OpenAI припиняє Sora: стратегічний поворот для інвесторів та креативних команд

[screaming] Have you all heard the news? Open AAI is deleting Sora, eviscerating, snapping it out of existence like Thanos. We got anti- AI rage baiters out there celebrating in the digital streets claiming that the AI bubble is popping. But listen, plenty of people who are actually into AI and use it every day think that this is a great thing as well. Not because they see a bubble popping, but because they know the actual reason Open AI is deleting and cancelling Sora. They understand this is all about compute and open AI realizing they need it to bring on the future of AI technology where it's the most important and the most profitable. But before we dive into any future implications, let's talk about why this is happening, what the actual situation is, and some of the lessertold overlooked effects. Here is the official announcement from the official Sora app page. As we explore this topic today, something to pay attention to is how serious Open AI was about the future of Sora. I find it hard to believe they didn't want to keep this running for longer. Anyways, the tweet reads, "We're saying goodbye to the Sora app. It's flying high in the sky. It's going out to the back of the barn. It's going to be in a better place soon." Added after the fact 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. Feels like a typical canned response because people are emotionally upset about this. I was because I actually used the Sora app quite frequently, weekly at least. It was one of my favorite places to get creative with AI. That is not a lie. But I am rational enough to prevent this from clouding my judgment. I think internally OpenAI is very thankful that people used Sora for sure. I'm sure many members of the dev team had an amazing time with this project, but the training data that came from Sora is extremely valuable to OpenAI and is probably the reason they internally justified the cost of having Sora exist. It's extremely expensive. We'll get into how much it likely cost them, but they weren't making money. That training data though, that is good stuff. So, they're going to share more soon, including timelines for the app, which is going to disappear, both Android and iOS, and the API, which appears to also be disappearing, likely later. Also, details on preserving your work. I hope that they just give us like a big download button on an email to download all my Sora videos. Uh, for those with cameos, if I could download all of the cameos that people have made of me, that would be interesting. Although, there are thousands. And you know, that's something to think about because the talented few, some of us have built real followings on this app. I'm not one of them, but I did obtain 2,000 followers and people love to make cameos with me. Now, some of these cameos, in fact, a lot of them are super creative and a ton of fun. But that is not the vast majority of them. The vast majority are simply ads. I'm selling all kinds of stuff, guys, in all kinds of different scenarios. Guys, I even get to hang out with the dogs. The point is that the signs are on the walls. Fun creative cameos of you, your friends, and your family on a social media app just isn't what makes money. Once the initial Sora novelty wore off for most, putting yourself in a couple of silly situations, seeing ridiculous, unbelievable photorealistic videos that have this like punchy Tik Tok effect, the shine just died for the average person. Monthly app downloads lessened. Monthly users did retain pretty well, but there weren't that many, and it was still costing them a fortune. Clearly, not enough to justify further investment. The expense of the technology really was the Achilles heel for Sora. At the end of the day, like I said, I never stopped posting on here. Sora 2 is just my jam. This is my type of app, my type of experience. I love getting creative, bringing impossible ideas to life, just messing around with whatever pops into my head, putting weird twists on things. I also love remix culture, so it is at the end of the day something I am sad to see go, but I have no doubt it will eventually basically be replicated and replaced. Sora did indeed have a lot of active users, people who made original content that was different and would catch attention, make people remix it. Here's an example. Rorow, also a prominent user in my Discord server, which you should join by the way. 35,000 followers here on Sora. Almost half a million likes, and most of this is just genuinely original ideas and content thought up and put into a prompt generated with Sora. This is the closest thing to a Sora YouTuber, let's say, or content creator. So, it sucks that all those followers and everything Rorow built is going to disappear. You have to stop and think about the message that this sends from OpenAI to maybe its most potent core user base of consumers. The people that test all the latest technology, find every flaw with it. The people that make the best original content for your experimental Sora platform. If you can a new AI powered app only months after it's released, people who spend their time actually building something with that platform, something of real value are going to be less inclined to do so on your next experiment or venture. In some ways, this is a bad look for OpenAI. Some of their core user base is going to be hurt. And in a time where OpenAI is already struggling with public appearance, I don't think this really helps very much. But what it is is a clearer line in the sand. I think OpenAI says they want to focus more on enterprise businessto business solutions. They're looking for where the money is. They realize they need to become actually profitable. And while their products absolutely do provide upfront value, the price to make the better versions of the future is extremely high, ventures like Sora, they just can't exist right now for for the company, right? That is clearly the case. Do I think OpenAI is making the decision that is best for the future of their organization? Yes, I definitely believe that they were ballparking a loss of about $15 million a day on Sora. That's insanely high for a social media app with so few users. But