Audio, video, and visual learning outputs
Generating and customizing audio overviews, cinematic videos, slides, mind maps, infographics, whiteboard-style formats, and interactive explainers from notebook sources.
40%
Best tweets about NotebookLM
50 Best Tweets About NotebookLM (2026)
Explore the best tweets about NotebookLM, including source-grounded research, audio overviews, study workflows, prompts, and practical use cases.
Specific NotebookLM workflows, features, research methods, source handling, and product updates supported by firsthand examples.
- Creators
- 38
- Updated
Original Xholic analysis
What 50 top NotebookLM posts reveal
The dataset primarily discusses NotebookLM workflows built around user-provided materials: posts describe uploading documents, asking cross-source questions, and producing study or media outputs from those materials. Announcements and individual workflow accounts are common; claims about outcomes in study, legal, tax, and medical contexts should be treated as user-reported examples rather than independently validated results.
- Dominant tone
- Positive
- Median score
- 27.0
- Leading format
- Announcement
- Recent posts
- 22%
86% of posts
All-time engagement
32% of posts
Published in 90 days
Conversation map
The themes creators return to
Source-grounded research and synthesis
Uploading curated PDFs, notes, papers, transcripts, and documents to ask cited questions, compare sources, surface relationships, and produce reliable research outputs.
38%
Product updates, Gemini integration, and roadmap
NotebookLM feature announcements and reported development around Gemini Notebooks, deep research, planning mode, Canvas, connectors, autosync, notifications, and personal intelligence.
30%
Study, revision, and exam preparation
Using course materials, books, lectures, past papers, quizzes, gap analysis, and targeted prompts to build understanding, predict assessments, and focus revision.
24%
Personal knowledge bases and note systems
Combining NotebookLM with personal notes, reading archives, Obsidian, Readwise, agents, and CLIs to query and reuse private, curated knowledge.
16%
Research-to-build automation
Turning notebook research into technical prompts, websites, code, private podcasts, or agent-driven projects, then feeding outputs back for review and improvement.
14%
Professional document analysis
Applying NotebookLM to legal case files, depositions, medical records, tax guidance, and other high-stakes document collections for contradictions, timelines, explanations, and preparation.
10%
Content production and business intelligence
Analyzing customer calls and research to identify objections and themes, maintain living source libraries, and repurpose material into marketing, podcasts, websites, and multi-channel content.
6%
Tone and stance
Sentiment Positive leads
Author posture Supportive leads
Performance benchmark
- Median likes
- 57
- Median reposts
- 7
- Median replies
- 7
- Median views
- 5.3K
Posts with media make up 74% of this collection. Their median all-time score is 36.0, compared with 18.3 for text-only posts.
Format mix
- Announcement 32% · score 69.3
- Case Study 24% · score 16.3
- Tutorial 24% · score 27.6
- Story 12% · score 148.8
Consensus and debate
Where creators agree, and where they do not
Shared view
User-provided sources anchor many described workflows
Posts describe loading readings, slides, papers, transcripts, books, or case materials into NotebookLM before asking targeted cross-document questions. A legal user explicitly describes finding cases in Westlaw or Lexis first, then loading those materials into NotebookLM for targeted questions.
Shared view
Study posts emphasize connections, gaps, and question generation
Study-oriented posts describe prompts for connecting current and prior material, probing weak understanding, generating quizzes, and examining likely or challenging questions. These are reported workflows, not evidence that they improve learning outcomes or predict assessments reliably.
Shared view
Notebook sources can be repurposed into multiple outputs
Official posts announce Cinematic Video Overviews and customization of video overviews in the mobile app. Other posts describe using notebooks for outputs such as slides, mind maps, quizzes, flashcards, audio, and websites; those broader output claims come from individual users or third-party posters.
Shared view
Some users describe research-to-production loops
Posts describe taking NotebookLM-based research into website, code-building, podcast, or content workflows. One post specifically outlines uploading a completed build back to NotebookLM for review, security feedback, and design feedback.
Open debate
Source grounding still requires review
A legal user says results depend on supplying their own sources and pairs NotebookLM with legal databases. Separate posts mention classmates' hallucination concerns and a user making corrections to generated slides, indicating that source-grounded workflows do not eliminate the need to verify outputs.
Open debate
Released features and reported tests should be distinguished
Google-linked posts announce Cinematic Video Overviews, mobile video customization, and Gemini-app notebooks. In contrast, third-party posts characterize Canvas, Connectors, Planning Mode, and Personal Intelligence as hidden, in testing, or anticipated rather than confirmed generally available features.
Open debate
High-stakes examples are anecdotal
Posts describe legal contradiction checks, a tax explainer, and medical-record review. These accounts do not independently validate the stated results. The legal workflow also explicitly relies on the user first providing relevant case sources.
