Priyal Walpita’s Post

MIT Technology Review’s What’s Next series looks across industries, trends, and technologies to give you a first look at the future. You can read the rest of them here. When OpenAI revealed its new generative video model, Sora, last month, it invited a handful of filmmakers to try it out. What’s next for generative video

1:12 p.m. screen time screen time OpenAI’s New Product, Sora, Is Already Producing Wild Videos By John Herrman It sure seems like a big jump forward in text-to-video technology.: … wearing a red wool knitted motorcycle helmet, blue sky, salt desert, … , and is capable of producing short clips from text prompts … these shortcomings into motion, producing impressionistic videos that look as … of a single company’s product, it’s reasonable to … #motorcycle #motorcycles #motorbike

🎉 Exciting News Alert! 🚀 I am absolutely blown away by OpenAI's groundbreaking new release! 🔥 Say hello to the future of video creation: OpenAI's Text-to-Video model, Sora! 🎥✨ Imagine this: With just a simple prompt and a compelling storyline, you can now generate stunning movies and advertisements, all without the need for expensive camera gear or elaborate shooting locations. 🌟 For below video given prompt was: A drone camera circles around a beautiful historic church built on a rocky outcropping along the Amalfi Coast, the view showcases historic and magnificent architectural details and tiered pathways and patios, waves are seen crashing against the rocks below as the view overlooks the horizon of the coastal waters and hilly landscapes of the Amalfi Coast Italy, several distant people are seen walking and enjoying vistas on patios of the dramatic ocean views, the warm glow of the afternoon sun creates a magical and romantic feeling to the scene, the view is stunning captured with beautiful photography. These remarkable clips, all courtesy of OpenAI's Sora model, showcase its unparalleled ability to create lifelike and captivating scenes entirely from scratch. 🎬✨ Here's what makes Sora a game-changer: 👀 Detailed Video Generation: Sora can effortlessly bring to life intricate scenarios, whether it's a chic woman strolling through the bustling streets of Tokyo or a gripping movie trailer. ✔ Consistent Visual Quality: Rest assured, Sora maintains top-notch visual quality and coherence throughout the entire video, ensuring a seamless viewing experience. 👉 Real-world Understanding: Sora comprehends the laws of the physical world, ensuring that characters and scenes behave in a believable and realistic manner, adding depth and authenticity to your creations. 🧠 Personalization at Its Finest: Users have the power to guide Sora's memory, tailoring future interactions and enhancing the level of personalization in every video generated. 🖼 State-of-the-Art Technology: Powered by a cutting-edge diffusion model and transformer architecture, Sora boasts unmatched scalability, effortlessly extending videos and breathing life into still images with astonishing accuracy. Get ready to revolutionize the way you create videos! With OpenAI's Sora at your fingertips, the possibilities are truly endless. 🌈✨ #OpenAI #Sora #TextToVideo #Innovation #FutureOfCreativity

The Future of Video is Here! 🚀 OpenAI's Sora Generates Stunning Videos from Text Prompts, revolutionizing creative industries! 🎥 Craft high-quality, 60-second videos with unprecedented accuracy in mood, lighting, and believability. 🤯 🌐 Sora's Game-Changing Features: 1. Unprecedented Accuracy: Say goodbye to awkward transitions – Sora captures facial expressions and reflections seamlessly for truly believable visuals. 2. Broad Impact: Filmmakers, graphic designers, storytellers – brace yourselves! Sora transforms workflows and unlocks new creative possibilities. 3. More than Just Video: Text-driven worlds, characters, and narratives – welcome to the future of content creation! Just the Beginning: Industry experts predict Sora could make stock footage obsolete. OpenAI takes a cautious approach, releasing it to a limited group for safety testing. 💬 Join the Conversation: 🤔 What are your thoughts on Sora and its implications? 🌐 How will this technology impact your industry? 🤖 What ethical considerations need addressing? 🗣️ Engage with us in the comments and let's explore the future of video, together! 🌈 #OpenAI #Sora #TextToVideo #AIInnovation #CreativeIndustries #TheFutureIsNow 🚀

Sora is able to generate complex scenes with multiple characters, specific types of motion, and accurate details of the subject and background OpenAI’s Video Generator Sora Is Breathtaking, Yet Terrifying https://flip.it/g7nK.A