we know why AI video technology is expensive. So my question would be why did they bother going through with making this app in the first place? Well, like I said in the beginning, obviously it's to train the next iteration of their models. Even that that video data, that training data from the Sora app could be valuable for chat GPT. All of this data connects across domains and they still may very well pursue video generation models in the future. There is a very healthy demand for it. But maybe OpenAI is going to go back to the drawing board and try and figure out an architecture that works better, doesn't cost so much, and produces great videos. I think when Sora initially released, they should have marketed it as an experiment. Google has its AI labs, literally even called experiments. Sora was marketed as a new social media app that's going to change the game. It wants to sit next to Instagram and Tik Tok. And maybe that was OpenAI's goal, but I just don't see how they realistically, especially considering how much it costs them, could have thought the app was going to grow to that size, especially when it took them so long to figure out the restrictions, preventing copyright while also giving users creative control. Pair that with inconsistent daily generation allowances and then also building in new features. I mean, look at this, guys. There's still a preferences beta section. this app is now announced to be closing and they still have a beta feature on it, like it's getting updates. This all adds up to basically them kind of just coming across like they didn't know what they were doing and that this was just winging it. The truth is that we're all just winging it at the end of the day. No one really knows what's going on or has full control over any situation. If you believe that, you're lying to yourself. But I don't know, OpenAI, I think you could have done a better job, pre-planned things a little bit more and hid your inconsistencies a bit better and maybe come across as a little bit more professional. Google's been at it for years. They've already learned this lesson. So, the big question that everyone's trying to answer is, what's the future going to bring? A lot of the AI haters, you could call them, say that this is the beginning of the popping of an AI bubble. First, you have to prove that there is a financial bubble. And there is actually plenty of proof of that. But there's no way to definitively claim there is a bubble until I suppose it's popped. And if an AI bubble does exist financially, then it's popping is really just a deflation to where it should actually sit monetarily. Just like with.com, people still use the internet websites. They're very important, very valuable. But just because we know a technology is going to be important doesn't mean we dump infinite amounts of money into it. And OpenAI isn't deleting Sora because it lacks any potential or that everybody hates AI. No, it's because they need the compute for other things that are actually going to make them money. The Sora app and its concept does have plenty of potential, but the technology needs to get cheaper for the output quality you get, and it needs a more dedicated investment that's less willy-nilly. Let's dive right in. Try to make something on the spot and more pre-planned, probably brought in by a Tik Tok or a Meta, something you're going to see made by a company that already has been doing social media and has very, very deep pockets and resources. OpenAI wants to be able to jump around and be fluid with its compute to wherever it feels is necessary to bring out new state-of-the-art tech. So, what is that state-of-the-art tech? Well, apparently the next architecture or the next model is cenamed Spud. Taking a look at this tweet by Tour, you can see there's clearly a shift of focus happening at OpenAI. Sam focusing on raising money, supply chains, building data centers at a massive scale, safety and security, changing places, and Spud expected to be very strong and a model that can accelerate the economy. Tibbor also notes that it shelved plans to bring video features into chat GPT, but I think based on the reaction of the Sora user base, they're going to have to bring some sort of Sora capabilities into chat GPT, although we don't know to what degree. Some ways, it feels like OpenAI is now playing catchup with Sora research shifting to long-term world simulation. Again, real world use cases that are going to bring in money like robotics. Sam noting things are moving faster than expected, but he always says that the truth is that this might reveal OpenAI was set back a little bit by bringing Soore into existence as an app. Gemini 3.0, Opus 4.5, those initial announcements kind of embarrassed Open AAI and its GPT5, which was pretty lackluster, especially on launch, and is now just improving to a level that I think is really awesome. Like, I love the new GPT 5.4 for thinking model, but it's taken them a while to get back on track. Here's my final thoughts. Overall, this is the smart move. It's the smart play, but it comes at a cost, and they kind of screwed themselves over by committing so heavily to the Sora app in the first place and leading users in with that. It's like giving us a shiny new toy and then just yanking it away. Maybe this was OpenAI's lost teenage years at the end of last year leading into the beginning of this year and they finally find their place and they've drawn their line in the sand like let's focus on enterprise. Let's focus on businessto business still produce like some light consumer experiments because we need the training data but mainly we got to focus on bringing money in because it's costing a ton of money to make leading edge AI. competition is fiery and it is not just in the United States. Sora, you will be missed as an app for me creatively bringing back that remix culture that I haven't felt since Flipnote Hatenna on the Nintendo DSi. Shout out to those that remember that. But like I said, I mean, we still have the internet, we still have tools, we still have creativity and things happening in the AI space. So, be on the lookout. I've got plenty to talk about this week. Thanks for watching and I'll see you in the next one. Subscribe and goodbye.