What performs
Patterns behind standout posts
Announcements led the measured format mix
Announcements represented 32% of posts and had a median all-time score of 69.25, compared with 27.63 for tutorials and 16.28 for case studies. The Cinematic Video Overviews announcement was an outlier with an all-time score of 3007.95.
Stories had the highest format median
Stories had the highest median all-time score among the listed formats, at 148.84. The context-stacking study post was the largest supplied outlier, with an all-time score of 6266.94.
Posts with media had a higher median score
Media appeared in 37 posts, or 74% of the dataset. The media-post median all-time score was 36.012, versus 18.34 for text-only posts. This is a descriptive comparison and does not establish that media caused stronger performance.
Professional-document-analysis posts had a comparatively high theme median
Professional document analysis had a median all-time score of 45.869, above the 15.92 median for source-grounded research. The theme includes posts about legal materials, a tax explainer, and medical-record review; their claimed outcomes remain individual reports.
Statistical standouts
- View standout post 1 Score 6266.9 · 232.37× median
- View standout post 2 Score 3007.9 · 111.53× median
- View standout post 3 Score 1762.8 · 65.36× median
- View standout post 4 Score 771.0 · 28.59× median
- View standout post 5 Score 392.2 · 14.54× median
Creator landscape
Who shapes this conversation
The five most represented creators account for 20% of the selected posts.
-
1. Louis Gleeson
@aigleeson
2 posts
-
2. Jainam Parmar
@aiwithjainam
2 posts
-
3. Amit | Frogomo | AI 🐸
@frog_omo
2 posts
-
4. Gemini Notebook
@Gemini_Notebook
2 posts
-
5. Glenn Gabe
@glenngabe
2 posts
-
6. Shalini Goyal
@goyalshaliniuk
2 posts
The listed Gemini Notebook account had the highest top-voice median
Gemini Notebook posted two evidence tweets and had a listed median all-time score of 1561.87 among top voices. Those tweets announced Cinematic Video Overviews and mobile video-overview customization.
Ihtesham Ali's two posts included two supplied outliers
Ihtesham Ali had a listed median all-time score of 4014.84 across two tweets. The study-workflow post and legal-document post are both included in the supplied outlier list; the performance and workflow outcomes described in those posts are not independently verified by the dataset.
Participation was dispersed across creators
The dataset contains 50 tweets from 38 creators, and the top-five placement share is 20%. Repeat voices in the supplied top-voice list cover study workflows, legal use cases, feature tracking, and research-to-build tutorials.
Since the previous snapshot
What changed since Aug 12, 2026
- 80% of the selected posts remained.
- The creator count changed by -3.
- The leading sentiment remained stable.
How this analysis was made
Themes, sentiment, stance, and post format are classified per tweet. All counts, shares, medians, creator concentration, freshness, and performance comparisons are then calculated directly from the published snapshot.
Xholic's all-time score compares engagement while accounting for reach, post age, and creator consistency. It is used for relative comparisons within this collection.
This report analyzes the exact 50-post snapshot shown below. AI identifies editorial categories and drafts explanations; all statistics are calculated from the snapshot, and every narrative claim is checked against cited posts before publication.
Best NotebookLM tweets
Top NotebookLM tweets from 38 creators
Ranked 01–50
- 01
@ihteshamali ·
A MIT student figured out how to compress an entire semester of lecture content into one 90-minute study session. He calls it "context stacking," and it's the most unfair thing I've seen done with NotebookLM. I asked him to walk me through it. He did. I haven't studied the same way since. Here's exactly what he does. Two days before each lecture, he uploads everything into NotebookLM. The assigned readings, the previous week's slides, 3 or 4 related papers he finds himself, and any problem sets that are still open. Most students wait for the lecture to explain the material. He walks in having already built a mental model of it. That's step one. But it's not the move that makes it unfair. The first prompt he runs across all of it: "What are the 5 core concepts this week's content is built on, and how do they connect to what I studied last week?" Not summarize. Not define. Connect. NotebookLM pulls threads across everything he uploaded simultaneously. It surfaces relationships between ideas that would take a normal student weeks of review to notice. He gets that map before the lecture even starts. Then he runs the prompt that does most of the work. "What would I need to genuinely understand about this material to be able to teach it to someone with zero background in this subject?" That question is doing something most students never force themselves to do. It exposes exactly where his understanding is solid and exactly where it's hollow. The gaps show up immediately, and he spends the rest of the 90 minutes filling only those gaps. Not reviewing what he already knows. Only fixing what he doesn't. The final prompt is the one that separates context stacking from every other study method I've heard of. "What question could a professor ask about this material that would expose a student who understood the surface but missed the underlying logic?" He's not studying for the exam he expects. He's studying for the exam designed to catch people who only think they understood it. By the time he sits in the lecture hall, the professor is not teaching him anything new. The professor is confirming what he already mapped, filling in a few details, and occasionally surprising him with something he didn't anticipate. That surprise is the only thing he writes down. Most students leave a lecture hoping the material will eventually click. He walks in with it already clicked, and uses the lecture to find out what he missed. That's not a study hack. That's a completely different relationship with learning.
