
About this episode
Get every episode summarized
Each time The Journal. publishes, we email you a written briefing from the transcript — the topics, who appeared, and any specific claims, with the ad reads skipped.
Email me new episodesFree for 3 shows. No card needed.
Hosts & guests
Transcript ready
268 searchable segments. Every word is indexed and playable.
Full transcript
The Journal. — More Coding, Less Slop? Why OpenAI Ditched Sora. Machine-transcribed; use the interactive transcript above to jump the player to any line.
Hey everyone, it's Jess. And Ryan, do you have plans on April 28th? If so, you should cancel them and come to our next journal live event, which is taking place in Los Angeles on April 28th. And we all know that LA is great in April. So come on out, it's gonna be at the L-Ray Theater. Jess and I are gonna interview some really interesting people. We're gonna have a lot of fun on stage and then we'll stick around afterward to hang out and chat with all of you. Tickets are still available, but not for long. So, grab yours now via the link in our show notes. So, grab yours now via the link in our show notes. If you're plugged into the AI zeitgeist, you've probably heard of Sora. So Sora is named after the Japanese word for sky. And it's open AI's AI model that is able to generate high quality video from text. Our colleague, Berbergen, covers the AI industry. He distinctly remembers when the public first got a glimpse of Sora back in 2024.
Is this the future of content creation? Meet open AI Sora to the next moment. So it was kind of a magical and also a bit of a scary moment. At the time, it was the first really sophisticated video generation tool that a company had previewed to the world. With just a few words, you can create hyper realistic videos of literally anything you can imagine. And it really just kind of showed the power of the technology. That video is not real. It's completely AI. If that watermark wasn't there, I genuinely would have thought this was real. I remember open AI posted these actually really beautiful clips. There's one that always sticks in my mind, which is like a pack of woolly mammoths, like galloping through a snowy field, because I used to love watching Ice Age as a kid. So it was one of those wild moments I had covering AI, where I really just realized that this was more than just about building a chatbot.
People were so impressed with Sora's videos that many wondered if it would change the entertainment industry forever. All the conversations about whether we'd be watching AI-generated movies and there'd be no more actors and actresses and like no more screenwriters, like that was because of Sora. But then, just a few weeks ago, open AI did a 180 on Sora. We got some breaking news out of the business world. Open AI, saying it's shutting down its video generation platform, Sora. You see the poster on Sora? This was a shock, I think, for everyone. Sora felt really integral to Sam Altman's vision for the company to kind of branch into the realms of popular culture and entertainment. And so when they shut it down, it was this very abrupt, sudden, unpredictable moment that I think also kind of speaks to just how quickly Open AI's identity is changing as a company.
Welcome to The Journal, our show about money, business, and power. I'm Jessica Mendoza. It's Tuesday, April 7th. Coming up on the show, the end of Sora and what it says about the future of the AI industry. For Open AI CEO, Sam Altman, products like Sora were key to his vision for the technology. A lot of Silicon Valley companies were built around, like making HR workflows easier or building software that can make it easier for you to do note-taking or the slacks of the world communicate with colleagues. That was never really his belief in what Silicon Valley could be.
And Sora was part of this broader vision that he had to kind of take these more ambitious moonshot bets with the technology. When Sora first launched, it was only available to paid chat GPT subscribers. But in September of last year, Open AI launched a Sora app with a social media feed where people could make in post videos. That's where some of the more like AI slot videos came in where people were splicing their faces into different scenes and kind of using it for more silly, goofy videos. I met Bigfoot or whatever type of stuff. And then the one that was kind of not great for Open AI was a lot of users started using Martin Luther King Jr.'s likeness. I have a dream that Xbox Game Pass Ultimate was still $20. And having him do kind of very silly things. I have a dream that Sam Oldman saves humanity instead of making AI slot.
