Claire Xue

Seedance 2.0 now available in 1080p. Turns out size does matter.

We tested 3 elements + prompts, swapped sunglasses for headphones, and compared results. Pika Labs = creative & fun; Kling AI = precise & flexible. ✨ 📩 Full breakdown: buff.ly/4hm6oga Thoughts? 👇 #CreativeAI #AIgeneratedVideo

Everything you love about generative models — now powered by real physics! Announcing the Genesis project — after a 24-month large-scale research collaboration involving over 20 research labs — a generative physics engine able to generate 4D dynamical worlds powered by a physics simulation platform designed for general-purpose robotics and physical AI applications. Genesis's physics engine is developed in pure Python, while being 10-80x faster than existing GPU-accelerated stacks like Isaac Gym and MJX. It delivers a simulation speed ~430,000 faster than in real-time, and takes only 26 seconds to train a robotic locomotion policy transferrable to the real world on a single RTX4090 (see tutorial: genesis-world.readthedocs.io/en/latest/user…). The Genesis physics engine and simulation platform is fully open source at github.com/Genesis-Embodi…. We'll gradually roll out access to our generative framework in the near future. Genesis implements a unified simulation framework all from scratch, integrating a wide spectrum of state-of-the-art physics solvers, allowing simulation of the whole physical world in a virtual realm with the highest realism. We aim to build a universal data engine that leverages an upper-level generative framework to autonomously create physical worlds, together with various modes of data, including environments, camera motions, robotic task proposals, reward functions, robot policies, character motions, fully interactive 3D scenes, open-world articulated assets, and more, aiming towards fully automated data generation for robotics, physical AI and other applications. Open Source Code: github.com/Genesis-Embodi… Project webpage: genesis-embodied-ai.github.io Documentation: genesis-world.readthedocs.io 1/n

@doganuraldesign I thought @LumaLabsAI V1.5 don’t offer image to video yet!

Just testing different video AI tools. Luma // Runway // Kling Static shot: She's standing still. It’s raining, and lightning is striking. Dark cinematic atmosphere. Tbh, I just feel lucky to have these tools. Thoughts?

a single AI-driven product image into multiple video clips? 🌟 Here’s a step-by-step breakdown of my latest experiment: 1️⃣ Select and prepare your AI-generated background. 2️⃣ Integrate the product image using Photoshop. 3️⃣ Crop the final image into three parts. 4️⃣ Use @luma_ai keyframe to create video clips from the cropped images. 5️⃣ Stitch the clips together for a seamless video. Disclaimer: This experiment features a product from Tiffany, but it’s not commissioned by them. This is purely for educational purposes. While this method showcases the potential of AI in creative production, it’s important to note that it still can’t match the quality of a professional production team. The current technology has its limitations, but it’s exciting to explore and push the boundaries of what’s possible with AI. 📰 Want to stay updated with the latest in AI and creativity? Subscribe to my newsletter, ‘The AI Stuff Creators and Artists Should Know,’ for weekly insights, tips, and more!

🚨 “Most people are 100% wrong about AI-driven content.” 🚨 👉 Think AI tools are creating content on their own? Think again! 👈 The reality is, AI-driven content isn’t just about AI tools. It also requires a stack of traditional tools like Photoshop, After Effects, Adobe Premiere Pro, and other editing apps. To achieve that polished, high-quality final deliverable, we need both AI and traditional tools, along with domain expertise to know what content works for specific use cases. Anyone can use AI tools, but creating production-ready content takes skill, knowledge, and a combination of various tools. It’s time to switch up our mindset and recognize the real creatives behind AI-driven content. Want to stay informed about the latest in AI and creativity? Subscribe to “The AI Stuff Creators and Artists Should Know” for weekly updates on innovations, tutorials, and industry insights. Link in bio! #AIDriven #CreativeCommunity #AIMindset #AIandCreativity #Newsletter #StayInformed #CreativeTools

🚀 AI-Driven Product Videos with ST. George Gin 🍸✨ Curious about how AI can transform still images into engaging motion clips? Here’s a look at my latest personal experimental project with @LumaLabsAI to create content for @stgeorgespirits While not perfect, the results are intriguing and showcase the potential of this technology. Step 1: Setting the Scene 🔹 Started with an AI-generated background image displayed on my big monitor. 🔹 Positioned the transparent glass gin bottle in front of the screen to pick up realistic reflections and lighting. This step was important for achieving a natural look due to the bottle’s transparency. Step 2: Creating the Motion Clips 🔹 Uploaded the still image to @LumaLabsAI AI to generate various video clips. 🔹 Noticed some quirks with the three coffee beans in the martini glass – sometimes they floated, other times they vanished, or only one or two would disappear without instruction. Despite this, the anti-gravity effect looked cool and added a unique touch. Step 3: Observations & Insights 🔹 The coffee beans’ unpredictable behavior highlighted a current limitation of Luma AI: lack of control over fine details. 🔹 As the camera spun, the gin bottle’s color shifted slightly, influenced by the warm, orange-yellow tones of my environment. This created a visually interesting effect, even though the actual liquid should have remained transparent. Although we’re in the early stages of AI-driven product video methods, the rapid pace of technological advancement means that by the end of this year, we will likely see more polished AI-driven product commercials and campaigns. These will feature more consistent frames, on-brand images, and improved model performance, allowing more creators to produce high-quality videos without the need for expensive gear or large production crews. What’s Next? I’ll be conducting more experiments and sharing my findings. Follow along for more insights into AI-driven product images and videos! 📬 Join My Free Newsletter: Get the latest AI news, tutorials, and insights—free! 🚀 C.A.S.H. Camp: Our 5-week boot camp now includes bonus lessons on AI-driven product videos. Don’t miss out—link in bio!