- 02
@Gemini_Notebook ·
Introducing Cinematic Video Overviews, the next evolution of the NotebookLM Studio. Unlike standard templates, these are powered by a novel combination of our most advanced models to create bespoke, immersive videos from your sources. Rolling out now for Ultra users in English!
Watch video
- 03
@ihteshamali ·
I spotted a lawyer recently at a big law firm doing something on his laptop between depositions. Took me a second to realize what I was looking at. He had NotebookLM open with 6 years of case files uploaded. Here's what he was actually doing. I watched him paste in a fresh deposition transcript and run one prompt: "Cross-reference this testimony against all prior statements in this case and flag every contradiction with exact page citations." What used to take a paralegal team 2 days came back in 90 seconds. But that wasn't the part that broke my brain. He uploads every opposing counsel's past filings into a separate notebook. Then asks: "What argumentation patterns does this attorney rely on and where have those arguments failed in court before?" He walks into every hearing already knowing how the other side thinks. I asked him how long he'd been doing this. "Since I realized billing hours for document review was making me dumber." His win rate in summary judgment motions is up. His prep time is down 60%. His partners think he just got sharper with experience. He told me the experience part is true. He just has a 6-year memory that never forgets a page number.
- 04
@aiwithmayank ·
🚨BREAKING: Google just merged Gemini and NotebookLM into one unified workspace and it changes everything about how you use AI for deep work. It's called Notebooks in Gemini and it's the personal knowledge base that power users have been begging for. You create a notebook for a project, drop in your files, PDFs, and documents, give Gemini custom instructions, and every chat you have stays organized in one place. No more hunting through old conversations. No more re-uploading the same files every session. The wildest part is the sync. Anything you add in Gemini automatically appears in NotebookLM. Anything you add in NotebookLM automatically appears in Gemini. One source of truth. Two powerful apps. Zero friction switching between them. So you can start a research notebook in Gemini, ask it questions all week, then flip to NotebookLM to generate a Cinematic Video Overview from the same material. Next morning, open Gemini and ask it to write a full report on exactly what you just watched. That workflow used to take three apps and a lot of copy-pasting. Now it's one notebook. Rolling out this week to Google AI Ultra, Pro, and Plus subscribers on web. Mobile and free users coming soon. What do you think?
Watch video
- 05
@GoogleResearch ·
Ever wonder if a computer can simulate brain activity? Google Research just curated a new #NotebookLM notebook asking just this. It features sources on: ✅ Predicting neural activity with AI ✅ Nanoscale brain mapping ✅ Synapse-level reconstructions Start exploring the frontier of neuroscience today: https://t.co/oQJ5mMHv6C
- 06
@thisdudelikesAI ·
Harvard students have a NotebookLM workflow that replaces 6 hours of revision. They don’t re-read notes. They upload lectures, slides, and readings. NotebookLM builds custom quizzes, predicts likely questions, and explains only weak areas. It compresses weeks into one session. Here’s how they do it:
- 07
@aiwithjainam ·
A PhD student told me he uses NotebookLM to reverse-engineer how professors think. He uploads years of course material, past papers, lecture slides, reading lists, and assignment briefs into NotebookLM. Then he asks 5 prompts. By exam season, he understands the professor better than students who attended every lecture. I thought that sounded ridiculous. Then I saw the workflow. Here’s the exact system:
- 08
@sundarpichai ·
Some helpful updates from across Google this week, lots more to come! 🧵 @NotebookLM is introducing Cinematic Video Overviews for Ultra users in English. Distill complex information into amazing visual deep dives - take a look 👇
Watch video
- 09
@testingcatalog ·
NotebookLM will be overpowered 👀 > Google is working on a new Canvas experience and Connectors support for NotebookLM. > Canvas mode lets users generate any visual representation of the data from notebook sources. > A new Connectors option has been added to the settings menu recently (currently hidden). > Looks like at Google I/O, we will see a new NotebookLM with Canvas, Connectors, and Personal intelligence support.
- 10
@sundarpichai ·
Nothing quite like the blank pages of a fresh notebook! Notebooks are now rolling out in the @Geminiapp. Organize conversations, notes, and other sources related to a single project (and easily go back and forth between @NotebookLM to go deeper). Rolling out today, starting with Google AI Ultra, Pro + Plus subscribers on the web.