Which was a bit of a PR nightmare for Open AI. King's estate complained about the videos, which led the platform to announce that it had removed his likeness. Did the company expect to make money from Sora? So the bet that they were making was that they really wanted this thing to go viral and then figure out how to chart it into a money-making business afterwards. And Sora was generating a lot of hype, so much so that even Disney wanted in. The entertainment giant announced it would partner with Open AI to license hundreds of Disney characters so people could generate videos with them. For Disney, I think it was an important moment for them because they felt like AI was going to change the way that entertainment companies operated and changed their creative processes. And for Disney, the deal with Open AI was a landmark moment that kind of showed that there was a future where they could coexist and work with an AI company
in a way that they felt respected their own IP. Was any other AI company doing anything like Sora at that point? Google has invested a lot of money into video generation and there are some startups that are also working on video generation technology. But very interestingly and throbic, Open AI's biggest startup rival avoided going into video generation, particularly because it just costs a lot of money. It costs more money to train and run video models and it does to run the language models that power chat bots. And so Open AI was really the face of it. So Open AI decides it's going to double down on Sora. It's rival and throbic decided to go in a different direction. Can you say more about that? Yes, so and throbic, obviously, this startup rival for Open AI that's putting them on their back heels. They are really focused on building AI tools for businesses
that are looking to integrate AI into their workforce. And throbic's approach was more focused on a gentick AI, the kind of AI that's like an assistant. It does the boring stuff for you, like scheduling emails, developing software or analyzing data. How do you think about the contrast between these two companies, Open AI and Anthropic? Like would it be fair to say that in this AI race, they define themselves in opposition to each other? They very much do. We've reported the very ugly history between Sam Altman and Dario Amade, who is the CEO of Anthropic. The most high profile rivalry in tech is heating up in 2026 as both Open AI and Anthropic race ahead. Anthropic released a series of ads that directly mocked chatubiti two of AI's biggest rivals Anthropics Dario Amade and Open AI Sam Altman are going viral after a public snub. So that's Open AI's Sam Altman and Anthropics Dario Amade, not holding hands during a group phone.
And I mean, in Silicon Valley, it just depends on who you talk to. Like, it's almost like there are two different camps in terms of who people are kind of sympathetic or allied to. Berber says the distinction really came into focus when Open AI launched the Sora app. That same month, Anthropic came out with a new version of its coding model. And starting last fall, they kind of took Silicon Valley by storm with the release of a new version of Cloud Code, which is their AI coding agent. If you talk to any software engineer in Silicon Valley, they will rave about Cloud Code, about how they don't have to work anymore, how AI's becoming so much smarter than them, how they think their jobs are going to be obsolete. And Anthropic really was driving that conversation. The two rivals are both planning to go public, potentially even this year. And Berber says, going into 2026, the huge success of Anthropics Cloud Code sparked panic inside Open AI.
They've been very much in code red mode for the past few months because they were seeing that they were losing the enterprise race. And they realized that their models were just not as good as Anthropics when it came to coding. And as a result, they just began taking a much more critical look at all the other projects that they had previously greenlit at Open AI. So how did Altman's vision for AI video end up on the chopping block? That's next. Open AI has a reputation for taking big swings. Altman has described the company's approach as startups within a startup. There's the team that's working on the new hardware device, there was a Sora team,
there was a team working on like an AI powered browser. And it kind of made sense because like, you know, AI is very unpredictable and like they wanted to empower their researchers to be able to spin up different techniques from proving AI. But all of those experiments and projects use a processing power. And the company only has so much of it. All the labs are basically rationing chips. Like every team inside Open AI is like begging for more computing resources. And what employees who are there at the time would say is that the way the AI chips were allocated between teams was very unpredictable. Like you could wake up one week and it would become very clear that like some new team had received a massive amount of computing resources. Sora was one of those teams because video generation uses a lot of computing resources. When we talk about it uses a lot of compute. Like do we have kind of like X number versus like a chat GPT?
