Transcription
Hello everyone. And today, together, we will look at one of the best products that has significantly advanced in its capabilities in 2025. And this product is NotebookLM. And today I will show you how, with its help, you can study any topic and make it so that you can explain this topic to anyone. What is NotebookLM? Yes, Google itself defines it as your research assistant for studying any topic. Let's go. Upon entering NotebookLM, your notebooks are loaded, and a notebook is, as it were, the key entity with which work is done within NotebookLM. Let's start from scratch, create our first notebook. What does Google offer us? It offers us to connect any sources that we can see. You can upload files, you can specify YouTube links, you can upload your files, and generally connect Google Drive, insert text. But let's try to use the most interesting feature. This is, of course, searching directly on the internet. And since 2026 is ahead of us, let's see what trends await us in 2026. You can search the internet on your drive. You can select deep research. Let's choose deep research so that it selects the maximum number of sources for us, based on which we will further work with them. Find all reports from major consulting companies with forecasts for 2026 on technological trends. Let's launch. What is Google LM? Yes, we have a name. Let's give it a name. Trends 2026. On the left are our sources. Now we are searching. And the middle is a classic chat for working with sources. You can ask questions, and on the right is the studio, i.e., a set of already defined functionalities that can be used to work with our sources. So, step number one. NotebookLM has selected 41 sources for us. Let's see what it offers us. We see Gartner here. We see Deloitte, IDC, Forester here. Well, in principle, quite a lot of good sources. Let's import them all into our project. What is happening now? That is, Google analyzes them all and adds them so that we can use both text and images from these sources. Here the first ones are starting to appear. How is it presented here? Yes, there is a brief guide to the source, we can familiarize ourselves with it. And there is the source itself with tables, i.e., the maximum recognized information for our material. And in total, we have almost 41 sources. Here we have some sources that did not load. Possibly paid links, possibly something else. So, well, there is an option in Google itself to delete all sources that are unavailable to us. Let's delete them, we don't need them. Most people who use NotebookLM simply start asking questions in the chat and interacting with it. But, I think, those who are familiar with data and any topic of analysis know that the most important thing is how high-quality our sources and materials are. Therefore, we will go through the following steps, which are recommended to be done before starting any research. We have searched for sources, and now we need to validate them. Are these sources generally relevant, do they have any significance? Perhaps there is some bias, and they do not look at our problem systematically and comprehensively. How do we do this? We ask NotebookLM in the chat the following question. Prepare a table listing all sources. So that they are sorted by file name, as they are included in this report. And next to each source, indicate how relevant it is. That is, we need the date of this source. We need to understand if this is a primary source, or a news item, or some opinion, or a reference to another source. And, please, indicate how much we should include this source in our report. or not. And let's launch this request. Let's see what we got. All our sources are listed in the same order as they are presented here. Let's see which ones are suggested not to include. Reference, you see, press release of the report, optional, brief version, and new blueprint. Okay, let's just delete it altogether. We don't need it anymore. Optional Accenture Vision 25. Let's disable the optional tariff. Everything else, in principle, is worth including. Excellent. Let's leave it like that. So, we now have 31 sources. You see, if we click here, the number of sources we will ask questions about changes. Let's move on to the next step. The next step is significance, right? How significant are these sources after all? So, let's ask this question again. Among the remaining sources, check if there are any that refer to existing ones. Also, indicate in these sources if there are any that are referenced and are most cited among the sources we have already connected. Look at what we got. We have a number of sources in the reports that refer to other reports. Here we can make a decision ourselves, to remove these reports if we want to focus only on key foundational reports from consulting companies, or to keep them to reinforce some opinion. This is already at your discretion, this is a task of how we detail the approach to what we will use in our reports. I will not clean them up now, let's leave them, let them be further. Our next step. We still want to check if we have covered everything, all directions, and let's ask a simple question: study these key primary sources, which are referenced, for bias and possibly some narrow focus. Our task is to understand how broadly we are looking at the problem of technological trends in 2026, if something is missing, so that we can launch an additional search. If you think there is more than enough information, then just write that we can proceed to our report. In my opinion, this dialogue with our database of sources allows us to determine for ourselves what further relationship we will have with them. So, look at what we got, right? That we got that it can be divided into four groups, each with its own specifics, right? That the Big Four are more focused on C-suite, right, on executives, and present trends more as challenges. At the same time, if we look at technological analysts, then here we see truly new terms, which Gartner loves to come up with some new term every year. Specialized expert companies, narrow specialization, engineering academic communities. Well, in general, in general, again, for our current task and quick immersion, this is more than enough. We will not change anything further based on these sources. Let's leave these 31 as they are. Although no, let's also see what else we can add. Look how it works. So, for example, Gartner has a video presentation from the last summit on YouTube. And we just take this video and add it as a source. And if you look, then here we get the entire transcript of the presentation. So, the classic approach to working with NotebookLM. First, we have settings, i.e., we can immediately configure our notebook so that it helps us in research in the way we need. Choose a communication style. So, let's write our option. You are a technology consultant for the boards of directors