Watch video
- 11
@Gemini_Notebook ·
Also, you can now customize your video overviews in the NotebookLM mobile app! We're all watching videos on our phones anyway... might as well make them educational, grounded in your sources, (and kawaii!) 💕
Watch video
- 12
@IamEmily2050 ·
I used NotebookLM to study Google new breakthrough with TurboQuant and used Video overview to study the subject, best learning tool in the world at the moment. TurboQuant: Redefining AI Efficiency with Extreme Compression Google Research has introduced TurboQuant, a suite of advanced algorithms designed to dramatically compress the data used by large language models and search engines. By utilizing specialized techniques like PolarQuant and Quantized Johnson-Lindenstrauss, the system transforms complex information into a compact "shorthand" that requires significantly less memory. This innovation specifically addresses the key-value cache bottleneck, allowing AI models to process massive amounts of text faster without losing accuracy. Testing demonstrates that these methods can shrink data size by six times while actually increasing operational speed on high-end hardware. Ultimately, these theoretical advancements enable more efficient semantic search and high-performance AI applications at a global scale.
Watch video
- 13
@EXM7777 ·
you're sitting on your most valuable data source and not using it... if you run an agency, every call with a lead or customer should be recorded here's what i do: > record every single interaction > dump it into NotebookLM > let it analyze patterns in objections, questions, language they use then i use that to: - sharpen my pitch using Gemini's insights - refine my offer around real pain points i missed - create content google meet will automatically take notes for you but chances are you'll never read that instead, build a video database of customer intelligence that compounds over time + it's always fun to look at yourself 10y later...
- 14
- 15
@DallasAptGP ·
I spent an hour with Google NotebookLM and it made a better Opportunity Zone explainer than anything I have ever seen online. Just good source docs and a clear prompt. It walks through the full OZ framework: defer, reduce, eliminate, and no depreciation recapture (which most people miss). The first version was generic and rushed the math. A few prompt revisions later it nailed the pacing, got the 1245 recapture trap right, and explains it all pretty well.
Watch video
- 16
@goyalshaliniuk ·
NotebookLM just became much more agentic. 1. It can now run Deep Research agents that discover, evaluate, and organize sources, then turn them into structured reports. 2. Chat gets stronger reasoning, 1M-token context, longer memory, and goal/persona-based behavior. 3. It now creates reusable outputs like reports, tables, mind maps, flashcards, quizzes, slides, infographics, audio, and video overviews. 4. Google AI Ultra unlocks the biggest limits: larger notebooks, more collaborators, priority Gemini models, deeper research, and watermark-free visual artifacts. NotebookLM is moving from “AI note assistant” to a full research OS.
Watch video
- 17
- 18
@nikunj ·
Random models that are so good and yet still somewhat underrated.. > Gemini 3 Flash for long context and structured outputs. For the cost, this thing is so freaking good. Even Claude was like yeah one shot it bro, this multi agent orchestration is unnecessary. I have agents running this continuously and I barely spend any money. > GPT 5.4 Pro for problems that require really hard thinking. The latency sucks, only available on ChatGPT but the output is beautiful. No model comes close to how smart it is and how human like the reasoning output is. API wen!? > NotebookLM for slides. The ability for it to compress huge number of sources into beautiful slides is unparalleled. I use it regularly to explain research papers. It uses nano banana under the hood so you get a frontier model mostly for free. > Open source - If you didn’t know, open router has a bunch of free models and inference. It’s fun to try them just to we what shape each model has and where it’s good. Minimax 2.5 and GLM (both not free) are quite incredible for the price. I’m excited for DeepSeek 4 and Kimi’s next models. If there’s a model and use case that has surprised you, drop it here!
- 19
@ai_for_success ·
A son built a “vibe-coded” AI workflow to help his mother navigate stage 4 cancer and catch critical medical errors. Sadly, she passed away, but what he built with AI changed how she was cared for in her final days. - Pulls daily medical data from the hospital’s Epic system to track real time treatment changes - Managed 1,600+ pages of records using NotebookLM - Found 2 misdiagnoses and 3 instances of the wrong cancer type in official scan reports - Detected a life threatening pulmonary embolism just from observed symptoms - Uses Claude to translate complex medical terms into plain English for the family - Prepares sharp questions for doctor visits to push back when needed - Identified diet patterns that were triggering internal bleeding - Acts like a second opinion layer in a system that almost discharged her without a plan
- 20
@ordonez_adan ·
Been using Notebook LM more than CoCounsel. I think it's better at finding the "needle in the haystack." If you load your own sources, it's more accurate than most RAG-based legal tools I've tested. You ask it a specific question and it pulls the right excerpt with pretty solid accuracy. The tradeoff is that you have to give it the sources yourself. It's not connected to a legal database. My current workflow is: find your cases in Westlaw or Lexis, load them into Notebook LM, then ask your targeted questions there. I've found that combination works better than relying on either tool alone.
