OpenAI's Sora AI model can also make impressive video collages
What's the story
OpenAI recently unveiled Sora, a cutting-edge AI model that turns written prompts into eye-catching videos up to a minute long. Sora's remarkable video quality distinguishes it from other text-to-video generators, and experts believe it could revolutionize video creation. OpenAI employee Bill Peebles has now demonstrated Sora's ability to create video collages. The text-to-video generator produces multiple videos simultaneously crafting them into a collage form, as shown through a social media post.
Twitter Post
Take a look at Peebles's post
Sora can generate multiple videos side-by-side simultaneously.
— Bill Peebles (@billpeeb) February 17, 2024
This is a single video sample from Sora. We didn't stitch this together; Sora decided it wanted to have five different viewpoints all at once! pic.twitter.com/hGae80tM5a
Highlights
A mosaic of viewpoints within one composition
Peebles expressed his enthusiasm for Sora, stating, "Sora can generate multiple videos side-by-side simultaneously." "This is a single video sample from Sora. We didn't stitch this together; Sora decided it wanted to have five different viewpoints all at once!," he added. Certainly, this feature is quite impressive as it allows for a variety of perspectives in one composition.
Better than other models
Decoding Sora and its relation to other generative AI models
Sora joins other notable AI models like OpenAI's ChatGPT, DALL-E, and Midjourney. It uses similar technology to generate original content, such as creating videos from still images. Unlike others, though, Sora tackles the more complex task of AI-generated videos. OpenAI has not yet made the tool available to the public, as they are currently working with policymakers and artists before launching the official product. However, the company has provided a peek into its capabilities through a collection of video samples.
Insights
Community engagement and limitations
OpenAI CEO Sam Altman recently engaged with the online community on X by asking for prompt ideas and sharing realistic videos made by Sora. While Sora's capabilities are impressive, OpenAI recognizes some limitations, such as spatial and cause-and-effect elements. The text-to-video generator model also has some serious flaws, which may pose challenges in video generation.