Howdy again. It’s Barsee, and welcome back to AI Valley.

Today’s climb through the Valley reveals:

• 🎬 Google has a new tool just for making AI videos

• 👥 Google 3d video chats

• 💻🧑‍🎨 Google made an AI coding tool specifically for UI design

• 🤖 Plus trending AI tools, posts, and resources

Let’s dive into the Valley of AI…

🔐 Secure Your AI Future at Snyk Launch 2025

AI is revolutionizing application development, accelerating innovation at an unprecedented pace. However, this rapid evolution brings forth new security challenges that demand immediate attention. Join the new Snyk Launch 2025 to explore how to build and maintain secure applications in the age of AI.

Why attend Snyk launch 2025?

• Discover strategies to match the speed and volume of AI-generated code, ensuring robust security without hindering development velocity.

• Learn how to identify and mitigate new vulnerabilities introduced by AI-powered applications and large language models (LLMs).

What you'll gain?

• Hear from Snyk's product experts and security professionals on the latest tools and practices for securing AI applications.

• Experience live demos showcasing Snyk's AI-enabled security solutions in action.

• Leave with practical approaches to implement AI security measures within your organization.

Don't miss this opportunity to fortify your applications against the evolving landscape of AI-related security threats.

Save your spot now

_{*This is sponsored}

THROUGH THE VALLEY

🎬 Google has a new tool just for making AI videos

Google has launched Veo 3, its latest AI video model featuring audio support for realistic sound effects and dialogue, designed to rival OpenAI’s Sora. Available through Google AI Ultra ($250/month) in the U.S. and for enterprise users on Vertex AI, Veo 3 offers improved text and image prompting with real-world physics and accurate lip syncing. Alongside Veo 3, Google introduced Imagen 4, which generates highly detailed images with fine textures like fabrics and water droplets.

Google also unveiled Flow, an AI filmmaking tool that combines Veo, Imagen, and Gemini, allowing users to create and stitch together 8-second video clips from text or image prompts using a simple scene builder. Flow is accessible on Google AI Pro and Ultra tiers in the U.S., marking a significant advance in AI-driven video creation.

Source

👥 Google found a way to make virtual meetings suck less with 3d video chats

Google has unveiled Beam, an AI-powered video communications platform designed to make video chats feel as lifelike as meeting in person, announced at Google I/O 2025. Building on its earlier Starline project, Beam uses a six-camera array combined with AI to create a real-time, 3D, life-sized hologram-like experience with near-perfect millimeter-precise head tracking at 60 frames per second (all without special glasses or equipment).

Developed in partnership with HP, Beam aims to overcome common video call issues like lack of eye contact and Zoom fatigue by offering a deeply immersive interaction. The platform may also feature speech translation like Google Meet and is initially targeted at large organizations looking to revolutionize videoconferencing. Early Beam devices are expected later in 2025.

Source

💻🧑‍🎨 Google made an AI coding tool specifically for UI design

The company also launched Stitch, a new Gemini 2.5 Pro-powered AI tool that turns text prompts and reference images into production-ready UI and frontend code in minutes. Available in Google Labs, Stitch generates multiple design variants based on your descriptions and even exports directly to Figma for refinements. It supports natural language input, visual references, and customizable styles, offering a fast, collaborative way for developers and designers to move from rough UI ideas to real, working interfaces.

Source

🤖 Google is bringing an ‘Agent Mode’ to the Gemini app

They also announced Agent Mode for the Gemini app and major updates to Project Mariner, Google’s AI agent tool. Agent Mode can complete tasks on your behalf, like finding apartments on sites like Zillow and even adjusting filters using Mariner. Mariner itself now handles up to 10 tasks simultaneously and introduces Teach and Repeat, where it learns from a single example to automate similar tasks later. An experimental version of Agent Mode will roll out soon for Gemini subscribers.

Source

👗 Google will let you ‘try on’ clothes with AI

Google is upgrading virtual try-ons: US users can now upload a full-length photo to see clothes on their own body (not just on models). Powered by AI that understands fabric textures and body shapes, this feature is live in Search Labs. Meanwhile, AI Mode in Google Search can now personalize shopping by considering your travel plans or weather needs. Coming soon: a “buy for me” tool that completes purchases for you using Google Pay, turning search into a full AI-powered shopping assistant.

Source

📷 Google is bringing real-time AI camera sharing to Search

Google also announced new ways to access its AI-powered “Live” mode, letting users search and ask questions about anything their camera sees. Now part of Google Search’s expanded AI Mode, it’s also coming to the Gemini app on iOS, after being available on Android for a month. Originally launched as Project Astra, this feature lets the AI chatbot interact with your camera feed, enabling real-time conversations — like getting recipe ideas from fridge ingredients.

Source

🔎 Google AI Mode in search expanded

Google has expanded AI Mode, its chatbot integrated into Google Search, to all U.S. users after a gradual rollout starting in March. Powered by the custom Gemini 2.5 model, AI Mode lives in a separate Search tab and handles complex queries like product comparisons. Google plans to integrate AI Mode features into core Search, including Deep Search for detailed reports and task automation via Project Mariner to book tickets or reservations. Upcoming updates will add custom charts, personalized responses, app integrations, and shopping capabilities.

Source

Other headlines

• NotebookLM to get Video Overviews soon.

• Google Gemini AI is coming to Chrome this week.

• Google says its new image AI can actually spell.

• Google Gemini 2.5 Pro is outperforming OpenAI models in Mathematics (49.4%), Code (80.4%), and Multimodality (84.0%).

TRENDING TOOLS

• Gadget - The fastest way to build and scale apps with AI.

• Fluig AI - Instantly turn documents & ideas into professional diagrams.

• BnbIcons - Create and animate Airbnb-like isometric icons with AI.

• Bolto - Find, interview, and hire engineers with AI.

THINK PIECES / BRAIN BOOST

• The Codex of Ultimate Vibing.

• Ask HN: When will managers be replaced by AI? (Hacker News Thread)

• How I learned complex topics 10x faster with NotebookLM.

• Gemini Diffusion, DeepThink, Imagen 4, Veo 3, Flow, Jules by AI Google.

VALLEY GEMS

THAT’S ALL FOR TODAY

Thank you for reading today’s edition. That’s all for today’s issue.

💡 Help me get better and suggest new ideas at [email protected] or @heyBarsee

👍️ New reader? Subscribe here

Thanks for being here.

REACH 100K+ READERS

Acquire new customers and drive revenue by partnering with us

Sponsor AI Valley and reach over 100,000+ entrepreneurs, founders, software engineers, investors, etc.

If you’re interested in sponsoring us, email [email protected] with the subject “AI Valley Ads”.