When I opened Sora this morning, I was met with a flood of strange and disturbing AI-generated videos. On OpenAI’s video app, I ran through fictionalized scenarios of the Iran war and a series of fake Donald Trumps talking about Jeffrey Epstein. In my least favorite clip, I watched a man fry a baby. The app lets users create realistic-looking AI-generated clips—including their own faces—and then post them to a TikTok-like feed. It is not all of them they’re very annoying, and for better or worse, Sora has become an ample source of Internet viruses. Within days of its release, it jumped to the top of the App Store.
Now Sora will die soon. Yesterday, OpenAI said it was shutting down the program and ending public access to its video-making technology. The decision seemed sudden: Just a few months ago, Disney announced plans to invest $1 billion in OpenAI as part of a licensing deal to bring its characters to Sora, and earlier this week, employees from both companies were. it seems they still cooperate. (Disney has since rescinded its investment plans.) Even some of Sora’s own employees were reportedly caught off guard by the announcement. Online, people praised Sora for posting their favorite videos—like the one featuring a a row of penguins circling and another which Jesus walks on water to win an Olympic gold medal in swimming.
After OpenAI launched the Sora program in September, Sam Altman predicted that the community was about to make an incredible artistic revolution. “Innovation may be imminent through the Cambrian explosion,” he wrote online. But such a revolution has never happened. It’s not that people hate AI creep. In fact, if anything, people seem surprisingly curious — the latest TikTok trend is. mischievous soap games AI-derived fruit star. In response to a request for comment, an OpenAI spokesperson pointed me to a public statement that cites “compute demand” as a key factor in the company’s decision. Producing video is more expensive than producing text, and Sora may have been real money run: In autumn, Forbes estimated that Sora could be costing OpenAI millions of dollars every day, and Bill Peebles, who leads Sora, he said that the economy was completely unsustainable. (OpenAI declined to comment Forbesestimate at that time.)
The decision to move a project quickly and then suddenly pull the plug has become a common move for OpenAI. The company has spent the last few years cycling through new product features and business models at an impressive pace in an attempt to find a way to turn a profit. OpenAI seems to be finally learning that slope is not a business strategy.
Altman never had a good plan for how OpenAI would make money. “We don’t know how one day we can earn income,” Altman said at the 2019 event. He went on to explain that one day, AI will be intelligent enough that OpenAI will simply ask the computer how to generate a return on investment. “You can laugh,” he told the (justifiably) amused audience. “But it’s what I believe will happen.” After the success of ChatGPT a few years later, investors started pouring money into OpenAI, and Altman has done a great job of organizing investors’ money. The initiation is now value more than Toyota, Coca-Cola, and Disney together. But investors like to see revenue, and so far, OpenAI hasn’t done much to prove it can generate enough money to avoid bankruptcy.
That’s not to say it’s never tried: Over the past few years, OpenAI has explored almost every possible business model. Last season, Altman explained OpenAI as four separate companies—a consumer technology business, a large infrastructure project, an AI research lab, and an incubator for “new things,” including hardware. (OpenAI has a corporate partnership with Atlantic.)
The problem with trying to do everything is that sometimes you end up doing nothing well. Sora is the latest victim in a long line of sudden changes, about-faces, and seemingly sloppily executed projects. Last year, Altman announced a massive joint construction of AI infrastructure with Oracle and SoftBank called Stargate, but the effort it’s stuckit is reported following poor leadership and coordination. Altman he said in 2024 that combining ads with AI will be the “ultimate” answer – but then, earlier this year, the startup launched an ad plan. Last fall, OpenAI debuted a shopping feature, which allowed people to buy products directly within ChatGPT; yesterday, the company announced that it is killing the feature and focusing on product discovery instead. In January, the company he said that its long-awaited first device was “on track” to launch later this year, but weeks later, court cases. exposed that the company is unlikely to introduce its new hardware before 2027. OpenAI initially banned NSFW content, and then announced last year that it would make a difference to such material, even planning a December publication for erotica, and later putting erotica on hold indefinitely.
A certain amount of change in business plans is only natural for any company, let alone one in an industry as fast-moving as AI. But compared to its peers, OpenAI is particularly chaotic in its strategy. The company’s plans always seem tentative: No partnership or product roadmap feels guaranteed to endure. Earlier this year, Nvidia backtracked on a promise to invest up to $100 billion in OpenAI. At that time, The Wall Street Journal information that Nvidia CEO Jensen Huang was concerned with OpenAI’s “lack of discipline” in its business approach. (When asked about the report, Huang said it was “absurd” to suggest he was unhappy with OpenAI.)
OpenAI’s haphazard business strategy has left the company facing an identity crisis of its own making. OpenAI is losing ground to Anthropic, its main rival in the AI race, which has stuck to a focused approach to selling productivity-enhancing AI tools to other businesses. Anthropic has been very successful in its consistent approach to the business market. Now OpenAI is trying to copy Anthropic’s playbook. “We can’t miss this moment because we’re annoyed by the argument,” Fidji Simo, OpenAI’s head of applications, is reported he told employees at a company-wide meeting earlier this month, explaining that the company needs to nail “productivity ahead of business.” To do this, OpenAI is planning around twice its main calculation this year, including hiring a team of experts who will help other companies adopt its technology. Even at the product level, OpenAI appears to be copying Anthropic—OpenAI is apparently planning to launch a “superapp” to simplify the delivery of its products in a single app, possibly an attempt to compete with Anthropic Cowork and Claude Code. “We were spreading our efforts across too many programs,” Simo he wrote for workers last week. “That division has been slowing us down and making it difficult to achieve the level of quality we want.”
After scrolling through the fake Iran and Trump data trolling on Sora this morning, I came across Altman’s account on the forum. I was curious to see what the CEO of the company had to say about Sora’s ending. The last time Altman appears to have posted on the app was six months ago, when it launched. Maybe that should be an ominous sign. I kept watching more clips until the pop-up filled my screen. OpenAI wanted to know how using Sora was affecting my condition. The app gave me a choice between “Pipe-Up” and “Pipe-Down.” I called “Dole-Dole.”