- 21
@Freyabuilds ·
I accidentally discovered how to read a complete book in 30 minutes. A Harvard student showed me the workflow. Here's exactly what he does. He doesn't open a book and start reading from page one. He said that's the slowest, most inefficient way to absorb a book ever invented. You read linearly, your brain has no context for what matters, and by chapter four you've already forgotten chapter one. He does something different. He uploads the entire book into NotebookLM first. Then he runs one prompt before touching a single page. "What is the single central argument this book is making? What does the author believe that most people don't? And what are the 5 most important ideas I need to understand before everything else makes sense?" That prompt does something most people don't realize. It gives your brain a skeleton before the flesh goes on. You are no longer reading to discover what the book is about. You already know. Now every page you read is confirming, extending, or challenging something you already hold in your head. That is a completely different cognitive experience. The second prompt is the one that saves the most time. "Which chapters or sections contain the core ideas? Which ones are examples, case studies, or repetition of things already said?" Most nonfiction books are 60 to 70 percent padding. Not because the authors are dishonest. Because publishers want 250 pages, not 80. The actual argument usually lives in four or five chapters. The rest is illustration. NotebookLM tells him exactly which four chapters to read. He reads those. He skips the rest. He is not missing anything. He is cutting everything that was never the point. The third prompt is what separates this from summarizing. After reading the core chapters, he goes back and asks: "What questions does this book not answer? What would a hostile critic say is wrong with the central argument? Where does the evidence feel weakest?" This is the move that most people never make. They read. They absorb. They move on. They have opinions given to them by the author and they carry those opinions around as if they built them themselves. He stress-tests the book before he closes it. He knows where it holds and where it doesn't. That is not reading. That is thinking with the book as a sparring partner. The final prompt is the one I use every time now. "If I had to explain this book's core idea to a smart 14-year-old in three sentences, what would I say? And what is the single most actionable thing the author wants the reader to do differently after finishing?" That prompt forces compression. And compression forces understanding. You cannot compress what you do not actually understand. I read four books last month this way. I retained more from each one than I have from any book I read cover to cover in the last two years. The average person reads a 300-page book in six hours and forgets most of it within a week. He reads the same book in 30 minutes and can still argue its central thesis six months later. The book didn't change. The interface did. Most people are reading books the way they were designed to be sold. He reads them the way they were designed to be understood.
- 22
@Shruti_0810 ·
This Russian guy hacked learning Saved 1,460 hours NotebookLM + Gemini + Obsidian → Dump any content → AI removes repeats → Keeps only what you don’t know 20 videos = same 20% info He deletes the other 80% What took 1 month now takes 15 minutes
Watch video
- 23
@joshwoodward ·
New in NotebookLM: The Science Of Ben Franklin, a first-of-its-kind featured notebook This featured notebook was made in collaboration with The Royal Society, the oldest scientific academy in continuous existence. This one features some of Franklin's original papers, letters, and contemporary sources. Featured notebooks allow you to learn from curated collections of high-quality sources. Enjoy!
- 24
- 25
@Angaisb_ ·
I feel like I have a massive advantage with AI For example when studying, I watch my classmates spend hours doing the same thing I do, simply because every time they get something wrong they start searching through hundreds of pages of notes, find it, and sometimes don’t understand it so they waste more time asking and waiting for an answer When I get something wrong I just go straight to NotebookLM with my notes and in a minute or less I have the explanation And even when I give them the tool they say they don’t trust AI for studying because of hallucinations, because the only AI they know is the free ChatGPT from a year ago
- 26
@aigleeson ·
A Stanford researcher ran a 15-minute experiment that lifted students' exam scores by a third of a letter grade. Not by making them study more. By making them ask better questions before they started. Her name is Patricia Chen. She's a psychologist who kept meeting the same student after every exam. Smart. Worked hard. Completely blindsided by the grade. The effort was real. The strategy wasn't. Here's what she found. Most students treat studying like a sprint. Open the book, read until it's time to stop, close it. No frame going in. No target to aim at. Just motion. Chen tested what happened when students spent 15 minutes before studying answering three questions. What do I already know about this. What do I actually need to learn. Which resources will get me there fastest. Scores went up by a third of a letter grade across the board. Same students. Same material. Same amount of time studying. Just 15 minutes of strategic thinking at the start. The brain does not absorb information randomly. It absorbs it by hooking it onto something already there. Walk into a book with no question and the information has nowhere to attach. Walk in with a specific question and every page is either answering it or it isn't, which is itself useful information. Chen called it a strategic mindset. The orientation of asking yourself what you need before you start pouring effort in. Most people skip this entirely because starting feels like progress. It isn't. I read her research and immediately rebuilt my reading workflow inside NotebookLM around the same principle. Here's what I do now before I open chapter one of anything. I upload the table of contents, the introduction, and a couple of reviews of the book. Then I run this before I read a single page: "Based on what this book claims to cover, what are the 5 most important questions a serious reader should walk in trying to answer? What would I need to believe going in for this book to actually shift how I think?" That's it. That's the 15 minutes. Then I read. When I'm done I come back and run: "Did the book actually answer those questions? Where did it deliver and where did it dodge?" Chen's research says the framing before you start is doing more cognitive work than most people give their entire reading session. NotebookLM generates the frame in 30 seconds. Most people are still opening chapter one cold. The questions you ask before you read determine almost everything you take away from it.
