OpenAI's new video-generating model, Sora, is about to level up your gaming experience. This innovative AI model has a unique trick up its sleeve - it can simulate digital worlds, including rendering video games.
Sora, OpenAI's first video-generating model, is making waves in the tech world with its impressive capabilities. A recent technical paper revealed that Sora can generate videos of any resolution and aspect ratio, up to 1080p. But it's not just about creating high-definition videos. Sora can perform a range of image and video editing tasks, from creating looping videos to extending videos forwards or backwards in time, and even changing the background in an existing video.
The real game-changer? Sora's ability to simulate digital worlds. In a recent experiment, researchers fed Sora prompts containing the word "Minecraft". The result was a convincingly Minecraft-like game, complete with HUD, game dynamics, and player character control.
Sora operates as a data-driven physics engine, determining the physics of each object in an environment and rendering a photo, video, or interactive 3D world based on these calculations. This opens up exciting possibilities for the development of highly-capable simulators of the physical and digital world.
However, Sora isn't perfect. It struggles with accurately approximating the physics of basic interactions like glass shattering. And even with interactions it can model, there are often inconsistencies.
Despite these limitations, Sora's capabilities could potentially pave the way for more realistic, even photorealistic, procedurally generated games from text descriptions alone. While this is exciting, it also raises concerns about deepfake implications. As a result, OpenAI is currently limiting access to Sora.
With Sora, OpenAI is pushing the boundaries of what AI can do in the gaming world. We can't wait to see what comes next!
Made with TRUST_AI - see the Charter: https://www.modelprop.co.uk/trust-ai
Comments