Hi tech & AI enthusiasts, this week's highlights:
Gemini Omni Can Generate (and Edit) Video From a Conversation
Gemini 3.5 Flash Beats the Frontier and Runs 4x Faster
The Gemini App Becomes Your 24/7 Personal Agent
The Lab: How to Use Gemini Omni to Visualise Empty Rooms With Different Interiors
AI Image of the Week: Architectural photography of a newly built luxury residential interior
Let’s get into it.
▲ AI MODELS
Google launched Gemini Omni, a new video model that takes any combination of text, images, audio, and existing video as input and produces high-quality video as output. It’s available today in the Gemini app, Google Flow, and YouTube Shorts.
What you need to know
Edit video through chat: Drop in a clip and refine it through conversation. "Dim the lights." "Change the camera angle." "Make the violin invisible." Characters, physics, and scene continuity hold up across edits.
Combine inputs into one output: Feed it a reference image, an audio track, and a description, and it produces a video grounded in real-world physics and Gemini's world knowledge.
Includes Avatars and SynthID: Create a digital version of yourself that looks and sounds like you, and every output carries Google's invisible watermark so it can be verified.
Key Ideas
▲ Why it matters: Most AI video tools still feel like slot machines, where you press generate and hope for the best. Omni shifts video generation into the same conversational model you already use for everything else. For any business that markets with video (property, hospitality, e-commerce, agencies, training), the cost of producing a polished clip drops dramatically. The tutorial below shows exactly how a construction firm can use it.
▲ AI MODELS
Google released Gemini 3.5 Flash, the new default model in the Gemini app and AI Mode in Search. It outperforms Gemini 3.1 Pro on almost every benchmark, runs 4x faster than competing frontier models, and costs less than half what they do.
What you need to know
Built for long-running work: Designed for agentic tasks that used to take a developer days or an auditor weeks. Shopify is using it for merchant growth forecasts. Macquarie Bank is using it to read 100-page documents during customer onboarding.
Frontier intelligence at Flash speeds: Sits in the top-right quadrant of the Artificial Analysis index, the first model to deliver both top-tier intelligence and the fastest output speed.
Massive cost savings: Google estimates companies running 1 trillion tokens a day could save over $1 billion annually by shifting 80% of their workloads from other frontier models to 3.5 Flash. Pro version coming next month.
Key Idea
▲ Why it matters: This is Google's answer to "the smartest model is also the slowest and most expensive." For business users, the practical impact is that AI agents handling real work (customer onboarding, forecasting, document review) become economical to run at scale instead of one careful task at a time.
▲ AI TOOLS
Google rebuilt the Gemini app around two new agentic features: Daily Brief and Gemini Spark, plus a new design language called Neural Expressive. The app now has 900 million monthly active users, more than double last year.
What you need to know
Gemini Spark, your 24/7 agent: Runs in the cloud, keeps working when your laptop is closed, and handles multi-step work like flagging hidden subscription fees, sending a daily school digest, or drafting project kickoff emails. Integrates with Gmail, Docs, Slides, Canva, OpenTable, and Instacart at launch.
Daily Brief: A morning digest that reads your Gmail, Calendar, and notes, then prioritises and suggests next steps. Built to be skimmed in under a minute.
Voice and design overhaul: New Neural Expressive design language with fluid animations and integrated voice. The new Gemini app for macOS is rolling out with Spark and voice features this summer.
Key Idea
▲ Why it matters: Until now, AI assistants answered questions and waited. Gemini Spark is one of the first agents that does work in the background and reports back. If it lives up to the demos, the morning email-and-calendar triage that eats most professionals' first hour might genuinely be gone.
▲ EVEN MORE NEWS
Other big ideas from the past week.
▲ Antigravity 2.0 Launches as a Standalone Desktop App: Google upgraded Antigravity into a full agent orchestration platform with a new desktop app, CLI, and SDK. The new $100/month AI Ultra plan gives 5x the Antigravity usage of Pro.
▲ Google's Intelligent Eyewear Arrives This Fall: Audio glasses built with Gentle Monster and Warby Parker, powered by Gemini. Translate menus, navigate hands-free, take photos, and even have Gemini order your coffee on DoorDash while your phone stays in your pocket.
▲ Gemini for Science Lands With Backing From Nobel Laureates: Google launched Gemini for Science, a research toolkit covering hypothesis generation, computational discovery, and literature analysis. BASF is using it on supply chains, Klarna on ML training. Validation papers published in Nature today.
The Lab
▲ AI EDUCATION
This week you’ll learn how to create a walkthrough video of a newly finished but empty room, then use Gemini Omni to generate multiple interior design variations from a single clip. You can show clients three furnished options before any furniture is bought.
What you'll learn:
How to shoot a usable walkthrough video of an empty room (or generate one with Gemini Omni)
How to use Omni to dress the same room in completely different styles (Scandi, industrial, family-friendly)
How to keep the architecture, lighting, and camera path consistent across every variation
How to pair the output with Nano Banana stills to create a full pitch pack
🔬 Interested in learning about AI for business? Join The Lab waitlist.
▲ AI IMAGE SHOWCASE
Create this type of image with Midjourney
Architectural photography of a newly built luxury residential interior, freshly completed and empty. Pale oak herringbone floors, full-height crittall windows letting in afternoon light, exposed black steel beams across a high ceiling, one feature wall of unrendered brick. Camera at eye level, wide angle, soft natural light, no furniture, sense of scale, photorealistic, shot on a Sony A7R IV with a 24mm lens, f/4.Image Usage Suggestion: Construction firm portfolios, property developer marketing, project case studies, architecture awards submissions, hero imagery for interior design pitches, or starting frames for Gemini Omni walkthrough videos (as covered in the tutorial above).
💌 Reply to this email with your AI image generation or suggestion of what you’d like to see and I'll feature it in a future newsletter.
▲ SHARE YOUR THOUGHTS
Help shape the content you see here by giving feedback
Have specific feedback or want to get in touch? Reply to this email and I’ll get back to you.
Know someone who’d love this newsletter? Forward it to a friend and have them sign up here.
Thanks for reading, until next time.
Stay curious,
Matt Lok, Editor
Brought to you by Metalabs. Digital marketing consulting specialising in AI.