Sora is the most popular bot since ChatGPT.

For more than two years, every new AI announcement has been in the shadow of ChatGPT. There were no models from any company. that can overshadow or be compared to the aforementioned initial fever But perhaps the closest the company came to imitating this rumor was last February, when OpenAI teased its first video-generating AI model, Sora, a tantalizing clip – of a woolly mammoth kicking up snow clouds. Pixar-style animations of cute furry animals – promises a fascinating future where anyone can create high-quality clips by typing simple prompts. into the computer program

But Sora, who was not immediately revealed to the public, Still just a teaser, pressure on OpenAI is increasing During the intervening months Several major tech companies, including Meta, Google, and Amazon, have showcased their own video creation models. Today, OpenAI finally responds. “This is a launch we’ve been excited about for a long time,” the startup’s CEO, Sam Altman, said in the announcement video. “We’re about to launch Sora, our video product.”

In the announcement, the company said that ChatGPT paid subscribers in the US and other countries will be able to use Sora to create their own videos. It’s different from the video creation model of other tech companies. Still in preview or only available through enterprise cloud platforms, Sora is the first video creation product delivered directly to users by a major tech company. Chatbots and image generators like OpenAI’s DALL-E make it easy for anyone to create and share detailed content in just seconds. It is threatening entire industries and causing profound changes in online communication. Today, the era of video-generating AI models will only make those changes deeper, faster, and more bizarre.

OpenAI’s keywords this afternoon are product– The company charges money from Sora, not to advance research. But it is the consumer experience. This is part of the company’s ongoing trade stagnation. Founded in 2015, OpenAI was a non-profit organization. Mission To create digital intelligence “For the benefit of humanity as a whole Unconstrained by the need to generate financial returns, it now publishes products and business deals like any other technology company. OpenAI added a for-profit segment in 2019, and as of September, it did. According to the report consideration Completely revoking control of the non-profit board, Sora’s marketing has changed since February, when OpenAI introduced its video creation model. It’s an important step towards the company’s lofty mission of creating technology that’s smarter than humans. Bill Peebles, one of Sora’s lead researchers, told me in May that the video would help “push the road to AGI,” or artificial general intelligence. It allows the company’s programs to simulate physics and even human thinking. To create a video of a soccer match, Sora may need to model both aerodynamics and player psychology.

Today’s announcement is preceded by a review By Marques Brownlee YouTuber famous for his reviews of gadgets like iPhones and virtual reality headsets. Altman wore a hoodie emblazoned with the words Sora– Altman and the Sora product team talked for over 17 minutes; Peebles and another researcher spoke for a minute and 45 seconds, mostly praising the company’s release of a “turbo” version of Sora that was “faster and cheaper” to launch a “new product experience.”

The Sora launch comes on day 3 of the “12 Days of OpenAI,” where new products are launched or demoed to users every day. What the company has announced has more in common with a product than a breakthrough in computer science: a modern interface for creating and editing video. With features such as “Remix,” “Loop,” and “Blend,” Sora’s work so far has been impressive as well as surprising. The company has yet to create a new bot that’s smarter than the interfaces in the style of iMovie and Premiere Pro.

Videos created by OpenAI staff and early access users with Sora are already videos. drop Oven Social media and feedback from users around the world follow– For more than two years, cheap and easy-to-use generative AI models have turned everyone into a potential illustrator. Soon everyone can become an animator, too. That’s a clear threat to human illustrators and animators. Many people have been warning for a long time that the AI ​​generation will destroy their way of life. Sora and similar projects have also raised fears of disinformation campaigns (Sora videos come with an image watermark). But with OpenAI’s highest tier subscription, which costs $200 per month. Customers can create clips without watermarks)

But job shifts and distorted data may not be the most immediate or important impacts of day three of OpenAI. Both happened without Sora, although the program helped accelerate each problem: Production studios were experimenting with the product. Enterprise-grade AI already exists to create videos, for example. latest Coca-Cola’s holiday ads and low-tech, cheap methods of creating and distributing disinformation have also been a huge success. by yourself–

Where the widespread adoption of video-generating AI products may significantly change the way people express themselves online. Over the past year, memes, cartoons, caricatures, and other AI-generated images, sometimes called “Nonsense” has saturated the internet. Much of this content is clearly AI-generated, rather than intended to deceive. which is a medium for expressing rudeness It’s not a complicated trick. Possibly the biggest impact of technology on the 2024 presidential election, the fact that anyone can create such images provides an immediate way to express their feelings about the world being penetrated through images. It’s easy to understand, as my colleague Charlie Warzel wrote. Such content is intended to be consumed “for a moment and requires little to no thought beyond the initial response of the limbic system.”

Massive AI-generated videos may provide a more effective way to communicate visual confusion. Accused feelings Or persuasive propaganda – perhaps a much more realistic version of the latest version created by a low-quality AI? Video of Donald Trump and Jill Biden in a fight, etc. Sora may take over TikTok and similar short-form video platforms. Just as AI visualization models have distorted Facebook and changed the way people express support for X for political candidates.

Sora’s web takeover isn’t guaranteed Back in May, another Sora researcher, Tim Brooks, has since joined Google. It compared the current state of the program with GPT-1, the earliest version of the program that supports ChatGPT, which is currently in its fourth generation. OpenAI repeats the comparison today. That comparison breaks down as companies become more and more profit-driven. GPT-1 is preliminary research at a high level. It’s a concept before proof of concept. And four years removed from the launch of ChatGPT, Sora may not be nearly as developed. There’s no opening for AGI, but it’s only becoming a full-fledged product after OpenAI teased the model for nearly 10 months. Solving the climate crisis Or how else are startups claiming that AI might benefit humanity as a whole? But it might be just what OpenAI needs to increase its profits.