Pocoyo's AI Era

My interview with Sixte de Vauplane, CEO of Animaj

Pocoyo courtesy of Animaj

To my mind, we are facing two separate crises in animation right now. First, there’s the distribution crisis. Kids are watching cartoons on YouTube, TikTok, Roblox, some new platform that’s probably being invented this very minute, and our financing and production models haven’t kept up. Then there’s the technology crisis. I cover it here every week: AI is disrupting how animation and VFX are produced at breakneck speed.

To use a term of art, it’s a doozy.

Sixte de Vauplane not only sees these two issues as connected, he believes the technology crisis is the solution to the distribution crisis.

De Vauplane co-founded the French animation studio Animaj in 2022 with the goal to use AI to create animation in a fundamentally new way. The company soon acquired the enduring cartoon brand Pocoyo as their flagship property and a proving ground for their new AI tools.

You can read about Animaj’s approach to AI in this great Cartoon Brew article. Animaj doesn’t use the text-to-video models most people think of when they hear the term generative AI. Instead they use proprietary models that translate storyboard drawings into rigged CGI poses, and then use machine learning to inbetween in that distinctly Pocoyo way. The result is totally editable Maya files, but produced much, much faster.

Equally unique is how open Animaj has been about their use of AI. At a time when many studios are being coy or downright deceptive about how they use the technology, Animaj is at least being transparent, whether you love it or hate it.

I spoke with Sixte shortly after Animaj announced a new US$85-million investment in their studio. The conversation below has been lightly edited for length.

courtesy of Sixte de Vauplane

LIFE IN THE MACHINE: Thanks for taking the time to chat with me. Let’s start with the new funding round that was recently announced for Animaj. Congratulations. What are your plans for this new investment?

SIXTE DE VAUPLANE: We’re thrilled to get the new investors on board. I think it's a positive announcement in quite a turbulent industry. It shows that there is some hope. This new fundraising is going to be the fuel to help us accelerate in three ways.

First, we want to build a very strong and complementary portfolio of IPs. We've acquired three IPs so far. The idea with this round will be to help us fund other acquisitions in the coming months, using Pocoyo as a best template for the next IPs we want to acquire - Iconic, highly differentiated, and with a good digital footprint, or the ability to be revamped or scaled through our digital first-approach.

The second pillar will be to double down on AI. We are super bullish on it. The company has been built on the promise of AI, but [we are distinct from] other AI companies because we are a content producer, so we know exactly what it takes to create content, and we know the difference between poor quality and outstanding content. We really want to use AI in a way that can empower the artists and remove tedious tasks. That's why we've been investing over the past few years in the Sketch to Motion technology, which is a way to go directly from the storyboard to animation blocking and inbetween animation.

The third pillar is the monetization engine. We want to be in a situation where we can be fully independent, or at least not asking for permission from anyone else in the industry to produce a show. The old model of commissioning from broadcasters doesn't work anymore, and that's the current crisis the industry is facing right now. So, we want to be in a situation where we can build a flywheel model, and we can monetize on our own through YouTube, VOD, FAST channels, music, DSPs, gaming, etc. To be truly everywhere all at once, but also profitable, so that we can reinvest the savings into more content, better content, and more marketing in order to support the franchise in the coming years.

These are the three elements, and that’s just the beginning. I would say that the most difficult part is ahead, because we now need to deliver on these immense tasks.

LM: So, do you see AI as the key to being more independent?

SDV: Totally. Just for one reason: you cannot be successful on YouTube if you have a price point [that doesn’t work for the platform]. Let’s assume [you have a show] that you produce at US$15K per minute. That’s not crazy for a premium series, but it's far too expensive to be profitable on YouTube. On YouTube, you're getting paid $1 per thousand views on average, which means that at 15K per minute, you would need to do more than a billion views to be profitable.

You don't have a ton of shows that are able to hit that kind of milestone. Which means that you need two things. First, you need to have a very strong scale on this digital platform. And second, you need to have a production cost per minute which is in line with the new economy. So, you should target something at least 50 - 60% lower than the cost I mentioned before. That’s where the AI comes in.

Pocoyo and friends courtesy of Animaj

LM: In previous interviews you have stated that your AI tools would allow animators to produce somewhere close to 30 seconds of animation in a day, maybe upwards of 500 seconds of animation per day. Are you finding that the animators are actually able to hit those numbers?

SDV: I think that more than ever this target is possible to reach. You know better than myself, but you can reach any quota you want, it all depends on the quality, right?

LM: That’s true.

SDV: So, the real challenge we want to address is how we can increase the quota while keeping the exact same quality. That means that I still want the same Pocoyo as before. I want the audience to see they’re not getting a cheaper Pocoyo from a production value perception. That’s why any time we are focused on quota, we couple this with a quality metric.

LM: I remember when Pocoyo came out and all my animator friends were obsessed with it. It had such a fresh look and style that we hadn’t seen in CG animation before. It’s very pose-to-pose and snappy. Is there something about the style of Pocoyo that lends itself to the AI techniques you’re using? Can you see the tools you’re developing working for a variety of animation styles?