of non-tech companies. Why are we doing this? And let it answer longer. So that it explains in simple language. We are offered different options on how approaches to companies and management will change. and in 2026. Okay, let's see what it will write. We can delete the chat history. Chats are saved between sessions. This appeared relatively recently in NotebookLM. Before, all this disappeared. So, we get a lot of text that we can already work with and try to understand what the key value of NotebookLM is. It makes direct references to the text and effectively eliminates any hallucinations related to answering questions. We can always look at this text, check how correctly it is displayed here and how it is taken into account. Therefore, we can always confirm one fact or another. In this regard, NotebookLM allows us to quickly go through the fact-checking stage of the facts that it actually offers us. Well, let's move on. And the key thing that we need is the next step, immersion in materials. We have already familiarized ourselves with the chat. And the first feature that appeared in NotebookLM is, of course, reports. What are reports? That is, you can set your own format, specify the structure in which you want to receive this report. You can launch a brief overview, a manual, you can immediately prepare a blog post. And there are recommended formats that are generated after NotebookLM analyzes the connected sources. I very often use reports as follows. I just run all of them that are available, because this allows me to quickly review materials from different angles. The number of these reports that we can generate is unlimited. Therefore, in principle, there is nothing terrible in terms of using some limits. Therefore, let's launch these reports. Please note, I specifically did not remove the reports that our NotebookLM told us were optional. And if they are there, there is a risk that NotebookLM will include them in our report requests. What happened in this case? Yes, we had 31 sources that we left, and the reports started to be created based on thirty-eight sources. That is, it included all that were optional. Therefore, when you perform validation, you must delete the sources that you think you will not need. In our case, it is not critical, so I did not do it. I also showed how NotebookLM works if you don't do it. Look, we already have our first reports. Analytical Brief. We won't dive into it now. The very fact that it can be done. A textbook on technological trends. That is, here are already question-and-answer sessions, right? That is, how we can work with it. The most interesting thing, of course, for immersion is the various formats in which we can present information. And therefore, what we are doing now is moving on to preparing materials, working with the studio. There are five types, besides reports, of ways we can present information. We can create infographics, we can make presentations, we can create tables with data from these materials, make audio podcasts and video presentations. Each format has its own specific cool features. To start, let's see what they are. We can launch a default infographic. Let's do that. Let's see what it will do for us. One of the latest features that appeared in NotebookLM is, of course, data tables. That is, we can tell it: "Prepare a table listing all technological trends, their current state in 2025, and a forecast of their state in 2026. If there are any financial indicators for revenue or market volume for this technology or trend, include it in a separate column for subsequent analysis." and say generate. Audio summary. We have the option to launch an audio summary in four variants. That is, detailed analysis, brief summary, review, debate. Debate is one of the most interesting features. And therefore, here we will give it the following prompt. Prepare a discussion between participants where they argue about which trends were chosen and which are truly relevant or not. They will try to separate what is hype and what is truly worth attention. Generate. Our table with links has been generated, and we can immediately export it to a table. Google. We get all our information in a convenient interface that we are already used to. Let's return to our notebook. Look, we have our first infographic, which is made in a basic format. And, I would say, it looks trivially simple. Therefore, my first advice when creating an infographic in NotebookLM is to immediately specify your style in which you want to receive the information. For creating infographics, I recommend using your style. That is, the style I just created with the design system together with Google Gemini, right? That is, we set the color palette, we set the accents, the effect, graphic elements, and so on. That is, everything that it offered me. Therefore, I now say: "Create an infographic for me. Prepare an infographic with the top ten trends that are most frequently mentioned in these reports, in the design system that is presented to you." And let's launch the generation. Let's see what else might be useful here. First of all, of course, is a presentation. For presentations, I also recommend adding your style immediately. And here we can choose two options: either a detailed presentation, or slides for the speaker. Slides for the speaker by default, and let's launch the generation. The next thing we can do is a video summary. When creating a video summary, it makes either an educational video or a brief summary. We can choose, look, there is already our own option, right? That is, we can set a prompt. in the way we want to see it. Let's try. Let's insert our infographic. What should the presenter focus on? Analyze the most non-obvious trends that are least mentioned in the reports, but at the same time can have a significant transformative impact on our world. Well, all our reports are ready. Let's see what we got. So, first is the infographic. 