Well, we hadn’t heard about Sora in a while!
The AI video generation app from OpenAI initially shot to the top of app stores and immediately created IP headaches in Hollywood and the business world as people created Pixar-style characters and deepfakes of actors and public figures like Martin Luther King Jr.
Later came its big announcement with Disney to license its iconic characters, along with an $1 billion investment from the entertainment giant.
When Sora first came on the scene, it rattled many creator economy insiders who privately told us their worries about how it could upend creators’ businesses, from more competition for attention and revenue to the misuse of their likeness. Others were excited about the potential, especially with the Disney deal, to highlight more creators and user-generated content.
We were always skeptical, both of Sora’s staying power and OpenAI’s reasons for developing the app in the first place.
The end goal for Sora wasn’t necessarily to develop a social network that competed with the likes of TikTok or Instagram. (Though judging by the number of former Meta execs OpenAI has poached, it’s easy to jump to that conclusion. See our timeline above.)
In the short-term, Sora was a way for OpenAI to get people excited about AI by turning it into something fun, rather than scary. In the longer-term, Sora would have been a major source of first-party user data that OpenAI could use to train its models and build out its core business.
But that didn’t quite work out. The novelty wore off fast and by November, news headlines were calling the app “boring” as people complained that their friends weren’t there. Downloads of the app fell to 600,000 last month, down from 1 million in the first five days after its launch, according to data from mobile app tracking firm Sensor Tower.
That was to be expected. App download data isn’t a good indicator of long-term sustainability and it’s really hard to break into the social market. BeReal is a perfect example: It’s still around, but barely. Many people also don’t want—or need—an AI social app when there’s plenty of AI slop on the platforms they already use.
Running a social app also comes with a lot of headaches. Just ask Mark Zuckerberg: Earlier on Wednesday, Meta along with YouTube were found liable for designing their apps in ways that were harmful to children. (More on that below.)
And OpenAI already has plenty of headaches of its own. The company is working to unify its offerings, including ChatGPT, its coding platform Codex and browser into a single “superapp.”
Superapps don’t have a great track record outside of China and require a change in user behavior and a lot of trust, which OpenAI hasn’t built up. Sora is likely a casualty of those efforts.
But the end of Sora doesn’t mean the end of AI-generated social content or the challenges they bring for creators. (Hello, ChatGPT’s ad push.) Plus, the loss of the Disney deal shows just how precarious and tenuous the relationship still is between AI, creators and Hollywood. The news could make it harder for those types of deals to happen in the future.
In other OpenAI news…
OpenAI Poaches More Meta Execs
As we mapped out in the timeline above, at least five former Meta executives have moved to OpenAI over the past seven months.
What stood out to us is that OpenAI isn’t just recruiting AI researchers—it’s tapping talent from across Meta’s business, including ads and creator partnerships. These high-profile hires offer a window into OpenAI’s business priorities and future ambitions.
Two execs have joined during this month alone. David Dugan will lead ad sales, signaling that OpenAI is readying for a major commercial push. Meanwhile, Charles Porch, a 15-year veteran of Meta-owned Instagram who helped onboard many celebrities to the app, took on a similar partnerships role earlier this month at OpenAI. The hiring shows how the company is looking to forge more relationships in the entertainment industry. (Gulp, to the Disney news above, which is sure to complicate Porch’s pitch.)
One of the biggest announcements was OpenAI’s hiring of Fidji Simo, the former head of the Facebook app and CEO of Instacart. She officially joined in August as CEO of Applications, also pointing to OpenAI’s advertising ambitions. At Meta, she played a key role in building Facebook’s advertising business and scaling mobile ads.
In August, Instagram’s co-head of product Ashley Alexander moved to OpenAI to lead health products, an area where ChatGPT is already widely used (for better or for worse). In January, OpenAI also released a series of health-focused products for consumers and physicians and made a big ad push highlighting how people use ChatGPT for health and well-being, which included a spot featuring creator Gigi Robinson.
Science also seems to be a growing priority. About six months ago, Kevin Weil moved from chief product officer to a different role leading a new OpenAI for Science unit, which he described on X as an “AI-powered platform that accelerates scientific discovery.” Weil, who was one of the co-founders of Facebook’s Novi crypto division, left Meta in 2021 and joined OpenAI in 2024.
Obviously technical AI talent is still key. Amid an intensifying recruiting war among tech companies, OpenAI last month poached high-profile AI researcher Ruoming Pang, who had just joined Meta from Apple seven months earlier. At Meta, Pang oversaw AI infrastructure for the company’s AI division Superintelligence Labs.
If OpenAI is preparing for an IPO later this year, it’s going to need the right seasoned leaders in place. Some of these leaders were also around during Meta’s attempts to develop a “superapp” and could offer valuable insight into OpenAI’s effort to do the same.
We suspect these Meta execs won’t be the last to join OpenAI.
Meta and YouTube were found liable on Wednesday in a lawsuit alleging their app designs were addictive and harmful to young users.
As Jasmine told the AFP on Wednesday, the financial fines aren’t the problem: $3 million is a slap on the wrist for companies like Meta and YouTube, which are two of the biggest ad sellers in the world. In 2025, Meta made roughly $200 billion, which is over $22 million per hour.
What’s really at stake is whether social media companies will be forced to redesign their products. The endless scroll of content algorithmically tailored to a user’s interests and other features have helped these apps keep people’s attention.
A redesign poses an existential threat to their business models. Their ad businesses thrive off of engagement and if product overhauls change that, these platforms become much less valuable to advertisers.
This case could not only set a precedent for how similar cases play out (and there are thousands of consolidated cases against them), but also how social platforms do business. Because of their size, Meta and YouTube often set the pace for smaller rivals. We expect this case to have a ripple effect across the entire social landscape.
Before the verdict came in, we sat down with California Attorney General Rob Bonta to talk about the trial and whether he thinks youth social media bans could be a solution. Watch the full interview below to hear his response, or tune in on Spotify or anywhere you get your podcasts.
