Openai's SORA 2 pushes the boundaries of AI video creation and is completely in spoken language

Openai has released the Sora 2, a video creation model that takes AI video to new heights. For now, access is limited, apps and websites are free and have usage restrictions based on available processing power.

https://www.youtube.com/watch?v=lecg6aj6dvy
SORA 2 addresses the problems of previous AI video models by focusing on physical realism. Unlike the mostly visual first Sora, this produces synchronized conversations, sound effects, and background soundscapes. Prompts for two climbers screaming in a snowstorm produce pictures of wind whip coats, ice skin faces, snow crunches, and despair in their voices. You can control your model with specific commands across multiple shots, with consistent characters, settings and actions throughout the scene.

The new SORA iOS app launched with the model adds a social aspect to video creation. This is the first invitation-only site to be available on iOS in the US and Canada. Create 10-second videos, view curated feeds, and remix other people's works. The stream shows content from videos where the people you follow and the videos that inspire your own thoughts, so you don't get lost in endless scrolls.

https://www.youtube.com/watch?v=V2OHXIHNSQW
Another cool feature is the “cameo”. This allows you to put yourself and others in the video. After one video and audio recording to confirm your identity, the model is transported to any setting whether you are hiking through snowy mountains or becoming friends with friendly Bigfoot. OpenAI prioritizes user controls. Decide who is available to use your images and you can cancel or delete your videos at any time.

CHATGPT Pro users get a high-quality “Sora 2 Pro” model, but the Sora 1 Turbo is available to everyone else. Openai is working on Android apps and will soon be rolled out to more countries. The API could allow other developers to experiment with the model.
[Source]