📔 Google launches Featured Notebooks in NotebookLM 📔
Curated content partnerships with experts, publications, and nonprofits

Google has introduced Featured Notebooks in NotebookLM, a new feature that provides curated collections from respected authors, researchers, publications, and nonprofits to help users explore various topics with high-quality sources. T

he initial lineup includes longevity advice from Eric Topol (author of "Super Agers"), expert predictions for 2025 from The Economist's World Ahead report, life advice from Arthur C. Brooks' Atlantic columns, a science guide to Yellowstone National Park, human wellbeing trends from Oxford's Our World In Data, science-backed parenting advice from Jacqueline Nesi's Techno Sapiens newsletter, The Complete Works of William Shakespeare, and Q1 earnings reports from the top 50 public companies.

Each collection supports all NotebookLM features including AI-powered Q&A, Audio Overviews, Mind Maps, and citations from original material.

The launch builds on last month's public sharing feature, which has generated over 140,000 public notebooks in four weeks. Google plans to continue expanding featured notebooks with additional partnerships from The Economist and The Atlantic.

🔍 Mistral launches Deep Research mode in Le Chat 🔍
Plus voice interaction, multilingual reasoning, and advanced image editing

Mistral AI has significantly enhanced Le Chat with a suite of new features led by Deep Research mode, which transforms the assistant into a coordinated research partner that can plan, clarify needs, search, and synthesize structured reports with credible sources and references.

The update introduces Voice mode powered by the new Voxtral model for natural, low-latency speech recognition, enabling users to talk naturally without typing. Think mode leverages Mistral's reasoning model Magistral for complex multilingual reasoning, supporting code-switching between languages mid-sentence for thoughtful answers in any language. New Projects feature helps users organize conversations into context-rich folders with dedicated Libraries and persistent tool settings.

Additionally, Le Chat now offers advanced image editing in partnership with Black Forest Labs, allowing users to create and edit images with simple prompts like "remove the object" while preserving characters and details across series. All features are available at chat.mistral.ai and via mobile apps with no credit card required.

🌊 Windsurf releases Wave 11: Voice Mode and Browser Tools 🌊
AI code editor adds Cascade voice mode & deeper browser integration

Windsurf has launched Wave 11, a major update featuring voice support for Cascade AI assistant, enabling developers to dictate instructions using one-way voice input for handling long or complex prompts.

The update introduces deeper browser integration with screenshot capture, DOM tree access, and tab referencing through at-mentions, allowing Cascade to autonomously collect console logs and UI data.

New named checkpoints within conversations let users save and return to specific points, while the ability to at-mention past chats gives Cascade full context for follow-up tasks like generating tests. Additional features include Planning Mode by default, expanded JetBrains support, secure authentication for MCP servers, and auto-continue for complex queries that minimizes follow-up prompts.

INTERESTING TO KNOW

🎬 Netflix starts Using GenAI in Movies and TV Shows 🎬

Netflix has officially begun using generative AI in its shows and films, marking a significant milestone in Hollywood's AI adoption. Co-CEO Ted Sarandos revealed that Netflix featured the "very first GenAI final footage to appear on screen" in an Argentine show called "El Eternauta," where the internal production group collaborated with producers to use AI for creating a building collapse scene. The AI-generated scene was completed 10 times faster than traditional visual effects tools and at significantly lower cost. Sarandos emphasized that AI represents an "incredible opportunity to help creators make films and series better, not just cheaper," enabling real people doing real work with better tools. Netflix creators are already seeing benefits in pre-visualization, shot-planning, and visual effects, with advanced capabilities like de-aging now accessible beyond big-budget projects.

🤖 MIT creates Universal Robot Training Tool for Anyone 🤖

MIT engineers have developed a "versatile demonstration interface," a handheld, sensor-equipped tool that democratizes robot training by enabling anyone to teach robots new skills without coding expertise. The device attaches to common collaborative robotic arms and supports three training approaches: remote control (teleoperation via joystick), kinesthetic training (physically moving the robot), and natural demonstration (performing tasks while the robot watches). Previously, learning-by-doing robots trained using only one of these methods, but MIT's three-in-one interface allows users to choose whichever style best suits the task. Testing with manufacturing experts on factory floor tasks showed the tool's effectiveness in expanding training flexibility and the types of users who can interact with robots.