10 AI trends for 2026 based on our 38 sources. This is the style that was set. What do we see here? Agentic AI, well, I would still call it agentic. The Great Infrastructure Reassembly for the Era and Management as a Competitive Advantage. The Imperative of Authenticity. In general, not bad. The only thing to consider is that it sometimes makes mistakes with infographics. And we see it here too. There are quite a few typos. If the text is small, you need to keep this in mind. Therefore, when creating infographics, I, for example, use the classic Nanno Banana Pro, upload this information. That is, in principle, in NotebookLM, you can prepare a description for an infographic, extract all the necessary data, and then in Nanobana, launch and create an infographic in 4K quality, which will not have any typos. Next, yes, we received our audio file. Welcome to The Debate. Today we are diving into the future, analyzing forecasts of technological trends for >> it made debates for us for 15 minutes. We won't listen to it, but how I use audio. You can upload any book that interests you, one source, and ask it to create an audio format. and we will get a 15-minute shortened version of the book with key insights. It does it quite cool. Similarly, you can, in principle, upload any video, any information, any news and listen to it in the car while driving to work. I do exactly that. Let's see what our presentation turned out to be. The style was set to be closer to Adobe, so it generated it in that format. I can't say I like the style itself, but nevertheless, it's an example of how you can create presentations in the format you really want, not the basic one that is present in the studio. and video, >> to see what invisible. >> I'll say right away that >> it didn't pick up the format I set for it. >> It did it completely in the basic >> thresholds of the world of smart machines, capable of truly incredible. >> You know, something didn't like the options it gave us, so let's make another infographic and another presentation, but it will be said like this. We will choose vertical, maximum detail high, and say the following prompt. Prepare an infographic on the top five trends of 2026 in a New Year's style. Let the infographic definitely have Santa Claus. It's the year of the horse, a fiery red horse, snowmen. The mood is maximally festive. Tie it all together into a single picture. Generate. And let's also create a presentation. Prepare a presentation on the ten key trends of 2026. Each slide should be represented by characters from Soviet films, comedies, as suitable as possible for this trend. The photograph should fill the slide as much as possible against its background. Our infographic in New Year's style and trends in the style of Soviet films should be ready. Let's see what happened. Let's look at the infographic. Here, excellent, as requested, Santa Claus on a fiery horse, snowmen. Signs of the New Year. Yes, I even see here that there are a number of typos. Unfortunately, NotebookLM still makes mistakes with maximum detail, so I love using presentations here the most. So, our presentation. 10 key trends of 2026. Well, even some of the actors are recognizable. Invisible Revolution, Agentic Silicon Workforce. Physical, well, almost a pepelats, it's not very similar. The Imperative of Authenticity. Look, you see, it even tried to take quotes from films. Renaissance of blue-collar workers. Nadya, Fedya, Nadya, needs to. Well, excellent. Industry and the Crisis of Trust Verification Infrastructure Under Pressure We Will Sell Our Garage and Motherland, well, look, it tried to extract as many Soviet films as possible. Ruin is not in the closets, but in the heads. And here we see the maximum manifestation of generative artificial intelligence. And in principle, only your imagination will limit you in what form you will do it. Who Anna Smirnova is, I unfortunately don't know, but it was very funny that she appeared at the end of our presentation. We see the capabilities of NotebookLM in terms of loading materials, all the sources that we need, the ability to dialogue with it in the form of a chat, creating detailed reports on all the information that we have. Forming infographics, forming presentations, creating audio, video summaries. There is one more interesting thing, it's a mind map, as well as flashcards and tests. What is this for? I think NotebookLM is one of those tools that can be used in education, because with its help, you can quickly prepare tests to check the knowledge of those to whom you want to convey it. A mind map, in fact, breaks down all the technological trends for us. For each trend, we can look in more detail. I see that it took Gartner as a basis. Already recognizable detail. At the same time, for each trend, we can automatically generate a question in our chat, and it will continue the dialogue. That is, this is another way of structured learning of material, because you see how the key accents are already presented in your documents, and accordingly, you can go into more detail based on them. Well, here are our answers, right? What are sustainable skills, inversion of the skill hierarchy, resistance to automation, strategic thinking as steering. That is, we go further into each topic that interests us most. Again, we are offered the next questions, but this is already a familiar interface in many chatbots. Flashcards trends, right? That is, there are flashcards, what is the main reason why banks in 2026 are merging departments for combating fraud and cybersecurity? What new opportunity for banks does blockchain technology create? You can look at the answers. That is, these are more like flashcards for memorization. that you can create flashcards. Again, here they are created automatically, based on which we can further study the material more deeply. According to the 2026 trend reports, why is it becoming necessary for brands to demonstrate perfect imperfection? You can take tests, it suggests where the correct answers are, where we guessed wrong, you can explain. And thus, again, we go into the chat. That is, we get an additional tool to work through a particular topic more deeply. That is, we see what request was formulated, right? That we are taking a test, my answer was this, the answer is incorrect. Help me understand why my answer is not suitable. Automatically generated request to the chat, and then it provides a detailed description of why this was done. We can save our answers in notes, and then use them again within our sources, thus working with our notebook. In addition, we can share our notebook with other participants and invite them to work on this notebook. Moreover, give them the opportunity to see only the sources separately, or only, well, or the chat itself, or the artifacts that we have already generated. That is, this is, in the future, a full-fledged environment for collaborative study of large, complex topics, preparation of materials for subsequent use there in work. wherever you need it in your personal life. And I often use NotebookLM for, as I said, reading books, studying any materials that I need. Especially when it comes to quickly immersing myself in YouTube videos, when I want to upload several videos on one topic and quickly understand what it was about and what key insights can be highlighted from it now. After this, it's only your imagination. That is, the number and size of sources are practically unlimited. That is, I think 300 sources is already an overkill, because their size and quality will still require more serious work from you to filter out what is truly necessary, and based on them, achieve the desired results. But I think that's all. I'll say again that NotebookLM is one of the best products that I use almost on a daily basis. See you in the next videos in the new 2026. It will definitely be super.