New AI Play Toy – Coming Soon

byBeadCatz

492 followers

2 years ago

It’s always been predicted, and I always believed, that changes and new programs would come fast and furious once AI started. We now have a new AI tool in the testing stage that sounds so exciting, and I can’t wait to try it. Forgive me if this has been posted before, but I look forward to getting my hands on it. It’s a text-to-video AI module called Sora, released by OpenAI, which has already given us ChatGPT. Sora can generate a minute-long video and maintain the visual quality while adhering to the user's prompt input. While there are a few text-to-video tools out there, this one makes things look very promising. The videos posted on OpenAI make this new tool look promising.

According to OpenAI, Sora can generate complex scenes that include multiple characters, specified motions, and comprehensive subject and background details. This new tool is supposed to have a better language understanding, which helps it interpret prompts and generate characters and their emotions more accurately. Sora can also provide a single multi-shot video. The company states that there is room for improvement at this time as it struggles with simulating a complex scene and may not understand the prompts for cause-and-effect instances like changing the scene to show physical changes after a character may do something like taking a bit out of their food. Or spatial details in a prompt may get confused, like differentiating between left and right or delivering precise even descriptions as they take place over time, like in camera angle details.

However, despite the above limitations, Sora is a model for understanding and simulating real-world scenarios, which is an exciting feature for all AI. This new tool will generate videos from text instructions, extend an existing video and fill in missing frames, or use a still image to generate a video with accuracy and minute details.

Learn more about OpenAI’s new tool, Sora.