OpenAI Kills Sora Video Model

0:00We have some breaking news that OpenAI is going to discontinue their video service, Sora, their video model. Bayer was a post from the official Sora account that said we're saying goodbye to the Sora app. Today on the podcast, I'm going to be breaking down everything happening with the discontinuation of this. Why this had so many people shocked and where we can expect AI and video to go forward in the future. And maybe what some of the early warning signs were for this to begin with. I'm sure some people are celebrating. Some people have other mixed opinions. We're breaking down everything that's been said on X about this as well. So let's get into it. But I wanted to say before we do, as you probably know, I recently added video to my own platform AIbox.ai for video generation. And we actually have the Sora model on there. So before is deleted before the API is gone forever. If you want to try out the Sora AI video model, this may be your last chance. You can go for 899 a month. You can go check out AIbox.ai get access to Sora. And over 70 different AI models, including everything from

1:01Grock to Gemini to chat GPT to all of the top AI image audio video models. And it's 899 a month. So there's a link in the description to AIbox.ai. If you want to go try it out, I'll leave it linked down there. All right, let's talk about what's happening with Sora. So the official Sora account, which is Sora official app over on X tweeted out a couple hours ago and said we're saying goodbye to the Sora app to everyone who created with Sora shared it and built community around it. Thank you. What you made with Sora mattered and we know this news is disappointing. We'll share more soon, including timelines for the app and the API and details on preserving your work, the Sora team, which is what's interesting there is I mean, they basically spelled out, you know, saying as far as like timelines, it's not like what the future holds for the app or the API. It's like the timeline AKA they're going to kill the Sora app. They're going to kill the API is what which I mean, personally, it's kind of annoying since I just recently got the API all linked in with my own startup, but we'll just have to make some plans for killing that out. The top comment on

2:02this comes from Curianio, who says this is insane. Do you all know what you're throwing away here? Are we going to open source it or at least I think honestly that would be an amazing future for this product. You know, open AI probably could do some, you know, they can make a lot of developers. Happy if they open source it, but I highly doubt they will because I'm sure part of their reason here is they're like, Oh, look, people are making AI Slop. I also think that they're the that's probably like the excuse that they'll use even though there's like a hundred other AI models. But the real reason why they're getting rid of it, I think, is probably a little bit more interesting. But before we get into that, I wanted to read a little excerpt from the Wall Street Journal, that was just reporting on what Sam Alton said. So this is a quote from Wall Street Journal, CEO Sam Alton announced changes to staff on Tuesday, writing that the company would wind down products that use its video models. In addition to the consumer app, opening I will also discontinue a version of SOAR for developers and won't support video functionality inside of chat GPT either. Open AI is in the middle of a strategy shift to redirect the company's computing resources and top talents towards so-called productivity tools

3:07that can be used by both enterprises and individual users last week. Open AI announced that it was combining its chat GPT desktop app, coding tool codex, and browser into one super app. The company expects to the consolidated product to align its employees around a single vision. Now, someone that personally actually has both the codex and I also have their Atlas app, and you know, chat GPT has a desktop app. So, you know, kind of combining all those does make a lot of sense to me. I'm not sure why you need multiple. I also have the clawed app which has basically all of those features in one place as well. So, I think as far as consolidating the app, that makes a lot of sense. Now, killing off one of their, you know, major AI models, which to be fair, had a lot of use. I still remember when this thing came out back in October, SOAR hit a million downloads faster than chat GPT did. And of course, that's kind of probably riding on the back of a lot of these viral clips that people are making and kind of a lot of the viral campaigns that they had built around this. So, I think, I think SOAR had about 627,000 iOS

4:15downloads in its first seven days compared to chat GP, which had like 606,000 in its first week. So, obviously, like, this was a very popular app, but it seemed to be something that was popular for a second, right? It's kind of a novelty. A lot of people tried it. Personally, I tried it, and I remember it was super hard to get access to SOAR when it first launched, because you had to have an invite code, and then once you got on, you could invite like 15 people, and maybe you got like 10 invite codes that you could share. Anyways, it was, you know, it was kind of a headache to get access to this, but once you did, personally, I made like a handful of videos, it was kind of funny, kind of a novelty, and I moved on, because I mean, basically the place that they made this the most accessible was, they did, I believe, have SOAR dot com, or they do, but the app was kind of the main place. And to be honest, I don't think a lot of people had a lot of interest in just using an app that was exclusively AI-generated. I mean, really just like AI-slop. There was a post that someone recently shared of the app store downloads for SOAR in the United States. Actually, I think it was Canada,

