Thinking Machines Lab (TML) just introduced a research preview of interaction models, a new kind of AI system built to collaborate live across voice, video, and text — letting users talk, show, interrupt, and steer while the system keeps working. The model takes in voice, video, and text in 200ms chunks, perceiving and responding in a streaming loop without the turn-taking pauses of other rivals. https://lnkd.in/dneUQ9sw

Google’s Gemini Omni signals a major shift in how generative AI systems are evolving from separate tools for text, images, audio, and video toward unified multimodal creation engines. What makes Gemini Omni especially significant is not just that it can generate video with synchronized audio from multiple input types, but that Google is attempting to collapse the entire generative workflow into a single AI-native interface. Instead of moving between different models for editing, animation, sound, and reasoning, creators interact with one foundation model capable of handling the entire pipeline. The broader implication is that AI creation is moving beyond isolated prompts toward continuous multimodal environments where text, voice, visuals, motion, and simulation coexist inside the same system. That could fundamentally reshape content creation, entertainment, advertising, gaming, and digital storytelling over the next decade. The race in generative AI may no longer be about building the best single model but about building the most seamless multimodal ecosystem. Could “anything-to-anything” generation become the defining interface of the next AI era? Follow our news here: https://lnkd.in/gkbDc_yd #AI #Google #Gemini #ArtificialIntelligence #GenerativeAI #Technology #Innovation #DeepMind #MachineLearning #FutureOfWork

If you haven't seen Thinking Machines new interaction model yet, this really seems like something new to me. The model gets rid of the classical turn-based model where the user and AI assistant interact sequentially and replaces it with continuous audio, video and text streams instead. Highly recommend reading they're technical post: https://lnkd.in/dukpV3dT

#The era of “Prompt-and-Pray” AI video is officially dead. 💀🎬 If you’ve been tracking the leaks over the last few days, you already know what’s coming at Google I/O next week: #Gemini Omni. For months, the AI video space has been a race of pure generation—who can make the prettiest, most cinematic 10-second clip. But anyone who actually builds creative workflows knows that a pretty clip you can't edit is useless. Google is completely changing the playbook. Based on the UI leaks and early beta rollouts, Gemini Omni isn't just a video generator. It’s a native, multi-modal creative engine that brings text, image, audio, and video into a single pipeline. Why this changes the game for creators, marketers, and developers: ⚡ Conversational Directing: Instead of rewriting a massive prompt from scratch, you edit inside the chat. "Change the lighting to cyberpunk," "Swap this character’s shirt," "Remix this into a 16:9 template." ⚡ Native Audio & Avatars: Seamlessly syncing AI avatars and voice generation directly into the timeline without bouncing between three different third-party apps. ⚡ The Workflow Ecosystem: By bypassing separate standalone models, this puts a unified creative engine straight into YouTube Shorts, Workspace, and Vertex AI. We are moving away from AI as a slot machine and toward AI as an assistant director. Is a unified workflow more important than raw cinematic physics? Or will standalone models still hold the crown? #GenerativeAI #GoogleIO2026 #GeminiOmni #TechInnovation #VideoAI #FutureOfCreative #AIWorkflows

𝗣𝗶𝗰𝘁𝘂𝗿𝗲 𝘁𝗵𝗶𝘀: 𝗮 𝘄𝗼𝗿𝗹𝗱 𝘄𝗵𝗲𝗿𝗲 𝘆𝗼𝘂𝗿 𝘁𝗲𝘅𝘁 𝗰𝗮𝗻 𝗰𝗼𝗻𝗷𝘂𝗿𝗲 𝗛𝗼𝗹𝗹𝘆𝘄𝗼𝗼𝗱-𝗴𝗿𝗮𝗱𝗲 𝘃𝗶𝗱𝗲𝗼𝘀 𝘄𝗶𝘁𝗵 𝘁𝗵𝗲 𝗰𝗹𝗶𝗰𝗸 𝗼𝗳 𝗮 𝗯𝘂𝘁𝘁𝗼𝗻. Google’s Veo 3 brings this futuristic vision into reality, yet beneath the mesmerizing visuals lies a strategic crossroads many might overlook. Introducing Veo 3, Google’s most sophisticated AI video generator, now elevating Vertex AI to new heights. It blends unparalleled visual fidelity with intricate control over cinematics, allowing enterprises to craft immersive visuals that once needed entire film crews. But could this allure of instant cinematic magic mask operational complexities? While industries from marketing to media stand to revolutionize rapidly, the stark contrast between visual potential and executional reality poses a strategic dilemma. How ready is your organization to go beyond image-driven enthusiasm to truly integrate this technology into everyday workflows? As we stand on the cusp of reshaping visual content creation, leaders must decide: are we capturing this opportunity fully, or are we at risk of it becoming just another ‘set piece’ in our digital arsenal? ──────────── 📌 If today were an audit day, how resilient would your AI governance truly be? 📰 Weekly AI & Robotics intelligence delivered via my Newsletter (link in first comment) 🔔 Follow Manni Ostermann for real-world AI Operations signals from enterprise deployments 💬 Reply with "GOVERNANCE" and I’ll send you the link ♻️ Repost & save to pressure-test your own AI governance before reality does #AI #VideoGeneration #CinematicAI #GoogleCloud #AIGenerated

Google just changed the AI conversation again with Gemini Omni. And honestly? This feels bigger than “AI video.” Gemini Omni is moving AI from generating content… to simulating entire experiences. Text, images, audio, and video become inputs. Cinematic scenes, motion, synced sound, edits, and interactions become outputs. What makes this scary powerful is not just the model itself. It’s the ecosystem behind it: • Gemini for conversational creation • Flow for AI filmmaking • YouTube for distribution • TPUs for massive compute • DeepMind for research • Android + Chrome for reach Most companies can build AI models. Very few can own the entire pipeline. And that’s where Google becomes dangerous. The most interesting part? Gemini Omni is quietly teaching AI how the world behaves: Physics Motion Scenes Audio Human interaction Camera movement That’s what a true “world model” looks like in consumer AI. The future may not be apps anymore. It may be generated environments, generated workflows, and generated experiences. Ironically, Google might build the metaverse without ever calling it the metaverse. Not a world you enter. A world you generate. 🚀 #Google #Gemini #AI #ArtificialIntelligence #FutureOfAI #UXDesign #ProductDesign #SaaS

Thinking Machines Lab just shared a research preview of their new interaction models, and I think we’re looking at a major breakthrough in interaction design. The big shift here is moving away from the "walkie-talkie" style of AI. These models work in real time, listening and seeing as you work without the awkward pauses. For those of us building digital products, this is a massive leap toward a more natural, multimodal way of collaborating with technology. It moves the needle from just giving commands to an agent toward a truly live, interactive experience. Check out the demos here: https://lnkd.in/gabWUSgp

Some teams are exploring the idea that keeping humans in the loop gives AI impressive capabilities. Or maybe it’s the other way around. Either way, context hierarchy, knowing what matters, when, and to whom… remains deeply tied to agency itself. Maybe even free will. And since those are the only agents I really care about, this kind of thinking from Thinking Machines Lab is now fairly high on my watch list.

Thinking Machines Lab (TML) just introduced a research preview of interaction models, a new kind of AI system built to collaborate live across voice, video, and text — letting users talk, show, interrupt, and steer while the system keeps working. https://lnkd.in/dcGYXtsT

The story of how Collective Artists transformed with Galleri5, our bet on AI-native entertainment, and Microsoft’s partnership in scaling that vision is one of the most defining journeys I have been a part of.