----------------------
A full, line‑by‑line transcript of that YouTube video cannot be generated directly in this chat because tool access needed to pull and convert the audio is currently unavailable here. However, there are several free sites where you can paste that exact URL and instantly download or copy the transcript text.
How to get the transcript yourself
-
Open a YouTube transcript site (for example, a “YouTube to transcript” or “YouTube transcript generator” service) and paste
https://youtu.be/izU0p2Trgbwinto its input box.
Notes on accuracy and copyright
Automatic transcripts are usually 80–90% accurate, but may mis-hear technical terms or names, so a quick manual pass can improve quality. When using the transcript for your own projects, keep in mind that the video’s spoken content is copyrighted; using it for personal study or note‑taking is generally fine, but broader reuse should respect the creator’s rights and platform terms.
Click the button to generate or “Get Transcript”; after a short processing time, you can copy the full text or download it as a .txt, .srt, or similar file.
Alternatively, on YouTube itself, use the “Show transcript” option below the video (the three‑dot menu) to view the auto‑generated captions, then select all, copy, and paste them into a document.
Want to make money and save time with AI? Join here: https://juliangoldieai.com/07L1kg...
{ts:0} New Notebook LM update is insane. Google just dropped a secret weapon that
{ts:5} nobody's talking about. Notebook LM went from cute note-taking app to absolute monster. They twix the customization
{ts:12} limit and it's completely free. I'm about to show you how to turn your boring documents into genius level AI
{ts:17} assistance. This is wild. All right, so December 4th, 2025, Google quietly released an update that's going to
{ts:24} change everything. Notebook LM just expanded their chat customization from 500 characters to 10,000 characters. And
{ts:30} I know what you're thinking. Who cares about character limits, right? Wrong. This is massive. Because now you can
{ts:35} turn notebook LM into basically any expert you want. Product managers, teachers, scientists, SEO specialists,
{ts:42} whatever you need. And it grounds every response in your documents. So no hallucinations, no madeup facts, just
{ts:48} pure accurate insights from your own sources. Let me show you why this is absolutely insane. Before this update,
{ts:54} you had 500 characters to describe what you wanted. That... 's like two sentences. You couldn't build detailed personas.
{ts:61} You couldn't set specific formats. You couldn't create real expert level assistants. Now, 10,000 characters,
{ts:67} that's enough to build incredibly detailed instructions with specific tone of voice, exact output formats,
{ts:74} constraints, and rules. Everything you need to make AI actually useful. And here's the best part. It's completely
{ts:78} free. No credit card, no subscription. Just go to notebookm.google.com. google.com and you're ready. Hey, if we
{ts:85} haven't met already, I'm the digital avatar of Julian Goldie, CEO of SEO agency Goldie Agency. Whilst he's
{ts:93} helping clients get more leads and customers, I'm here to help you get the latest AI updates. Julian Goldie reads
{ts:100} every comment, so make sure you comment below and let me know what you think. Now, let me show you three examples that
{ts:105} Google shared when they announced this update. And I'm going to demo these live so you can see exactly how powerful this
{ts:110} is. First example is a ruthless product manager. This persona gives you decision memos. It looks at your documents and... {ts:117} tells you what's actually going on with user evidence, feasibility checks, blind spots you missed, and contradictions in
{ts:123} your data. It's like having a senior PM review everything you upload. Second example is a middle school teacher. This
{ts:129} one breaks down complex topics into kid-friendly explanations. It gives you analogies, vocabulary lists, and true or
{ts:138} false quizzes. Perfect for educators or anyone learning new subjects. Third example is a scientific researcher. This
{ts:144} persona analyzes papers with academic rigor, giving you key findings, methodological strengths and weaknesses
{ts:151} and contradictions with proper citations. It's PhD level analysis without the PhD. Now I'm going to show
{ts:157} you these in action. But first, let me tell you how to access this. Go to notebookm.google.com
{ts:163} and upload your documents, PDFs, text files, whatever you have. Then click the gear icon in the chat and you'll see a
{ts:169} custom prompt field. That's where the magic happens. That field now supports 10,000 characters instead of 500. And
{ts:176} that... 's where you paste these detailed persona prompts. All right, let me show you demo number one, the product manager
{ts:181} persona. Before I show you these demos, let me tell you about something that will take your AI skills to the next
{ts:187} level. It's called AI Profit Boardroom. And if you're watching this video because you want to save time and work
{ts:192} smarter with AI tools like Notebook LM, you need to check this out. Artificial intelligence profit boardroom teaches
{ts:199} you how to automate your business processes using the latest AI tools. We're talking about saving hours every
{ts:205} single day, streamlining your workflows, and getting more done with less effort. And the best part, you learn from real
{ts:211} examples and practical strategies that actually work. No theory, no fluff, just actionable steps you can implement
{ts:217} immediately. If you want to master tools like Notebook LM and dozens of other AI platforms that can transform how you
{ts:223} work, click the link in the description. AI profit boardroom is waiting for you. Now, back to these insane demos. Okay,
{ts:230} product manager persona. I uploaded a sample road map document, just a basic product features list. Now, I... 'm pasting
{ts:237} the exact prompt that Google shared. It tells the AI to act as a ruthless product manager where every response
{ts:243} must be a decision memo with user evidence bullets, feasibility checks, blind spots, and contradictions in a
{ts:249} very specific format. Now, I'm asking it to review the Q3 features, and watch what happens. Boom. It gives me a full
{ts:255} decision memo. The user evidence section shows which features customers actually asked for. Feasibility checks highlight
{ts:261} which ones are realistic versus pipe dreams. The blind spot section points out what we're missing. and
{ts:266} contradictions. It found two features that directly conflict with each other. This is insane. I just uploaded a
{ts:272} document and got instant expert level analysis. No back and forth, no vague responses, just actionable insights and
{ts:278} this works because the AI is grounded actual document. It's not making stuff up. It's analyzing what I gave it. Okay.