🎥 AI-Driven Product Videos: @runwayml Gen 2 vs. @LumaLabsAI 🚀 I put two powerhouse tools to the test: Runway Gen 2 and the newly launched Luma AI. Here’s the scoop! Clip 1: Runway Gen 2 🌟 🔹 Generated a Purity Vodka bottle video with three 4-second clips. 🔹 Utilized Runway’s motion brush for precise control over segmentation and motion. 🔹 Pro: The motion brush allows for targeted motion areas, helping bypass AI’s struggles with human features. 🔹 Con: Text on the label got messed up, requiring retouching in Adobe Premiere Pro. 🔹 Exciting news: Runway is launching Gen 3 Alpha soon, and I can’t wait to see the improvements in consistency and fidelity! Clip 2: Luma AI 🌐 🔹 Luma AI impressed me with its accurate 3D predictions from a 2D image. 🔹 Pro: It handled dynamic camera movements well, accurately predicting the back and sides of the bottle. 🔹 Con: Struggled with generating hands and fine details without distortion. Editing tools are currently limited, but there’s great potential for improvement. 🔹 Bonus: No retouching needed for the label text, a big win! Although we’re in the early stages of AI-driven product video methods, the rapid pace of technological advancement means that by the end of this year, we will likely see more polished AI-driven product commercials. These will feature more consistent frames, on-brand images, and improved model performance, allowing more creators to produce high-quality videos without the need for expensive gear or large production crews. Key Takeaways: •Runway Gen 2 offers more control but needs text label corrections. •Luma AI excels in dynamic shots but has limitations in detail generation. •Both models have their strengths and areas for improvement, paving the way for the exciting future of AI-driven content creation. ✨ What’s Next? ✨ Stay tuned as I dive into Runway Gen 3’s capabilities，want to learn more? 👇 📬 Join My Weekly Newsletter: Get the latest AI news, tutorials, and insights—free! 🚀 C.A.S.H. Camp: Our 5-week boot camp now includes bonus lessons on AI-driven product videos. Don’t miss out—link in bio!

💡 Imagine transforming a basic iPhone shot into an interesting AI-driven product image. Swipe to see the before-and-after transformation While AI-driven methods might not be flawless, they creatively blend traditional and innovative AI techniques to produce unique, eye-catching visuals. Curious for more? Join my weekly newsletter, The AI Stuff Creators and Artists Should Know. Link in bio. #AIproductimage #AIdrivenproductimages #CASHCamp #Midjourney #photoshop

@nickfloats Means a lot from you, Nick!

@Moodelier so good

Seven seconds of AI-driven video, two days of worth of work: can AI transform video production in the commercial space? In my latest AI-driven product video experiment, I edited three different AI-generated backgrounds to showcase the potential of AI in product video creation. This seven-second clip, which took two people nearly two days to produce, highlights the integration of tools like @Midjourney, @Adobe Photoshop, After Effects, Premiere Pro, and @runwayml . While the clip is far from perfect, with challenges like shadow editing, color matching, and maintaining high fidelity, it features our favorite product from @kineuphorics Combining real-life models and products with AI backgrounds is complex, but the creative possibilities are immense. Despite the imperfections, this project was a valuable learning experience in exploring AI-driven video methods. Though not yet client-ready, it demonstrates the future potential of AI in reducing production time and costs while offering innovative design solutions. If you’re interested in exploring AI-driven methods in the commercial space, follow me for my weekly newsletter, “AI Stuff Creators and Artists Should Know.” Link in bio

The Vox Media and OpenAI partnership is truly fascinating. OpenAI is providing its cutting-edge AI technology to Vox, enabling them to develop innovative recommendation tools, analytics platforms, and other audience-facing applications. In exchange, Vox Media's authoritative content will enhance ChatGPT's knowledge base, with explicit brand attribution driving visibility and potential new readers to their publications. It seems ChatGPT is rapidly evolving into a powerful alternative to traditional search engines. Media companies recognize the immense value in partnering with OpenAI to secure prominent real estate within ChatGPT's responses, tapping into its vast user base. This strategic move by Vox Media could be a harbinger of more such partnerships as businesses aim to leverage the growing influence of AI assistants like ChatGPT. Additionally, the partnership includes developing AI-powered products for better ad targeting, creative optimization, and content recommendations. Vox Media's platforms like Forte and The Strategist Gift Scout will integrate OpenAI's technology, and Vox will get early access to new OpenAI products.

Watch me blend real and AI magic for mind-blowing product images! Thanks @kineuphorics for sending these beautiful products to me 😇🦋🍹 #AIdrivenimage

When you don’t have actual furnitures, and you are an AI visual educator and creator. Exploring AI-driven video content currently. A lot need to be improved before I can share more process. But yeah the furnitures are obviously AI-generated in here. Are you interested in AI-driven video editing?

@nickfloats “Her” -Sam Altman

Am I the only one that gets the ick from how flirty this is? It’s wayyy too much

@nickfloats Her

Would you ever consider licensing your video data for training purposes? If so, what terms and conditions are most important to you, and what would be your price per hour?

@Photoshop The beta gen fill and reference image isn’t working today:/

The latest features in Photoshop (beta) have arrived! The release is packed with new features and capabilities like Text to Image, Generate Similar, Reference Image, and more. Update your app today to expand your creativity. #PhotoshopAI adobe.ly/4dburNG

Over 13% of all images on Adobe Stock are AI generated. Most of the generated content comes from Dalle and Midjourney. Media tagged "fantasy" is 43% AI generated. Other tags are even higher This is a large portion of the training data that fuels Adobe's AI, firefly