CoreWeave’s Post

Mixing Houdini's proceduralism to generate synthetic data and train models with the power of running on Conductor on CoreWeave, kickstarts what VFX artists can do with AI/ML. DM me if you want to chat more about this. #sidefx #houdini #ai #ml #conductor #coreweave #vfx

Life Will Never Be the Same After OpenAI's Sora OpenAI's Sora, introduced on February 15, 2024, is revolutionizing the way we interact with artificial intelligence. Unlike its predecessors, Sora can transform text prompts into one-minute videos, unlocking new possibilities for creative expression. Sora is not just another AI model; it's a gateway to a new realm of imagination. By seamlessly translating text into dynamic video content, Sora empowers users to bring their ideas to life in ways previously unimaginable. Behind Sora's magic lies a blend of advanced technologies. From the foundational Transformer model, enabling nuanced understanding of language, to Vision Transformer (ViT) and Video Vision Transformer (ViViT), capturing visual context, each component contributes to Sora's ability to create compelling videos. OpenAI's Sora is not merely a tool; it's a catalyst for creativity. By bridging the gap between text and video, Sora opens doors to endless possibilities in storytelling, education, marketing, and beyond. As we embrace this new era of AI-driven creativity, Sora stands as a testament to the power of innovation in shaping our digital landscape. #soraopenai #sora #ai #generativeai #aiautomation #chatgpt #openai

I had a quiet moment on Sunday (thanks to my kids deciding to sleep in) where I realized that every day I work with some form of AI in my role at Wisdom Analytics Inc. It's been incredibly transformative for me, including seeing our #AI pipeline plowing through massive amounts of data, creating these amazing models that answer questions that are out of bounds for others (shameless plug, I know!) #AI alone isn't going to change the world. It's the science in the Art+Science equation that has been around forever - it's just a different sort of science - and it's constantly being updated. In the video below, the discussion is around the use of OpenAI #sora - an AI video generation implementation - combined with the Unreal Engine - which I think has been one of the most influential game engines for many years. Game engines like Unreal have been used in the past for simulations in business, such as medical trials. With the overlay of Sora, there's significant potential for businesses to use this type of combination to create so many other applications. Think marketing, product showcases, virtual events... the list goes on. The video is a little "techy", but hopefully you will also see the potential AI has in changing how many aspects of life can operate. Let me know your thoughts! #wisdomanalytics #videoai https://lnkd.in/gwc9ZhPY

Running LLM/GenAI training workloads on cloud based infrastructure can be fairly involved. There are multiple considerations to factor in: about compute, network, scaling, costs, etc. In this talk at the Scale by the Bay conference, Oleg Avdeëv and Riley Hun went into the details of this topic and discussed how Metaflow enabled such workloads at Autodesk. Check it out!

Imaging generating video just by writing prompts for them ! OpenAI just dropped a bombshell in the AI world with SORA - a groundbreaking leap in video generation models! 🚀 This game-changing release trains massive generative models on diverse video and image data, paving the way for a new era of AI-driven simulation. Check out this read : https://lnkd.in/d75DS_-E #OpenAI #SORA #VideoGeneration #AIInnovation 🤖

To anyone who wants to use 3D as a guide for more control on AI outputs, Ardy Ala just dropped an amazing detailed tutorial on his youtube channel. Check it out!

🌌 Unveiling the Future: OpenAI's #Sora Revolutionizes #AI with Unmatched Physics & Emergent Abilities! 🚀 Dive into the future with OpenAI's groundbreaking Sora model! 🌠 Wes Roth takes us on an exhilarating journey, revealing how Sora is not just a leap but a quantum jump in AI video generation. Imagine a world where AI-created videos are indistinguishable from reality. That's Sora for you! 🎥✨ Sora stands out by simulating 3D environments with physics so real, you'll have to do a double-take. From the delicate balance of ice cubes in a glass to the dynamic battles between pirate ships in a cup of coffee, Sora's capabilities are nothing short of magical. 🏴☠️☕️ But here's the kicker: what we see is just the tip of the iceberg. Behind the scenes, Sora is a powerhouse, blending data-driven physics engines with intuitive physics, long-horizon reasoning, and semantic grounding. It's not just creating videos; it's simulating entire worlds, real or fantastical. 🌍🔮 Dr. Jim Fan from #Nvidia hints at the complexity and potential of Sora, suggesting it might even leverage Unreal Engine 5 for its breathtaking simulations. This isn't just about video generation; it's about building the foundation for #AGI through intricate rendering and intuitive physics. 🧠💡 Join Wes Roth as he unpacks the marvel that is Sora, offering a glimpse into the future of AI that OpenAI is shaping. This isn't just an advancement; it's a revolution. And we're here for it. 🌟🛠️ 👍 30 min.⏱here: https://lnkd.in/eq3yCcgr #OpenAI #SoraRevolution #AIFuture #PhysicsSimulation #AGIPotential

GPT vs. StarCraft II – 4 – Prompt Engineering & Monitoring This is the fourth in a series of webcasts covering the development and experimentation of using GPT algorithms to control the high-level strategy of a StarCraft II bot. In this webcast I’ll be using TensorBoard to monitor the statistics of the game and experimenting with prompt engineering to influence the strategy decisions made by the GPT model. https://lnkd.in/dzNpVf3H

OpenAI keeps on leading innovation and is on the forefront of an accelerating wave of #AI overtaking more and more aspects of the industry and our lives. #Sora, OpenAI's text-to-video product, shows unprecedented capabilities. It can create minute-long realistic video examples. Open AI has released examples (see also: https://openai.com/sora ) which demonstrate Sora's detailed rendering, from human textures to environmental consistency. It is clear that the model actually has knowledge about and is aware of the physics of the world it is rendering. Apart from the direct impact on the "creative" audiovisual industry, the fact that OpenAI's models seems to show non-linear growth in their capabilities is disconcerting and should make those who doubt the real, lasting impact of these technologies ("not real intelligence¨) at least question their preconceptions. https://lnkd.in/dvFnDwXJ

Following up on my recent subaquatic project with Houdini & ComfyUI, I've created a YouTube video to shed some light on the process and the challenges I encountered. 💡Here’s a brief overview: In the video, I guide you through the entire workflow. Together, we test each step, analyze the results, and identify solutions for various challenges, including discussing how prompts can be enhanced by incorporating LoRAs. Additionally, I briefly touch on upscaling and share my future plans for these projects. 🙋♂️While I'm not an expert in AI, I've kept up with the latest updates and best practices. I'm continuously working on various projects, and I plan to improve this setup as I progress. 🚨AI, as the name may imply, isn't a simple one-click tool; it requires continuous data evaluations and creative assessments. The extent to which you utilize it depends entirely on your discretion. 🎁How to download my ComfyUI workflow: By subscribing to our newsletter(newslounge.co), you'll receive the JSON file via email. Thank you and I hope this information has been useful to you. -Ardy 🔺Youtube video: https://lnkd.in/gHgizhd2 🤿Project Breakdown: https://lnkd.in/g_YBStf4 ✨Original post: https://lnkd.in/gEtWZse7 --------------------------------------- 👩💻 sign-up to our free newsletter: newslounge.co/subscribe --------------------------------------- #vfxartist #stablediffusion #comfyui #aiart