OpenAI launches Sora 2 model along with new Instagram like app

OpenAI has announced the launch of Sora 2, its advanced text-to-video generation model, alongside a new social media platform that harnesses this innovative technology. This release positions Sora 2 as a direct competitor to Google's Veo-3, which gained immense popularity shortly after its introduction earlier this year. According to OpenAI, Sora 2 significantly pushes the boundaries of what previous video generation models could achieve. The company emphasized in a blog post that this new model excels in adhering to physical laws—an essential feature for realistic video creation. For example, unlike earlier versions where a basketball could unrealistically teleport into the hoop after a missed shot, Sora 2 accurately depicts the ball bouncing off the backboard. The capabilities of Sora 2 extend to improved controllability, allowing it to follow complex instructions across multiple shots while maintaining consistent world states. It is designed to perform exceptionally well in various styles, including realistic, cinematic, and anime. Similar to Veo-3, Sora 2 can generate background soundscapes, dialogues, and sound effects that enhance the viewing experience, providing a greater sense of realism. An intriguing feature of Sora 2 is its ability to seamlessly integrate real people or objects into the generated videos. While OpenAI acknowledges that the model is not flawless and still makes errors, it represents a significant step toward accurately simulating reality through advanced neural networks trained on video data. The company is also aiming to create a buzz with a new social media application, Sora, available on iOS. This platform enables users to create and remix video content, discover new videos in a personalized feed, and incorporate themselves or friends into the app's scenes through a simple recording process. To address concerns about excessive screen time and content addiction, OpenAI has implemented features that give users control over their feeds. A newly developed recommendation algorithm allows users to guide content suggestions through natural language. By default, the app prioritizes videos that are closely related to users' interactions and the accounts they follow, focusing on content that inspires creativity in their own video projects.