Ohneis - The Pattern
Tech, AI, design, and the hidden logic behind it all.
Ohneis - The Pattern
Happy Horse Just Crushed The AI Video Race | with Laura
Use Left/Right to seek, Home/End to jump to start or end. Hold shift to jump forward or backward.
Technology moves fast, design makes it matter, AI changes everything. This is oh nice. Have you ever tried to shoot a promo video for your small business and realized halfway through that you needed a decent camera, good lighting, a person who doesn't look uncomfortable on screen, editing software, and about six hours you absolutely do not have. Yeah, most people give up before they even start. The promise of AI video generation is that one day you just type what you want and it appears. A golden retriever in a chef's hat baking a pizza in a sunny kitchen. 10 seconds, done, flawless. For a while it looked like American companies were going to hand us that magic button. But this week, the entire industry got completely sideswiped. A mysterious new tool nobody saw coming just topped every AI video leaderboard on the planet. It's called Happy Horse, and it was built by Ali Baba. Yes, the company most people associate with cheap phone cases and bulk orders of random stuff. I've got Laura here with me today, and we are going to dig into why OpenAI just blinked, why China is taking over the AI video race, and what happy horse actually means for the way content gets made going forward. Laura, let's start right at the beginning. When you first heard the name Happy Horse topping these leaderboards, what was your actual reaction?
SPEAKER_00Okay, so my first reaction was, I thought it was a joke. Like I genuinely thought someone had made up a fake tool name just to troll the AI discourse online. Happy horse. It sounds like a children's cartoon, not a piece of technology that's apparently beating everything OpenAI and Google have built. But then I started looking at the numbers and the benchmark results, and I was like, wait, this is real. And not just real, but like shockingly good. The videos coming out of it have this quality to them that's hard to describe unless you've seen it. Smooth motion, realistic lighting, objects that behave the way they're supposed to behave. And I think that's the thing that hit me hardest. Because generating video that looks convincing isn't just about pretty pictures, it's about the system actually understanding physics. How does a chef's jacket wrinkle when someone moves? How does steam rise off a pan? For a long time, AI video tools would get these things subtly wrong, and your brain would clock it immediately. That uncanny valley feeling. Happy Horse is getting those things right in a way that's making a lot of people very uncomfortable.
SPEAKER_01Right, and that's the technical leap that matters here, isn't it? Because this isn't just a story about one company beating another on a scoreboard. The leaderboard itself is worth explaining, because it's not like some panel of tech journalists sat down and voted. The way it works is more like a blind taste test. Think the old Pepsi challenge. You're shown two videos generated from the exact same prompt, side by side, and you pick the one that looks better. You have no idea which company made which, no branding, no hints, just the output. Happy Horse didn't win because of marketing or hype. It won because regular people looking at the actual product kept choosing it. And then the wild part, it didn't even have a public announcement, no big press release. It just kind of appeared at the top. Which raises an obvious question. Laura, Alibaba is not who anyone expected to pull this off. Like, what is actually going on inside that company right now?
SPEAKER_00So this is where it gets really interesting because Alibaba has been going through this massive internal overhaul that most people in the West completely missed. Their CEO has been publicly, explicitly saying that AI monetization is now the central priority of the entire company, not a side project, the engine. And to make that real, they restructured the whole business. New leadership, new teams specifically built around AI products and revenue. I think the analogy I keep coming back to is imagine taking a massive, successful shipping truck that's been running on diesel for decades, and while it's still moving down the highway, you rip the engine out and drop in a completely different one. That's the level of disruption they've done internally. And Happy Horse is basically the first major proof point that the new engine works. But here's what I think people are underestimating. Alibaba has something that a lot of pure AI labs don't have, which is a direct pipeline to commerce. They know how businesses actually use video. E-commerce product demos, marketing content, social media. They've been watching their sellers struggle with this for years. So they didn't build a cool research project, they built a tool.
SPEAKER_01That's a huge point. They're not just doing this to win a leaderboard, they have a commercial reason for every single feature.
SPEAKER_00Exactly. And that changes the incentive structure completely. When a pure research lab builds a video model, the goal is often to push what's technically possible. When a company whose entire ecosystem runs on sellers moving product builds a video model, the goal is make it fast, make it cheap, make it good enough that a small business owner with no production experience can use it today. Those are very different design philosophies. And I think that's partly why Happy Horse feels different when you use it.
SPEAKER_01Let's talk about the other side of the story. Because you cannot tell the Happy Horse story without talking about what OpenAI just did. Because the timing here is almost theatrical. Right as Alibaba is charging into the AI video space with everything they have, OpenAI, the company that was supposed to be the gold standard, appears to be pulling back. And look, OpenAI hasn't disappeared, but there are real signals that they're stepping away from the most expensive frontier of the video race. Laura, why would the company that is arguably the most famous AI lab in the world retreat right now?
SPEAKER_00I mean, this is the part of the story that doesn't get talked about enough. AI video generation is extraordinarily expensive to develop and to run. We're talking about compute costs that are almost incomprehensible. Training these models requires thousands of specialized chips running for months. And then every single time a user generates a video, that's more compute, more cost. OpenAI has been burning through money at a pace that even its investors find alarming. And video is the most expensive product category they could possibly be competing in. So the retreat, if that's what it is, isn't necessarily because they can't do it technically. It might be because the math just doesn't work right now. And that's a really uncomfortable position to be in when a competitor shows up who apparently has the resources and the commercial backing to absorb those costs and push forward anyway. Alibaba's cloud infrastructure is massive. They have data centers, they have the compute, and they have a business model that can subsidize the losses while the market develops.
