This is Atlantic Intelligence, a newsletter in which our writers help you wrap your mind around artificial intelligence and a new machine age. Sign up here.
Earlier this week, OpenAI launched the full version of its video-generating model, Sora. Hype has been building around this release since the startup teased the program nearly 10 months ago, but the final product doesn’t quite meet expectations.
OpenAI researchers said they had spent months making a “way faster and cheaper” version of Sora that the public could use—but did not say this version of Sora is more capable or intelligent. The company was eager to show off a number of features, such as “Remix,” “Loop,” and “Blend,” that might make Sora a legitimately useful short-video editor and generator but don’t suggest much about how this product serves the company’s ultimate goal of bringing about a supposed superintelligence. Indeed, my own tests of the model have been mixed, resulting in floating glasses of eggnog, vanishing cat heads, and Silly Putty–like arms. “The company hasn’t built a new, more intelligent bot so much as an interface in the style of iMovie and Premiere Pro,” I wrote after OpenAI announced Sora’s release on Monday.
This is all a far cry from the rhetoric of the initial Sora preview, in which OpenAI presented the program as a crucial avenue toward building smarter and more powerful bots. In May, I spoke with a Sora researcher who described the program as being in its “GPT-1” phase (in other words, it should be viewed as extremely early, conceptual research), and the company repeated the analogy in its presentation this week. It is worth keeping in mind, then, that if GPT-1 had launched as a product in 2018, it would have been very cool and not very practical, in the same way Sora is now. Of course, anyone who might have written OpenAI off then would have been very surprised by ChatGPT’s success just four years after that; such a moment is far from guaranteed to arrive for Sora and its video-generating successors, but I wouldn’t bet against it, either.
The Most Hyped Bot Since ChatGPT
By Matteo Wong
For more than two years, every new AI announcement has lived in the shadow of ChatGPT. No model from any company has eclipsed or matched that initial fever. But perhaps the closest any firm has come to replicating the buzz was this past February, when OpenAI first teased its video-generating AI model, Sora. Tantalizing clips—woolly mammoths kicking up clouds of snow, Pixar-esque animations of adorable fluffy critters—promised a stunning future, one in which anyone can whip up high-quality clips by typing simple text prompts into a computer program.
But Sora, which was not immediately available to the public, remained just that: a teaser. Pressure on OpenAI has mounted. In the intervening months, several other major tech companies, including Meta, Google, and Amazon, have showcased video-generating models of their own. Today, OpenAI finally responded. “This is a launch we’ve been excited for for a long time,” the start-up’s CEO, Sam Altman, said in an announcement video. “We’re going to launch Sora, our video product.”
What to Read Next
P.S.
The Silicon Valley hype cycle that immediately preceded generative AI was all about cryptocurrency, and while the many coins and tokens have all failed as functional currencies, their legacy as financial instruments is now clear. “Cryptocurrencies have minted a generation of millionaires, billionaires, and corporate war chests,” Charlie wrote on Wednesday. “And now they’re using their money to influence politics.”
— Matteo