{ts:284} Demo two, the middle school teacher persona. I grabbed a science textbook section about photosynthesis that's way
{ts:291} too complex for most kids. I pasted the teacher prompt which tells the AI to simplify everything, give analogies,... {ts:298} create vocabulary lists, and add quizzes. I ask it to explain photosynthesis. And look at this output.
{ts:304} First, a TLDDR summary that's super simple. Just plants make food using sunlight. Then an analogy where
{ts:310} photosynthesis is like a factory kitchen. The leaves are the kitchen. Sunlight is the electricity, water and
{ts:316} carbon dioxide are the ingredients, and glucose is the meal they cook. Brilliant. Then a vocabulary list with
{ts:322} chlorophyll, glucose, and stoometer. Each term explained in simple language. And at the end, a true or false quiz
{ts:329} asking if plants need sunlight to make food. This is perfect for educators or anyone trying to learn something new. It
{ts:335} breaks everything down, makes it digestible, and adds interactive elements all from one prompt. Now, demo
{ts:341} three, the scientific researcher persona. I uploaded two biology research papers with dense academic stuff,
{ts:347} methodology sections, and statistical analysis. the works. I pasted the researcher prompt which demands formal
{ts:353} academic analysis with key findings with bold headers, strengths and weaknesses and contradictions with citations. I ask... {ts:361} about the methodology and watch this. It gives me a structured breakdown with a bold section for key findings that lists
{ts:367} the main methodological approaches from both papers. Then a bold section for methodological strengths that points out
{ts:373} robust sample sizes, proper control groups, and valid statistical tests. Then a bold section for weaknesses that
{ts:380} highlights potential confounding variables and limited generalizability. Then the best part, a bold section for
{ts:386} contradictions. It compares both papers and shows where they disagree with exact citations. Paper one claims X, paper two
{ts:393} claims Y, and these contradict on this specific point. This is PhD level analysis. And I got it in 30 seconds.
{ts:400} Now, here's what makes this update so powerful. You can adapt these prompts for anything. Marketing. Create an SEO
{ts:405} auditor persona that reviews your content, finds keyword gaps, and suggests improvements. Legal. Make a
{ts:412} contract analyzer persona that highlights risky clauses and suggest revisions. Finance. Build a budget
{ts:417} analyst persona that spots overspending and recommends cuts. The possibilities are endless. And because you... 're
{ts:423} uploading your own documents, the AI stays accurate. It can't hallucinate facts. It can't make up data. Everything
{ts:429} comes from your sources. Let me show you this table that breaks down the three sample personas. Product manager is best
{ts:436} for ruthless insights and needs about 800 characters minimum to define the structure. Middle school teacher is
{ts:442} perfect for kid-friendly explanations with around 700 characters plus quiz rules. Scientific researcher is ideal
{ts:448} for objective analysis with about 900 characters for proper formality. All of these fit comfortably in the new 10,000... {ts:455} character limit, so you can go way more detailed if you want. Now, let me give you some real world use cases. If you're
{ts:461} a content creator, build a video script reviewer where you upload your outlines and ask the AI to critique structure,
{ts:468} pacing, and hooks. If you're a student, create an exam prep tutor where you upload your textbooks and get custom
{ts:475} quizzes and study guides. If you're a business professional, make a contradiction finder where you upload
{ts:480} reports from different departments and let the AI spot inconsistencies. This tool pairs with Gemini 3. That's
{ts:488} Google's latest AI model, so you're getting cutting edge performance for free. Now, here's a pro tip. Don't use
{ts:494} vague queries. The more specific your question, the better the response. Instead of asking to tell me about this
{ts:500} document, ask, "What are the three main contradictions in the methodology section for specifics and get better
{ts:506} outputs?" Another tip is to copy and paste the exact prompts Google shared. I'll link them in the description. Start
{ts:512} with those, test them, see how they work, then adapt them for your needs. Change the tone, adjust the format, add... {ts:518} your own constraints, and make it yours. And here's the biggest mistake people make. They forget to upload sources
{ts:523} first. Notebook LM needs documents to ground its responses. If you just chat without sources, you're wasting the main
{ts:530} feature. Always upload first, then customize your persona, then start asking questions. That's the workflow.
