it offline or even upload it to Spotify or iTunes and create a full podcast around it.
But even newer than their audio overviews is the new video overview feature. It actually uses Google's Vids technology which sort of makes like PowerPoint slide presentation style videos. If I click on video overviews here, we actually have a handful of options. We've got an explainer video or a brief, a bite-sized overview to help you quickly grasp the core concepts. Once again, you can choose languages and you can steer it in the direction you want it to go. But they've also added some new visual styles like this classic style, more of a whiteboard style, a watercolor style, a retro paint style, heritage paper craft, and even an anime style for it to generate your explanation video in. I've never seen it do one in anime style. So, let's have it generate another video in anime style. But while we're waiting for that, here's the previous video overview that it generated.
You ever have one of those days that just seems to drag on forever? Or maybe it flies by? Well, it turns out for the planet Saturn that might literally be true. Today, we're going to crack open a real cosmic mystery. A puzzle that had scientists scratching their heads for decades.
All right, I'm going to scroll through the video a little bit so you can see that the slides actually do change as they're explaining things in the video. How long is a day on Saturn? And then it gives some explanation of how the amount of time a day is actually changing. And there's visuals to go along with like everything they're saying in this video. And when you're done, you can actually download the whole video as an MP4 file and use it however you use videos. And here's what the anime version of the video looks like. The voiceover is pretty much the same. So, let's just take a look at the visuals here. So, it's a slightly different style, and you have a few options to choose from, and they all look pretty good. I mean, it's not using V.O., so it's not like animated videos, but it's making slides for your entire presentation.
But I really love this ability to create podcasts and video overviews and things like that by just giving it information. In fact, some people have created entire podcasts that they put on Spotify and Apple and those places just using podcasts made from Notebook LM. For example, Andre Carpathy, one of the original founders over at OpenAI, created a podcast called Histories of Mysteries. And if you take a listen, it'll sound pretty familiar.
>> On this planet. You're telling me we can actually learn something about this ancient ancestor?
>> You'd be surprised what we can piece.
Now, I don't know how many people are listening to this podcast, but it's got 48 reviews and a 4.4 stars. And the comments on the side, all two of them are both positive. But you can see right here in the bottom of the description for this podcast, it says, "Audio is AI generated by Google's Notebook LM. Images were made by Ideogram." So, we have a full ongoing podcast entirely generated by Notebook LM here.
Now, from a content creator standpoint, think about all of the repurposing potential. If you have a product that can make podcasts for you, make videos for you, create documents for you, and do all of the other stuff that we just created, there's pretty much infinite possibilities for turning one piece of content that maybe you create originally into multiple other pieces of content. So, let's quickly kind of go through the list. And I know I'm not going to cover everything here, but let's see.
You can feed it a video and turn that video, like a YouTube video, into a written blog post. You can turn that video into an audio podcast. Let's say you're starting from a podcast. You can turn that audio podcast into a written blog post. You can turn that audio podcast into a video. You can turn that podcast into a second podcast where AI is commenting on your original podcast. You can feed it audio notes of you just rambling. Turn that into a blog post. Turn that into a video, turn that into a podcast. You can convert written blog posts into videos, convert written blog posts into audio podcasts, and you can convert everything that I just mentioned above into social media posts for whatever social media platform. So give it a blog post, give it a podcast, give it a video, give it an audio note of you rambling and then tell it to turn it into a LinkedIn post, a Facebook post, an X post, a Threads post, a mind map that you can screenshot and share on Instagram. I mean, just in repurposing from one style of content to another alone, I've just fed you like almost 20 ideas. I mean 19 ideas for just repurposing from one style of content to another. And again, I feel like these are fairly obvious use cases of Notebook LM. There's some use cases out there that are completely wild and outside the box. And if you have ideas like that, I hope you leave them in the comments so that people that are watching this video can skim the comments and find even more ideas and creative ways to use Notebook LM because there are a ton.
Before we wrap up, I want to give a few quick tips and reminders to help you get the most out of Notebook LM. First off, don't forget that if it gives you an output that you like and you want to remember that output, it's not saving it to any sort of history. There's no history of your chat here. You need to save it to a note if you want to be able to refind this message.
When you are adding sources, I highly recommend feeding it different types of media. Give it videos, give it blog posts, give it audio, give it PDFs, copy and paste stuff from other sources, and just feed it as much as you can to get the best results out of it.
Another quick tip is see how there's checkboxes next to every single note here. If we don't want it to reference a specific article, we can uncheck that specific article and it will ignore that article when you ask it questions, or you can unselect all of them and have it just look at one article at a time if you want. So you can really tailor what information the chat is grounding itself with by checking or unchecking sources on the left-hand side.
Four, you can actually add your own notes. So, if you're reading something and there's something you want to remember, but you don't want to save this entire text, you just want to save something simple. Over on the right, you can click add note, and it'll just give you a little text editor where you can add your own note here. We'll call it my note. And now we've got a simple, easy-to-access note here that maybe you keep your own little side notes on. And these notes over here aren't going to be something they factor into the output of the chat unless you click the three dots and convert it to a source. and then it becomes a source that's part of what you're chatting with.
So, if you've ever wanted like a personal knowledge base where you can have separate folders for separate topics that are interesting to you and then just chat with that one topic, that one subject, that one area of interest, this is for you. It's also great for uploading your own thoughts and ideas and then creating podcasts or having discussions with like your past self, the past thoughts that you had. To me, that's really fun and interesting and helps me sort of unlock a little bit about how my brain works. So, so many creative use cases and I really feel like this is an underrated, underutilized, undertalked about AI tool. Everybody's talking about ChatGPT and Sora and Claude Code and all of these cool tools. But in my opinion, Notebook LM is one of the most useful AI tools, platforms, whatever you want to call it, available today. It is an educational resource and guide for anything you want to learn. It's a way to better understand yourself. It's a way to repurpose and create new content out of content you've already created and so much more. I am absolutely in love with Notebook LM and just wanted to share all of the ways I use it because it is so dang powerful.
And if you like videos like this and you want to stay looped in with the latest AI news, but also get details, tutorials, and tips on how to best use and integrate all of this AI that's coming out into your own life or business, make sure you like this video and subscribe to this channel. I will make sure that more videos like this show up in your YouTube feed, and you will always be looped in and tapped in and able to integrate all of the newest technology into your own workflows. That's my promise. If you follow this channel, I'll do my best to fulfill on that promise. Thank you again for tuning into this video. I really, really appreciate you. Thanks for hanging out. Thanks for nerding out. Hopefully, I'll see you in the next one. Bye-bye.
Thank you so much for nerding out with me today. If you like videos like this, make sure to give it a thumbs up and subscribe to this channel. I'll make sure more videos like this show up in your YouTube feed. And if you haven't already, check out futuretools.io where I share all the coolest AI tools and all the latest AI news. And there's an awesome free newsletter. Thanks again. Really appreciate you.