- 27
- 28
@drxwilhelm ·
🔊 My paper "The Deleted Degrees of Freedom," compressed to audio. An AI-generated walkthrough of the full paper. No equations. No footnotes. Just the argument from start to finish: 🔸 Why electrodynamics has 16 components and uses 6 🔸 The experiment that proves the deleted components are physically real (Aharonov-Bohm, 1959) 🔸 The patented industrial device that transmits through solid metal shields 🔸 What we can engineer with the remaining 10 👉 https://t.co/HqwXDJVeng (Yeah, I usually don't like these NotebookLM generated videos as well, but I think it turned out pretty OK — helpful for those that don't have the time to read the full paper!)
- 29
@VaibhavSisinty ·
A MIT student figured out how to make NotebookLLM do the thinking before the lecture even starts.🤯 His professors don't teach him anything new. They confirm what he already built. He calls it "context stacking." It is the most unfair study method I've come across. Here's exactly what he does :point_down::skin-tone-2: Two days before every lecture, he uploads everything into NotebookLM. The assigned readings, last week's slides, 3-4 related papers he finds himself, and any open problem sets. Most students wait for the lecture to explain the material. He walks in having already built a mental model of it. But that's not the move that makes it unfair. The first prompt he runs across all of it: "What are the 5 core concepts this week's content is built on, and how do they connect to what I studied last week?" NotebookLM pulls threads across everything he uploaded simultaneously. It surfaces relationships between ideas that would take a normal student weeks of review to notice. He gets that map before the lecture even starts. Then he runs the prompt that does most of the work. "What would I need to genuinely understand about this material to be able to teach it to someone with zero background?" That one question exposes exactly where his understanding is solid and exactly where it's hollow. The gaps show up immediately. He spends the rest of the session filling only those gaps. Not reviewing what he already knows. Only fixing what he doesn't. The final prompt is the one that separates context stacking from every other study method. "What question could a professor ask about this material that would expose a student who understood the surface but missed the underlying logic?" He's not studying for the exam he expects. He's studying for the exam designed to catch people who only think they understood it. By the time he sits in the lecture hall, the professor is confirming what he already mapped. Filling in a few details. Occasionally surprising him with something he didn't anticipate. That surprise is the only thing he writes down. The lecture was never the learning. It was always the confirmation.
- 30
- 31
@Scobleizer ·
You can't read 25,000 posts a day, synthesize them into a web site. Then make a podcast automatically. Mine can: https://t.co/8L5xphk0qQ There is a Notebook LM button at the bottom. It will put a script into your memory. Go over to @NotebookLM and paste it in and click "create." (No additional prompt needed). A few minutes later this podcast, mind map, slide deck, and shortly a video, pops out: https://t.co/8J7YscA8bS It is a podcast completely created by you here on X.
- 32
- 33
@pauliusztin_ ·
Andrej Karpathy calls it an LLM Knowledge Base. I've been building one on my own notes for months without knowing that was the name for it. I use Obsidian for notes, Readwise for reading, and NotebookLM for research. No AI reaches across all three. Generic tools like Perplexity and Gemini Deep Research only search the public web. Everyone gets the same sources. The real edge is your own curated thinking. I built three Claude Code skills (/research_create, /research_search, /research_distill) running on the obsidian, readwise, and NotebookLM CLIs. No vector database. No RAG pipeline. Just the filesystem, Markdown, YAML, and progressive disclosure via index.yaml.