SPEAKER_01So basically, OpenAI looked at the bill and said, not right now. And Alibaba looked at the exact same bill and said, we'll take it. Okay, so let's bring this down to earth for a second. Because I think there are two types of people listening to this. There are the tech people who are already thinking about API integrations and developer ecosystems, and then there are the regular people who just want to understand what this means for their actual life. Let's talk to both of them, starting with the regular people. What does Happy Horse actually change for someone who's not a developer?
SPEAKER_00Okay, so for the everyday person, and this is the part I find genuinely thrilling, the barrier to creating high-quality video is about to collapse. Right now, if you want a professional-looking video for your business, your social media, your creative project, you either spend thousands of dollars hiring people, or you spend hundreds of hours learning to do it yourself, or you settle for something that looks kind of rough. Tools like Happy Horse start to eliminate all three of those constraints. You type a description, you get a video. And I know that sounds like a slogan, but the quality gap between what these tools could do six months ago and what they can do now is enormous. We are at an inflection point where good enough for a real use case has arrived. Think about a small restaurant owner who wants a video of their signature dish for Instagram, or a teacher who wants a short explainer clip, or someone running an Etsy shop who needs product lifestyle videos. These people cannot afford a production team, but they might be able to afford a subscription to an app that's powered by something like Happy Horse on the back end.
SPEAKER_01And that backend piece is key, right? Because Alibaba is planning to open up API access, which means developers can build those apps. The way I think about it, and tell me if this tracks, is like a ghost kitchen. You know how some restaurants don't actually have a dining room? They just rent space in a massive shared commercial kitchen, and they put their own brand on the delivery bags. The customer has no idea where the food was actually cooked. That's essentially what's going to happen here. Alibaba built the kitchen, Happy Horse, and they're going to let thousands of smaller apps and businesses rent access to it. So you might end up using Happy Horse's video engine through some completely different app with a completely different name and never know it. And if you want a perfect example of what that future looks like in practice, there's actually a really useful breakdown in an episode we did called the GPS method. All about letting AI do the heavy lifting in your workflow. It's worth a listen right after this one if you want to see how these tools fit into your actual day-to-day. But Laura, back to the big picture. What does this mean for the professional creative industry? Like Hollywood, advertising agencies, the people whose entire career is making video?
SPEAKER_00This is the complicated one. I don't think it's as simple as AI is coming for everyone's job. That framing is both too alarmist and kind of too boring at this point. The more interesting question is: what does the creative industry look like when the cost of producing a video drops by 90%? Because historically, every time production costs collapse, you don't get fewer videos, you get an explosion of video. More creators, more content, more formats nobody had thought of yet. The internet itself did this with text and images, YouTube did this with amateur video. But here's the tension: the people who currently get paid to do the work that AI can now automate, cinematographers, motion graphics artists, certain kinds of editors, they are going to feel real pressure. And I don't want to gloss over that because those are skilled people with real careers. The optimistic read is that the best ones adapt and use these tools to do things they could never afford to do before, like a solo filmmaker who can now afford visual effects. The pessimistic read is that a lot of entry-level creative work disappears and the latter gets pulled up. I think both things are probably true simultaneously.
SPEAKER_01And that tension is playing out against this bigger geopolitical backdrop that I don't want to skip past. Because this isn't just about which app wins a leaderboard. This is about who builds the infrastructure for the next generation of the internet. If Happy Horse becomes the dominant engine that powers video creation globally through all those Ghost Kitchen API integrations, then Alibaba, a Chinese company, is essentially running the factory that produces a massive chunk of the world's visual content. And that is a very different world than the one where American companies control that stack. I'm saying it's a shift with implications that go way beyond tech. Who controls the tools shapes what gets made and how.
SPEAKER_00Like when Google launched, nobody thought this is the company that will mediate humanity's relationship with information for the next 30 years. They just thought it was a better search engine. Right now, Happy Horse looks like a better video tool. But if Alibaba executes on the API strategy, if they get developers building on top of this, if they price it aggressively enough to crowd out competitors, then a few years from now, we might look back at this week as the moment the center of gravity in AI content creation shifted east. And I think the part that should make people pay attention is the speed. This didn't take a decade. This took months. The pace at which capabilities are advancing means that whatever you think the landscape looks like, it's already changing.
SPEAKER_01Yeah. That speed is the thing that keeps catching everyone off guard, including the companies that are supposed to be leading.
SPEAKER_00And I think that's actually the most important takeaway for anyone listening who's not a tech executive or a developer. You don't have to understand the code. You don't have to track every benchmark. But you do need to understand that the tools available to you are about to be radically different from the ones that existed six months ago. The ability to create high-quality video is becoming a skill that anyone can access. And in a world where video is how attention works, how businesses market themselves, how ideas spread, that is not a small thing. That's a pretty big deal for anybody who wants to build something, sell something, or say something.
SPEAKER_01Laura, that is exactly the kind of clarity this topic needed. The way you connected the infrastructure shift to what it actually means for a real person trying to create something, that's the frame I think a lot of people were missing. So thank you for that, genuinely. You can find Laura at, and we'll put the link in the show notes. So go follow her there and keep up with everything she's tracking in the space. And if this episode made you think differently about what AI video means for your work, your business, or just your understanding of where the internet is headed, do us a favor. Subscribe so you don't miss what's coming next. Leave a review if you want to make a podcaster's week, and send this to one person who still thinks Alibaba is just for bulk orders. The race for Who Builds Tomorrow's Screen just got a lot more interesting. We'll see you next time.