The Coolest (And Creepiest) Video Demos From OpenAI's Sora
Not quite content with changing our lives forever with its AI-powered chatbot, the folks at OpenAI are at it again. A little over a year after ChatGPT changed the way we search, research, and create (text) content on the web, the company has come up with its first-ever text-to-video generator called Sora.
OpenAI's text-to-video generator does exactly what it says: it generates a unique video (currently limited to less than a minute in duration) based on a user's text-based prompts. So, all you need to do is to enter a prompt into a text box — just like you currently do with ChatGPT — except this time, the goal is to prompt Sora to create a video that is the closest visual representation of what you had in mind.
To demonstrate its capabilities, OpenAI has posted unedited videos generated by Sora on the company's website. These videos span multiple genres and have evoked mixed reactions from people.
While most people were blown away by the photorealistic nature of these videos — especially those that show people, natural landscapes, and urban settings — some were concerned about the many ways in which Sora could be misused. Listed below are some of the most impressive — as well as a few of the strangest — videos generated by OpenAI's Sora.
The coolest videos generated by OpenAI Sora
The first Sora-generated video to gain mass attention also happens to be the video embedded alongside the tweet introducing Sora. The video showcases a couple walking the streets of a snowy Tokyo city, as petals from roadside Sakura trees (also called the cherry Blossom) fall onto the pavement.
Introducing Sora, our text-to-video model.
Sora can create videos of up to 60 seconds featuring highly detailed scenes, complex camera motion, and multiple characters with vibrant emotions. https://t.co/7j2JN27M3W
Prompt: "Beautiful, snowy... pic.twitter.com/ruTEWn87vf
— OpenAI (@OpenAI) February 15, 2024
Another impressive video shows a herd of woolly mammoths making their way through a snowy meadow. The prompt for this video also specifically insists on keeping the camera angle low for a more dramatic effect. It also mentions that the lighting should be such that the video should look as if it were shot at noon.
Prompt: "Several giant wooly mammoths approach treading through a snowy meadow, their long wooly fur lightly blows in the wind as they walk, snow covered trees and dramatic snow capped mountains in the distance, mid afternoon light with wispy clouds and a sun high in the distance... pic.twitter.com/Um5CWI18nS
— OpenAI (@OpenAI) February 15, 2024
Another video that has gained popularity also involves Japan and showcases what appears to be a stylish Japanese woman walking the impeccable streets of Tokyo. This, however, was also the first video that we found a few issues with. While the video centers attention around the woman, if you look carefully, many of the people that are also seen in the video have unusual gait and a rather strange way of walking, almost as if they're floating.
Prompt: "A stylish woman walks down a Tokyo street filled with warm glowing neon and animated city signage. she wears a black leather jacket, a long red dress, and black boots, and carries a black purse. she wears sunglasses and red lipstick. she walks confidently and casually.... pic.twitter.com/cjIdgYFaWq
— OpenAI (@OpenAI) February 15, 2024
Then again, given the fact that the woman and everyone else in the video do not really exist, what Sora has created is truly incredible — but it could feel borderline creepy for some.
More Sora-generated videos
While the dedicated page for OpenAI's Sora claims that all of the AI-generated videos were posted unedited, there was likely some process of curation involved in which videos got posted in the announcement. To demonstrate that Sora and its capabilities are real, Open AI chief Sam Altman took to X (formerly Twitter) to take prompt requests from people across the world and create videos based on them.
Some of the notable ones include the video that showcases a cooking lesson taught by a grandmother in a social media style video.
https://t.co/rmk9zI0oqO pic.twitter.com/WanFKOzdIw
— Sam Altman (@sama) February 15, 2024
Another light-hearted video shows two golden retrievers recording a podcast on a mountain.
https://t.co/uCuhUPv51N pic.twitter.com/nej4TIwgaP
— Sam Altman (@sama) February 15, 2024
Someone also asked Sam to use Sora and generate a video showcasing a futuristic drone race taking place on the surface of Mars as the sun sets on the horizon.
https://t.co/P26vJHlw06 pic.twitter.com/AW9TfYBu3b
— Sam Altman (@sama) February 15, 2024
Then there are some truly weird ones, the first one showcasing a half-duck half-dragon flying through the air, with a hamster dressed in adventure gear on its back, while another features sea creatures riding bicycles on the surface of the ocean.
here is a better one: https://t.co/WJQCMEH9QG pic.twitter.com/oymtmHVmZN
— Sam Altman (@sama) February 15, 2024
https://t.co/qbj02M4ng8 pic.twitter.com/EvngqF2ZIX
— Sam Altman (@sama) February 15, 2024
If you're still wondering why there isn't a deluge of user-generated Sora videos on social media, it is because the tool is currently being tested by OpenAI and is not yet available for the general public to try.