- 34
@VaibhavSisinty ·
The last 7 days in AI were the wildest we've ever seen. A model got banned by the US government. A welder became a millionaire. And a trillion-dollar company was born. Here's everything that happened. 👇 → Anthropic launched Claude Fable 5 their most powerful model ever. 72 hours later the US government pulled it. First time in history export controls were applied to an AI model itself. → Amazon Anthropic's own $4B investor found the jailbreak and reported it to the US Treasury Secretary. The model went dark globally. → Apple held WWDC 2026. Tim Cook's final keynote. Siri was rebuilt from scratch on Google Gemini. A $1B/year deal. → SpaceX went public. $2 trillion valuation. Elon Musk became the world's first trillionaire. 4,400 employees became millionaires including welders and ship engineers. → Jeff Bezos came out of retirement. Launched Prometheus at $41B valuation. Building an "artificial general engineer" for the physical world. → Kimi launched Kimi Work 300 AI agents running locally on your desktop in parallel. → Kimi K2.6 now available for free on NVIDIA. No subscription needed. → Google dropped Gemini 3.5 Live Translate real-time speech-to-speech translation across 70+ languages. Already live in Google Translate app. → NotebookLM got a massive agentic upgrade. Now finds its own sources, generates diagrams, exports to Excel and PowerPoint. → ChatGPT can now build interactive charts from a single sentence. Bar, line, pie. Works on mobile. → Canva moved inside ChatGPT. Generate an image and turn it into an editable Canva design without leaving the chat. → Figma dropped a Chrome extension that captures any live website and imports it as editable layers. → Claude can now turn its answers into videos using HeyGen Hyperframes. Right inside the chat. → Replit added Skills set your preferences once and AI remembers them across every project. → Hermes can now run jobs on a schedule. Pick a blueprint. Set it. It runs while you sleep. → OpenAI Codex now lets you bank your resets instead of losing them on a timer. This wasn't a normal week. This was a turning point. AI stopped being a tech story and became a sovereignty story.
- 35
@IamEmily2050 ·
I used Gemini Deep Research + NotebookLM to study the new Samsung announcement about femtosecond laser, and generated the video with NotebookLM, you can learn anything or do deep dive on any subject and we still early, Veo4 +Nano banana V2 Pro +Gemini Pro 3.5 will take NotebookLM to the next level in the near future 🔥 Samsung Electronics is upgrading its semiconductor manufacturing by implementing advanced femtosecond laser technology for its wafer-cutting processes. This strategic shift aims to enhance the production of HBM4, as the ultra-fast laser pulses reduce particles and improve cutting precision compared to traditional mechanical methods. The company has already ordered at least ten specialized machines for its Cheonan campus to boost yield and maintain a competitive edge in the AI memory market. While currently focused on high-bandwidth memory, Samsung is considering expanding this technology to NAND flash and system semiconductors. Suppliers such as EO Technics and Disco are expected to compete for these significant equipment contracts as the industry moves toward more refined fabrication techniques.
Watch video
- 36
@TimHaldorsson ·
google has an AI tool that turns 50 articles into a custom 20 minute podcast i fed it my articles, my website, and my customer profile it built a fully personalised podcast that sounds completely human the craziest part — you can join in live and ask the hosts questions mid episode no more hoping a podcast covers what you need just build your own in minutes with NotebookLM
Watch video
- 37
@nurijanian ·
you should be NotebookLM-maxxing i use an agent that generates divergent questions based on the book A More Beautiful Question, which then talks to any NotebookLM via notebookLM cli (wrapped with a skill) you can use this for a few things: - fill your local notes with ideas from Notebooks that fit your context. for instance, I have all of Andrew Huberman's podcast episodes in a notebook, and I use them to improve my health notes - train skills on any topic and validate them against the source. i've written about this before: https://t.co/RrWUjfUSHV - expand your agent's in-chat knowledge with unique context that would have consumed lots of tokens and context on grokking pdfs and transcripts if used directly - read books with AI by asking hyper-specific deep questions from the book that map to your life - feed agents user research notes inline and much more comes part of my PM OS (link in bio if you want)
- 38
@maxcapacity ·
My AI conspiracy theory simulator has a podcast on Spotify now. I ran the agents' research through notebookLM to generate the audio. The three AI editors are using the research to build The Plan. The research is objective, and cited, but The Plan is constructed by the three AI agent editors as they try to connect all of the threads of esoteric history to each other.
- 39
@yaroslavvb ·
NotebookLM education videos are becoming really good looking, here is one I got asking to explain Thompson's Area-time complexity for VLSI
Watch video
- 40
@JustinThomasAI ·
Google just merged Gemini + NotebookLM. You can now use Gemini to chat with your notebooks, find sources, and add them directly. Every answer it gives you is grounded only in the sources you have approved, so you are getting answers pulled directly from your own curated knowledge base.
Watch video
- 41
@glenngabe ·
Using NotebookLM? -> Google tests Whiteboard Animation format in NotebookLM (and a new "Short" format) "A standalone whiteboard animation format would likely produce an entirely different video structure, mimicking the hand-drawn, progressively illustrated style that made the RSA Animate series a viral educational phenomenon in the early 2010s." https://t.co/zLAzhX8Jol
- 42
@onenewbite ·
Just did something incredible: created a skill/workflow for @antigravity so that it will automatically create a new notebook in @NotebooLM , generate the audio overview, and then push it to my private podcast! (Yes, I also create my own private podcast platform…). Will share more details in a future video.