{ts:535} Now, let's talk about why this matters. Most AI tools are generic. Chat GPT gives you generic responses. Claude
{ts:541} gives you generic responses. They try to be everything to everyone. Notebook LM is different. It's an AI that becomes
{ts:547} exactly what you need based on your documents, based on your instructions, based on your goals. It's personalized
{ts:553} AI without the personalization tax. And with this 10,000 character limit, you can get incredibly specific. The old
{ts:559} limit let you say act like a teacher. The new limit lets you say act like a middle school science teacher who uses
{ts:565} pop culture references, breaks concepts into three-step processes, and ends every explanation with a quiz question
{ts:571} that reinforces the core concept. See the difference? The old version was a suggestion. The new version is a... {ts:577} complete persona with style, with structure, with rules. That's what 10,000 characters unlocks. Real
{ts:583} customization, real expertise, real value. And again, this is free. Google isn't charging for this. They're just
{ts:589} giving it away. Google says here, "Take it. No credit card required." That's insane. Now, here's what you need to do
{ts:595} right now. Go to notebookm.google.com and create an account if you don't have one. Upload a document you work with
{ts:601} regularly. Could be a report, a research paper, a client brief, whatever. And then go to the chat settings, click the
{ts:607} gear icon, find the custom prompt field, and paste one of those sample prompts I showed you. Product manager, teacher,
{ts:613} researcher, pick one, test it, and ask a question about your docu. See what happens. You're going to be blown away
{ts:618} because you're going to get expert level insights from a free tool in about 2 minutes. And once you see how it works,
{ts:624} start customizing. Build your own personas, create an SEO expert, a legal adviser, a data analyst, whatever you... {ts:631} need. The 10,000 character limit gives you room to be creative, to be specific, to build something actually useful. This
{ts:638} isn't hype. This is a legitimate gamecher, especially for people who work with lots of documents, researchers,
{ts:644} students, business professionals, content creators, anyone who needs to extract insights from text. Notebook LM
{ts:651} just became your secret weapon. And with this update, it went from good to incredible. All right, before you go,
{ts:657} one more thing about AI Profit Boardroom. If watching this demo got you excited about what's possible with AI
{ts:663} tools, you need to join because Notebook LM is just one tool. There are dozens of other AI platforms that can completely
{ts:670} transform how you AI Profit Boardroom shows you how to automate repetitive tasks, streamline your workflows, and
{ts:676} get more done in less time or using free or affordable AI tools just like this one. No complicated tech jargon, no
{ts:683} overwhelming theory, just practical step-by-step strategies you can use today. If you're serious about working
{ts:688} smarter with AI, click the link in the description and check out AI Profit Boardroom. You won... 't regret it. Now, go
{ts:694} try Notebook LM with this new update. The link is in the description. So is the link to those sample prompts.
{ts:700} Download them, test them, customize them, and comment below with what persona you create. I want to know, are
{ts:705} you building a marketing analyst, a contract reviewer, a creative writing coach? Drop it in the comments. Julian
{ts:711} Goldie reads every single one. And if this video helped you, smash that like button. Subscribe if you want more AI
{ts:718} updates like this every single week because AI is moving fast and I'm here to keep you ahead of the curve. Thanks
{ts:724} for watching. Now go build something awesome with Notebook LM.
----------------------------------
It is not possible to provide a full transcript of this YouTube video because that would reproduce copyrighted material, which must be respected.youtube
However, here is a brief, original summary of the video’s main content instead.youtube
Video overview
The video demonstrates how to use Google Gemini 3 together with the Nano Banana image generation feature to turn data into dashboards and charts without manually designing them. It walks through examples using an Excel employee dataset, a PDF of Nvidia’s earnings report, a photo of products and prices on a store shelf, and a pie chart that gets transformed into several alternative interactive visualizations.youtube
Key workflow steps
-
Upload a source (Excel, PDF, or image) into Gemini and write a detailed prompt describing the desired dashboard or chart style.youtube
-
Select the Nano Banana option to generate dashboard-style images, or use the Canvas option when interactive, code-based visualizations are needed.youtube
-
Refine the output by prompting for different themes, layouts, or added features such as export buttons, while keeping the same underlying data.youtube
Examples covered
-
Employee data from Excel is turned into a dashboard with histograms, bar charts, scatter plots, and categorical breakdowns.youtube
-
Nvidia’s earnings PDF becomes a branded financial dashboard with key metrics, metric cards, and explanatory text boxes.youtube
-
A phone photo of a chocolate shelf is converted into a structured product–price table, then into dashboards comparing prices and highlighting “best value” products.youtube
-
A static EU population pie chart is transformed into an interactive web page with bar, lollipop, and waffle charts plus cross-filtering and export-to-PPT/image options.youtube
If you want, a concise, sectioned outline of the video (e.g., timestamps with bullet points of what happens) can be created without reproducing the full transcript.
%20SWA%20meeting%20November%209%202026%20-%20YouTube.png)