5:19Japan, and the United States. Total downloads was about 4.58 million. Average monthly downloads is about 1.5 million. But the graph of kind of the daily downloads on this since November to January, and I don't even have the full chart beyond January for the last couple of months, but it was kind of in November, somewhere around 100,000 a day dropping off significantly to under 25,000 a day as of January. So I don't think, I don't think it's, you know, something that's basically had a major drop off in popularity. As I think basically, people just tried it, they used it, and you know, it's kind of a fun little novelty, but beyond that, it wasn't super, super exciting. Now, one thing that someone posted, Trunest on X, I said, is this not extremely pointless, someone will replace SOAR, and it'll be 10 times more AI-slop. When they say someone will replace SOAR, I think a lot of people kind of have this, are kind of coming from this line of reasoning that like, oh, opening eyes, trying to do like, I don't know, like a noble thing to get rid of AI,

6:22so we don't have AI-slop videos at the end of the world, or at the end of the day, like Google has VO3, which is a really powerful video model, and to be honest, I've used it for a bunch of different projects where it's been pretty useful. Even when you look at companies like Adobe, they have video models embedded into Adobe, and I've used those as well, you know, I recently, a few months ago, I had filmed a music video, and for my wife, she's a Christian, just musician, and the first few frames of the video, I realized they were like pointing at something that I needed to cut out, and so anyways, it was kind of a hassle, but I basically was able to use their video expand tool to add a few extra seconds to the beginning of the clip, which made my whole thing work better. Anyways, I feel like there's just a lot of these like really useful, you know, just kind of expanding videos, or zooming in, zooming out, adding like little little effects, even if you wanted to be like, you know, super authentic and raw footage, sometimes those things are make a huge difference. So beyond just AI slot, these AI video models are super useful for a lot of

7:27different things, and so anyways, I don't think it's, you know, I don't think that's the reason why it's going away. I think here, there's a couple different reasons why OpenAI is killing it. Number one, I think compute, they're really trying to not get bottlenecked on compute. I think there's a lot of shortages on, you know, memory and a lot of other different hardware for some of these AI models coming up, and I think video is something that's super, super intensive. Now, it's super cool, and it's kind of fun, and I think it puts you on the cutting edge, and a lot of people like when you're working on it, like video stuff, but I think it is quite a big struggle. And so if they were really to scale, I think with a model, then I think that would be a struggle for them. Now, I think the real reason is probably just because they're trying to consolidate their focus. I think that Anthropics Cloud is kind of catching OpenAI by surprise. It's having a huge run up right now. It's honestly just the best text model out there. It's really good at reasoning, it's really good at coding, and I think a lot of business professionals are kind of starting to

8:28use it. I think OpenAI is realizing that the highest pain tier of, you know, the highest pain user is kind of moving towards something like Cloud, where it's really focused on a lot of reasoning and productivity tools. Every time Cloud comes out with a new feature, you know, it tanks the stock of half a dozen companies that are kind of in that area when they come out with like health care or finance, or just a lot of these different features. And even recently, Cloud just came out with their Cloud computer use, which is an incredible tool. They can take over your computer, and I've been using that all day long, testing that out. And so I think OpenAI realizes that if they're not kind of consolidating and focusing on their core product, they will get beat by Cloud. And so I think that's where they're really trying to put a lot of their focus, even if it means killing off such a massive tool like SOAR, which is pretty crazy to think because, you know, like anthropic, they only really have texts that they're really strong at, and code is kind of a part of that, and the reasoning is kind of a part of that. But Cloud doesn't even do image generation,

9:29they don't do video generation, they don't do audio generation, like they're literally just code and reasoning. And they're just getting such a huge part of the market. I think a lot of people are using that and OpenAI is concerned. So it's interesting to see OpenAI roll back from successful popular products to focus on that core, where they think there's a lot of basically threat, where they're being kind of threatened by an AI that is able to reason really well. So it's going to be interesting to see what happens here, especially considering the fact that there's also a lot of other video models out there. We have seed dance from China, which just kind of put on pause a lot of their releases because people could create copyrighted content. Now, one other thing that was interesting with this whole story is that OpenAI was actually in the process of signing a, I think, like, $200 million deal with Disney. They were working out the details where they were going to have Disney characters and kind of content embedded into the SOAR model. And with this all being shut down, it seems like that deal is probably going to be off the table. So there's just a lot of

10:30moving parts that are happening here and a lot of consolidation that's happening in OpenAI. I'll keep you guys up to date on everything happening with SOAR and all of this. But again, if you want to try SOAR for the last time before it is deleted and gone forever, go check out aibox.ai. You can use SOAR and all the other video models, tons of other models, which was over 70 on there. And we'd love to see what you guys make. All right, have a fantastic rest of your day, and I'll catch you in the next episode.