- 43
@RoundtableSpace ·
NotebookLM just merged with Gemini. Your notebooks now live inside the Gemini app and your Gemini chats can become notebook sources instantly. Google's research and reasoning tools are becoming one thing.
Watch video
- 44
- 45
@JulianGoldieSEO ·
NotebookLM gives you the research. Google Antigravity turns that research into something real. Here’s the workflow: → Upload your sources into NotebookLM. → Ask it to analyze the idea. → Generate a technical build prompt. → Paste that prompt into Antigravity. → Let the agents create the files, code, and structure. → Upload the finished build back into NotebookLM. → Ask for improvements, security fixes, and design feedback. That’s the loop. Research. Build. Review. Improve. Launch.
Watch video
- 46
@frog_omo ·
10 notebookLM features most people miss: → index trick: ask it to index sources into topics, then explain each → cross-doc comparison: "where do these disagree?" → audio overview: podcast-style, can call in to ask questions → citations: click any claim, jump to source → flashcards/quizzes: one click from any PDF → auto timelines: pulls chronological events → "explain it like i'm 10": rewrites dense language → web pairing: research in perplexity, feed sources back in → notebookLM inside gemini: files + live web
- 47
@frog_omo ·
Yesterday, I watched my 7-year-old nephew do homework. He had a 40-page reading assignment. A whole chapter book due Monday. He looked at it, sighed, and said, "I can't read all this." His mom did something I didn't expect. She took the book and put sticky notes between each chapter. Five sections. Maybe 8 pages each. Then she said, "You don't have to read the whole book. Just read to the first sticky note." He finished in two hours. Asked if he could start the next book. I sat there, realising I'd been doing something stupid for years. I have 23 unfinished nonfiction books on my shelf. Each one I started with excitement. Read a couple of chapters. Then never touched again. I kept thinking I was bad at reading. Bad at focus. Bad at finishing things. Turns out I was just looking at 300 pages and letting my brain quietly give up before I started. I was missing the sticky notes. So I built my own version. With NotebookLM. Here's what I do now: I split the book into chapters first. Each chapter becomes its own source in a dedicated notebook. Now, a book isn't 300 pages. It's twelve 25-page sections. My brain doesn't panic. It just sees the next chapter. For each one I ask: → "What's this chapter actually trying to say?" → "What's the main framework here?" → "Give me 5 bullets and a mental model" I keep the book open on one side, NotebookLM on the other. Read a bit, ask a question, keep going. At the end, I zoom out: → "How do these chapters connect?" → "Where's the argument weakest?" → "Turn this into a 30-minute workshop" The shift that changed everything: I stopped trying to "finish books." I just finished the chapters. And somehow, I finish more books now than when I was trying to finish books. My nephew figured this out at 7. Took me years and a shelf full of abandoned books. The whole book is overwhelming. The next chapter is doable. That's it. That's the whole trick.
- 48
@JulianGoldieSEO ·
Google just turned NotebookLM into an SEO content machine. One tiny update eliminates one of the biggest content workflow bottlenecks. What's New: → Google Drive Auto-Sync updates NotebookLM the moment you edit Docs, Sheets, or Slides. → No more uploading the same file over and over. → Your research, SOPs, and content briefs always stay current. Why This Matters: ✓ Turn one research document into a YouTube outline, newsletter, LinkedIn post, and community post. ✓ Build a knowledge base that answers questions using only your trusted documents with citations. ✓ Keep every piece of content aligned with the latest version of your research. The biggest takeaway: NotebookLM isn't just a research tool anymore. It's becoming a live content workspace where one document powers your entire content system.
Watch video
- 49
@glenngabe ·
Personal Intelligence in NotebookLM -> Google NotebookLM is testing Personal Intelligence and AI note editing options, hinting at upcoming feature releases "The first, Personal Intelligence, would give NotebookLM a better memory. It would draw context from past conversations, retain it, and lean on it later, with controls to turn it off and inspect what has been stored. This version looks bound: it appears to learn only from activity inside NotebookLM rather than reaching across Gmail, Docs, or other Google surfaces, a deliberate limit suited to the privacy expectations of researchers and enterprise teams. It has appeared in testing before, at both the account level and per notebook, and its return suggests Google is edging it toward release." https://t.co/hBmzO5cCZa
- 50
@Creator_Toolbox ·
Google @NotebookLM added push notifications for finished notebooks. This lets you leave a page and be pinged when they are ready.
Watch video
Best Tweets by Topic