Breaking: OpenAI's new text-to-video model 🤯 Imagine a world where videos are not just created but are brought to life with an understanding of real-world dynamics. That's exactly what OpenAI has achieved with Sora. This groundbreaking model is capable of generating detailed, realistic, and imaginative video clips directly from complex prompts - no modifications needed! Here's a glimpse into the magic that Sora brings to the table: 🌆 Vivid Storytelling: Whether it's a stylish walk through the bustling streets of Tokyo or crafting an entire movie trailer, Sora brings your visions to life with unparalleled detail and creativity. 🎨 Consistent Quality: Throughout the video, the visual quality and consistency are maintained, making each scene a masterpiece. 🌍 Real-world Understanding: Sora has a grasp of physical laws, ensuring characters and scenarios act and interact believably. 💡 Customized Memories: Users have the power to influence what Sora remembers or forgets, allowing for personalized content creation that gets better over time. 🖼 Advanced Technology: Built on a sophisticated diffusion model and transformer architecture, Sora not only scales beautifully but also extends videos and animates still images with astounding accuracy. The potential of Sora to revolutionize how we approach video content creation and solve practical problems by understanding real-world dynamics is simply staggering. OpenAI continues to push the boundaries, and I can't wait to see where Sora takes us next. Let's embrace this incredible journey of innovation together! #OpenAI #SoraModel #Innovation #VideoGeneration #AIRevolution

#OpenAI’s #Sora can generate videos up to a minute in length and features realistic scenes and multiple characters, all based on typed-in instructions.

How many white collar Income Taxes do you need to lose to Ai before govts and policymakers take note? How many democracies compromised? Oh hark the breakneck speed of Ai and LLM dev - last week I was trying to explain to non-tech friends the profound speed using the size of context windows and Google's Gemini as a guide: Nov'22 ChatGPT3 free version released with 2048 (2k) token window Jan'23 ChatGPT 3.5 turbo free version 4096 (4k) and enterprise 32k Nov'23 ChatGPT 4 =Free 8k... Plus/teams 32k... Enterprise = 128k Apr'24 Google's Gemini 1.5 Pro 1m token window That 1m tokens is ~700k words (3x James Joyce's Ulyses, almost all Harry Potter series). 2k to 1m in ~18mths. Even then, by the end of the same week a Google research paper suggested infinite windows were being tested. OpenSource LLMs and smaller parameter models are performing as well propietary large ones. This and token window size matter, because only the week before Anthropic (at the tip of the spear with OpenAi, Google, Meta, Mistral and Microsoft) released research showing the larger a context window, the easier it was to jailbreak ANY LLM using the many shot prompt technique plus other tricks. 🤕 Dec'23 an Ai startup was taking single images and turning them into simple lip syncing avatars. 16 Apr'24 Microsoft released a research paper about VASA-1 trained on VoxCeleb2 - zero latency, infered 3d, any angle, expressions, head tilts, blinking etc from a single image https://lnkd.in/e7u5mg2A Meantime actors and writers striked last year to get their industry to protect them and that was before Sora or VASA-1 etc. In music industry artists are up in arms after similar pace of development with Ai created music and song writing (pls try out UDIO or SUNO). Klarna is replacing people with LLM sales+call agents. Microsoft showed just 15secs of any voice are enough to create a clone. And the laggard topic in Ai? The advance of super alignment and ensuring none of the above can be used against human interests etc. We just need a few bad actors, not the matrix, sentience, AGI or super intelligence. Why is any of this important? Because ego, human avarice and an unwillingness to consider 2nd or 3rd order consequences are our living legacy (fossil fuels, debt?). We realise given a 2nd shot we'd handle social media (SM), screen time, addiction tech etc for adults let alone children differently. The mental health damage from SM is now a quantifiable economic and societal burden. And I've not even opened the pandora's box of Ai agents being a force multiplier with LLMs and Deepfake video. Autonomy to rewrite code, Ai training data or LLMs that can evaluate entire social feeds, A/B test and target replies at machine speed. Unlike the human speed of Cambridge Analytica and putin troll meddling in the referendum or US election. In 2024+25 we+policymakers need to be fully aware+act faster. The genie has never been closer. #amazingai #respectai #tmrw_is_too_late

🚨 OpenAI Releases New Text-to-Video Model, Sora ➡️ Sora can transform text into HD videos, animate still images, fill in missing frames and augment previously generated videos. 🌐 Read more about it on the blog: https://lnkd.in/ep4vyCij

Did you read this earlier? OpenAI's Sora is coming to video creation. Or should that be coming 'for' video creation? #SoraVideoCreation #EfficientVideoMaking #TextToVideo #CreativeAutomation #RevolutionizingVideoProduction