Introducing a more powerful NotebookLM 🚀 Massive upgrades deliver agentic capabilities in chat, more advanced reasoning, and a suite of new output formats. Tackling complex, multi-step research problems has never been easier. Rolling out now to Google AI Ultra subscribers.

Notice anything different about the NotebookLM mobile app recently? 😉 Well, we’re excited to REPORT that you can now create briefing docs, study guides, and blog posts on-the-go! 📱✨ Are there any other report formats you'd want specifically for mobile? Let us know!

A guy named Luis built a free alternative to Google’s NotebookLM that you can run completely on your own computer. Its called Open Notebook and gave it away on GitHub. You can upload: > PDFs and Office documents > Videos and audio > Websites > Research papers > Notes and other files Then ask questions, search across everything, and get answers linked back to the original sources. But Luis did not stop at copying NotebookLM. Open Notebook lets you: - Choose from more than 18 AI providers - Use OpenAI, Claude, Gemini, or DeepSeek - Run private models through Ollama or LM Studio - Create podcasts with one to four speakers - Control the speakers, voices, scripts, and episode style - Connect it to other tools through API or MCP The software is fully MIT licensed and has already crossed 37,000 stars on GitHub. You only need Docker to install it. Bring your own API key, or use a local model without paying API fees. NotebookLM gives you Google’s models on Google’s servers. Open Notebook lets you choose the brain, control the data, and run the entire thing yourself.

I accidentally dropped my competitor's YouTube video into NotebookLM. It told me exactly why it went viral, what his audience still wanted that he never gave them, and wrote me a script built around that gap. Here's how you can do the same in 60 seconds: Open NotebookLM. Paste the URL of the highest-performing video in your niche. The one you have been quietly jealous of every time it shows up in your feed. First prompt: "What made this video work? Break down the hook, the pacing, the emotional peaks, and the exact moment the viewer decides to stay." It does not give you vague advice. It tells you which sentence landed hardest and why. It surfaces the structure underneath the style. Second prompt: "What gaps did this video leave? What did the audience want that this video never fully delivered?" The comments section of a million-view video is a product brief your competitor wrote for you without knowing it. NotebookLM reads it like a researcher and hands you a list of unmet needs you can build your next video around. Third prompt: "Generate five viral video ideas that cover the gaps you just identified. Give me the hook, the core argument, and the emotional payoff for each one." The ideas it gives you are not generic. They are built from the actual evidence of what your audience already wanted and did not get. Your competitor cannot reverse-engineer them, because they came from studying him. Fourth prompt: "Write me the first 60 seconds of the strongest idea. Make it hook harder than the video we just analyzed." Most YouTubers spend three weeks planning and one hour studying what actually performs. This workflow inverts that entirely. The creators winning right now are not better on camera. They are better at reading the evidence their competitors accidentally left behind. You just got the playbook.

Here’s what launched this week: — Gemini 3.5 Live Translate our latest audio model for live speech-to-speech translation — @NotebookLM got a major upgrade including agentic capabilities in chat, more advanced reasoning, and a suite of new output formats — Project Genie from @GoogleLabs is now available to Google AI Ultra 5x subscribers globally — Notebooks in @GeminiApp are now available in the European Economic Area, United Kingdom, and Switzerland — DiffusionGemma, our newest experimental open @googlegemma model that explores text diffusion, an exceptionally fast approach to text generation

Catalyst Brands is building the future of American retail—using Gemini Enterprise, NotebookLM, and Veo to fast track product design to the closet. See how we're helping transform its powerhouse portfolio ↓

Time is the scarcest resource. Grok bot automations multiply it. They run continuously in the background compiling research insight, pulling latest data, executing full prompt chains, and generating video while human attention stays locked on irreversible problems. I ran the identical workflow through NotebookLM today. Grok still compounds further. This is not convenience. This is structural leverage on the only input we cannot manufacture more of. Automation that never stops is how the multiplanetary future gets built.

The top AI tools everyone should use (August 2026 - bookmark this): → Claude: best all-round reasoning, writing, and daily use. Start here if you're new to AI. → Claude Code: turns Claude into an agent that can actually touch your files, run code, and automate real tasks. → ChatGPT: strong for coding and voice mode, good to have alongside Claude for different tasks. → Perplexity: research and search with real citations, better than a normal search engine for anything you want sourced. → Hermes Agent: free, self-hosted AI assistant that remembers everything and runs 24/7 for a few dollars a month. → Grok Bot: an agent team with its own cloud computer per agent, connects into your actual apps (Gmail, Slack, CRM, etc). → Whisper Flow: voice-to-text that's faster than typing, changes how you prompt entirely. → Granola: auto-transcribes meetings and feeds notes directly into your AI's memory. → NotebookLM: great for digesting long documents, videos, and research into something you can actually query. → TradingView + an MCP connector: for anyone touching markets, connects your AI directly to real charts and data. → Obsidian: pairs with any AI as a local, ownable second brain.

It’s been a busy couple of weeks! ICYMI, here’s the recap ⬇️ — Gemini Robotics 2 from @GoogleDeepmind brings whole-body intelligence to robots — Gemini 3.5 Flash-Lite is our fastest, most cost-effective model optimized for high-speed agentic workflows — Gemini 3.6 Flash delivers faster, more accurate performance across tasks while using significantly fewer tokens — Gemini 3.5 Flash Cyber is our new, cost-effective model optimized for finding and fixing software vulnerabilities at scale, available exclusively to governments and trusted partners — Nano Banana 2 in @GoogleEarth lets you reimagine places and generate custom images using satellite, aerial, and 3D imagery — Lyria 3.5 is the newest music model from Google Deepmind, now powering @googleflowmusic — @Gemini_Notebook (formerly NotebookLM) launched Collections, a new way to organize your notebooks