We don't and the company doesn't disclose it. But what I can say is that if you talk to researchers at Open AI, I mean a lot of them were a little bit surprised by the amount of computing resources that the Sora team got. So Sora was this really interesting case study for that culture because the team worked separately from the research division. A lot of other opening employees didn't really know what Sora was up to. Berber says this sort of compute power could be justified for an AI model or app that was bringing in money or new users. But Sora wasn't really doing either. Sora had a very brief moment in the sun. The week after it was released it rocketed to the top of the app store and stayed there for a short while after but then it immediately plummeted and kind of flat lined out and around a little bit under 500,000 users a day. From a high of over a million and I think Open AI has realized that
recreating that chat GPT moment has been a lot harder than they expected. Open AI was hoping that people would love using the app, refer their friends. It would kind of grain itself in the pop culture and that just never really happened. This was all especially problematic with Open AI tightening its purse strings ahead of that potential IPO later this year. And so Sora's fate was sealed. The Sora team was about to start another very expensive training run to basically build the next more sophisticated version of Sora and then Open AI executives saw how much it would cost and we're just like we can't really afford to make this trade off and so let's just kill the whole thing. And that's the thing that was really surprising because it wasn't a huge business for them but just because it was so tight to the company's identity like I think it would have been hard to predict them getting rid of video generation altogether.
After the announcement Altman wrote a message to employees explaining the decision. He said the move was a necessary sacrifice for the good of the company a quote difficult trade off. And it's interesting I mean he very much seemed to realize just this new moment that Open AI was in where ahead of an IPO with tremendous competitive pressure they had to kind of ditch a vision for the company that he really believed in. Was it kind of an alpha Altman? And yeah it was an alpha Altman. Because I mean he was really personally invested in Sora. What about Open AI's big partnership with Disney? When Open AI pulled the plug on Sora what did mean for that deal? That whole vision for Sora it was kind of whiplash for them right because that's what they thought would be the future of this partnership and that just went away overnight. And I think for Disney they just have to they're kind of back to square one in
terms of thinking about how they're going to live in this AI world. Now that their main partner in Silicon Valley has basically done a 180 and pulled back. So if Open AI was sort of on the cutting edge of this and now they're pulling out, what does that mean for AI generated video? It's not like the whole world of AI video generation tools is dead, but it definitely kind of reshuffles the playing field a bit. And it's just like raising these questions again of like how soon is this technology coming? Who now is going to be bringing this technology to us if not Open AI? And like what's our strategy now for making sure that we are prepared for that world? Meanwhile, Berber says Open AI is now all hands on deck building a super app. It would combine chat GPT with an AI browser and the company's coding tool codex. It's interesting though it sounds like with this pivot Open AI and Anthropic are going to be
focusing on kind of the same things which is clarifying their core business models. So are these two rivals essentially just going to be walking in the same direction, doubling down on a very similar suite of products? The short answer is yes, the money now is in these agentic tools specifically agentic coding and that's a huge focus area for them. Chat GPT is still a much larger business than the consumer version of cloud. And so that's the advantage that Open AI has but right now they are both very actively battling for the same set of customers and I think compared to a year ago, they were obviously still competing but they were focused on different areas of the AI market. But Open AI is still trying to find ways to differentiate itself from its rivals. Last week the company announced that it's purchasing a popular video podcast about Silicon Valley which executives hope will help them shape the narrative around the company.
They've also said that the show will remain editorial independent. And investors seem excited about the direction Open AI is going. The startup recently closed a historic funding round of $122 billion. So Open AI is going strong. With that said, what do you think this moment means for the company? At the level of Open AI's identity, I think he kind of represents the death of a vision that Sam Altman had for Open AI that was kind of about using AI not just as like a basic productivity tool but to change the way people interacted with technology and specifically AI is an enabler of creativity. Are we just generally having to reorient on the promise that we the consumer have been told about how amazing and fun this tech is going to be and actually we're just going to be seeing it at work all the time?
Yeah, I think it's definitely like this weird moment where the kind of vision that people have of AI is actually a little bit different than how these companies want us to use these tools. I think being able to create a bot that's able to monitor your inbox and schedule appointments for you is very different than being able to conjure up like a fantastical world of like woolly mammoths trekking across the snowy meadow. That's all for today, Tuesday, April 7th. The journal is a co-production of Spotify and the Wall Street Journal. Additional reporting in this episode from Keach Aegee and Jessica Tuchel. Thanks for listening, see you tomorrow.
More episodes
More from The Journal.

My Monday Morning: Twiggy
The Journal.

The MAHA Shake-Up at Steak ‘n Shake
The Journal.

What Happened to the Affordable Car?
The Journal.

A Data Center Offered $26 Million for Her Land. She Said No.
The Journal.
