Hello and welcome back to Equity, TechCrunch's flagship podcast about the business of startups. I'm Rebecca Balon and this is the episode where we bring on industry experts to help us explore a trend in the tech world and dive deep. AI generated video has gone from novelty to creative tool in what feels like the blink of an eye. And nobody understands this better than Runway. They're a startup that's based in New York and they're punching way above their weight in terms of their model releases, coming out with models that are competing with some of the most well-funded labs like Google and OpenAI. As Runway scaled its video generation models, it learned that its applications are extending way beyond creativity. They are advancing into general world models with applications in use cases like gaming, robotics, and perhaps a more generalized intelligence. Here to talk about this today with us is Runway's co-CEO and co-founder, Cristóbal Valenzuela.
Chris, welcome to the show.
Yeah, really excited to chat to you again. We've been chatting all week. I've been chatting with your team about what you guys are doing at Runway, but I wanted to kick things off by bringing up something you said recently that AI could help Hollywood make 50 films instead of one hundred million blockbuster. Now I think with you know catchy headlines like that people might misconstrue or they get attached to something like oh these AI guys are just trying to disrupt and displace you know Hollywood and film making but I want to know what you're actually proposing right like what is the more nuanced argument than just treating filmmaking like a content factory.
Yeah, that's a good question. And I agree sometimes like headlines do tend to be a bit more clickbaity. I said that in the context of a lot of other things within a conference and of course, that feels like much more interesting to publish. And which I agree it's actually not that far from what I was actually like arguing. And my reasoning here is kind of actually pretty simple to be honest. It has to do with the way media and like feature films or just content in general has been made for the last couple of decades. And it's a process that's filled with a lot of technicisms and processes and budget constraints and time constraints. It's a process that's driven a lot by approvals, by convincing the right people. The industry itself, I would say, is in a tough spot. Hollywood and film making is mostly because it's getting just way too complex to approve and make anything. And I'm not the only one who's argued this. I think many people have said this before. You need like at least a couple millions just to like make a movie and then at least a couple more millions to like promote the movie. And so unless you're willing and able to fund raise or convince someone to give you at least 50 to 100 million to make a movie, then it's not probably going to happen. And of course there's exceptions and this is not a legislation for everyone, but like I would say in the large part I think most people agree with that sentiment. Just go to LA and you'll see that all over the place. And the reason there's a lack of interesting projects or lack of diversity of films is in part because of that.
Yeah. What I think is happening now and the new ingredient in the mix is well what happens when you have a technology that allows you to fast track many of those stages and steps right so it will help you iterate way faster on previsualization on pre-production and of course on making the actual frames for the movie and that's already our scene today like AI is getting deployed at scale in studios and also people and independent filmmakers. And you start to see kind of trajectory wise. It's a place where if you have a really good idea, if you have a really good script, if you have something the world wants to hear, you might be very soon in a position where you can tell that story and you can tell it for a minimal fraction of the budget that otherwise would require you something else that feels like probably at that stage like insane that we were like worth that way.
Yeah. And so my summary of that was like well yes with hopefully with 50 million or 100 million you're going to be able to make a thousand films like or hopefully many more like the constraint would not be a technical one it will be a storytelling one. Do you have something good to tell if you do then like technology will help you do it in the first place.
Yeah, you know, I think that like the concern is among many people is that this will just create more AI slop, right? And as someone who's a writer, I've worked on film sets. There's like a fear, there's a visceral fear that a lot of that will go away. A lot of that industry will go away. That collaborative nature of creating films. On the other hand, I'm not sure if you saw I got an ad for this today. There's a new scary movie coming out like this like the satire film scary movies. There's a second Practical Magic. Practical Magic is my favorite movie. I don't know that I need to see it again. My point is that there's all this existing IP that keeps getting produced and other amazing stories are being sidelined because, I don't know, whatever industry mechanisms say that we need to keep producing existing IP with the same actors you've seen for 30 years and that's the only way we're going to get a return on our investment. I'm hopeful that tools like yours can change that. But, you know, before we go into it, I want to take a step back because Runway, there are other companies in this space, right? Luma is a competitor. OpenAI is making video generation models. So is Google. And one thing that I find interesting about you guys is that you don't have that typical Silicon Valley background, right? You co-founded Runway in 2018. you and your co-founders Alejandro and Anastasis came out of NYU's art school program. Can you talk a little bit about that?