SDV: We are going to announce a new IP acquisition in a few weeks, which in terms of style will be very different from Pocoyo. We know that our model works pretty well with it, simply because as long as we have a mesh, a rig, and a storyboard, we can use our model.

But to your point, Pocoyo was the best IP to work on in terms of Sketch-to-Motion technology, because most of the time [spent in production] was on posing and animation. You have no backgrounds, you have almost no props, which means that once you've been able to reduce the time going from storyboard to posing and then from posing to inbetween, you have reduced a lot of the cost. That might not be the same for a feature film, or for other IPs where the cost mix is more balanced.

That's why in addition to our existing AI tools, we're working on others. How can we accelerate the time to create assets, backgrounds, designs, camera positioning, layout, etc.? Because this is where we're going to create value, down the line.

LM: Do you see yourself keeping these AI tools proprietary, or could you imagine an offshoot where you develop tools for other productions?

SDV: So far, we've been focused on using them for proprietary workflow and proprietary IPs. We don't want to be in the licensing business [when it comes to AI tools]. But we believe that we can work with other partners to either create IPs together, or boost IPs together.

LM: When people working in the animation industry hear all this talk of efficiency, the main thing they hear is that they’re going to lose their jobs. How do you respond to people who are afraid that this technology will ultimately reduce work for animators?

SDV: I think about this in two ways. First of all, like it or not it’s happening. So even if it were destroying jobs, it is happening right now. I think that the most important thing we should do collectively is [to ask] how we can adapt. It's like when we went from silent movies to [talkies]. Jobs were totally transformed. History has shown that the ones who survived were the ones who managed to adapt, not the ones who were reluctant to embrace new technology.

Secondly, I believe we are going to enter a world of more and more personalized, interactive content. The industry is totally fragmented with a lot of platforms, and that means that a lot of content. A lot of content leads to more work as well.

But I think that the real question is, how can we make sure that as an industry we are going to stand out compared to the sea of cheap junk that's going to hit the market? Because that's what we are fighting right now, AI or not. More than ever, it creates an opportunity, or at least a responsibility, to be focused on high quality content. So I believe that AI is going to be an amazing opportunity for great artists who will have the ability to scale their work. I truly believe that in a world that is totally over-flooded by poor quality content, [high] quality will stand out. That's the best business plan ever.

Courtesy of Animaj

LM: So far, the AI tools you’ve announced like the Sketch-to-Pose tool, and the inbetweening tool have been trained exclusively on your own material. How important is it to you that your AI training is ethical?

SDV: First of all, we don't want our AI to have been trained on work that is not owned by us. We don't want to steal. Some other players in the AI industry are maybe less ethical in that sense.

Secondly: creative purpose. It doesn't make any sense to train your AI model on, let's say Bluey if you want to generate Pocoyo, right? If you want to generate Pocoyo, the only dataset you need is Pocoyo. Because, as you said, it's snappy, it's supposed to pause. You have a style of animation, so you want to make sure you're going to replicate that style of animation. That's why, just for a technical reason, it wouldn't make sense for us to go after IPs that we don't own because we are only focused on our own IPs.

Thirdly, we want to make sure that we can build something for the artist and be fully transparent. With our use of AI, it's not a black box. We know exactly what we've been using to train on and what the outcomes are.

It’s important that we have an AI model which allows you to have full controllability over the output, which is not the case from other models. This enables us to have a hybrid type of animation - traditional AI empowered. It's not black and white. You can choose depending on the type of shots you have.

LM: One last question. Your studio has been around for three years now. If look three years into the future from now, where do you see the technology of animation heading?

SDV: Wow. I mean, if I knew I would be rich.

The funny thing is, when we created Animaj, we started with the deep conviction that AI was coming and it was going to disrupt the entire industry. At the time someone asked me, “Sixte, when is it going to happen?” And I told him, “It’s going to happen in one to two years.” Six months later ChatGPT was released, and we all know what’s been happening since then. Everything is moving faster than you can anticipate.

And I think that's why the industry needs to adapt because it's increasingly becoming better and better. Even if there are some issues like constructability or consistency, just look at where AI technology was three years ago – we’ve been able to increase the quality of the models like crazy.

I think we are not so far away from interactive content, which I believe is going to open up a ton of new revenue streams for IP owners, plus a ton of possibilities for the audience to be truly connected and engaged with the brands. I'm a huge fan of how AI can unlock interactive content in order to really blur the frontier between gaming and animation. I believe that's going to be the next wave in coming years.

LM: This has been fantastic. Thank you so much.

SVP: Cool! Thank you, Matt.

---

I hope you enjoyed my conversation with Sixte de Vauplane. Do you see Animaj’s approach to animation as the future of the medium, or do you think their use of AI more of an outlier? I’d love to hear your thoughts.

Seeya next time,

Matt Ferg.