Sure. Yeah, happy to. Also to the point before I think on the quantity of content. I think that we've seen that before. I think people sometimes forget this is actually there's a law for this called Sturgeon's law. It's like the most of the things that we see are low quality then small percentage of it are good quality. But my point is that we've seen that before. The field goes back to the 40s and maybe even before that. But deep learning and neuronet networks working weren't really a thing in 2016, 2017 and 2018 when we started working on some of these ideas of runway. And so in a way when you speak about competitors and others building here when we started there was no competitors because there was no industry altogether. And so there's no one we could like compare to or look or inspire or think that we were like competing with. And I think that helped us shape a little bit of this independence and unique way of thinking about why we do this in the first place. And you're right, we started this as I would say artists and engineers. We come from a background, the three of us, from a background that mixes art and science. We were either engineers with an interest in art or an artist who wants to become an engineer. And for a long time, if I'm honest, like it was kind of strange and weird and we didn't feel at place in many different like times and moments because it was hard to classify us. What are you guys in art school, but you're software engineers, so you're publishing papers on AI, but also you're like speaking about like impressionism like it just doesn't fit. And I think that in a way has become maybe our biggest strength over time is that yes, it doesn't fit. That's exactly the point.
Yeah. The point is that if you want to make good products and good technologies and you want to create your own destiny, maybe don't live your life by labels, but like create on your own label. And I think we've the company itself has embodied that culture over time. And so it comes from a place of care what we this is what we have dreamt about doing for years. And so the fact that we can do it as a job feels also like a dream.
Yeah. And you know, another thing that I think is important to point out is that you're based in New York, headquarter in New York. You have offices everywhere, of course, but
New York strong. I really appreciate that as a New Yorker. Always fun for me. It means I can visit you in the office, which I did. Lovely office, by the way.
Thank you. One thing I noticed while I was walking around looking at the way people work is everyone seems to be I mean, it seems really flat, right? your your desk is amongst it with you know all of your other workers desks and it feels like there's film making people next to engineers and like are you how do you consider yourselves are you a creative tool an AI lab both.
Yeah I mean you think too much about the best way to describe like the label of a company but if you look at it like very practical we are a deep research lab an AI research lab that's pioneering and building like large models so yes Think of us as a large research lab. We also create products and create like exciting tools. And I think we're a software company in some domain. We make movies. We like making shows. We like making just interesting pieces of content. I don't know how you can classify that as a studio creative agency set of artists. Again, I think like the sweet spot for us is like what happens when you have all of those things in the same place. to your point, you can go into the office and you literally have someone who's working Hollywood and visual effects for the last 10 years sitting literally like centimeters apart from like a research scientist working on like pre-training on large autoregressive models like and those people funny enough can speak the same language which is not that common and they can speak that same language technically and from a value perspective because some of our researchers and engineers are artists they're musicians they're photographers they they also make films and some of our artists and creatives are also engineers. So, it's a nice combination of things and I think yeah, it's a I think we've built a pretty special place from a cultural standpoint.
You know, one of your early investors told me that you guys refused at the beginning at least, I don't know if you're still doing this, to hire people with a PhD at first just because you were like, 'No, we're doing things differently here.' Is that true?
Maybe it wasn't like that we refuse to like hire people with PhDs. It's more just generally I think we we tend to not look at credentials as much as maybe others might have like some of our best people might not have any of the credentials you traditionally look for and that's kind of the point like I think talent is equally distributed in the world and if you happen to get into Stanford somewhere else like might not necessarily mean a lot about how much you can do and how much you can execute and build or learn I think we want to avoid that like it's a meritocracy like we we judge you by the outcomes of and what things you can make, not by who you are or where you went to school or who you vacation with or who you know maybe because like that's where we started. If we we judge ourselves by who we knew and where we were and our last names and our titles, then we wouldn't be here. And so it's a representation, I would say, of this very hard work driven meritocracy that we've built internally as well.
And I know that that has also had its effects on on you know how you've had to be lean as a company and how you've had to you know start revenue generating pretty early. So like really the revenue generation was happening alongside the research. But I do want to talk a little bit about the research, right? Because I was, you know, in the course of researching you, I was looking at some of your, you know, the first video gen models that came out and, you know, of course for the time they were impressive for the time, like two years ago or whatever, four years ago, and it just reminded me of that Will Smith eating spaghetti thing, like we've come a long way in video generation, right? And one piece of research that I believe came out of runway was about dual process image generation, right? Having this kind of like judge vision language model assessing AI image generators in real time. Is that something that you guys were working on?
A couple things I guess on how far we've come. Yeah, we we've come very far. I mean I think the first ever video model we put out that maybe seen some of the outputs was around 2023. So like two and a half years ago which in AI landfill like three decades ago. And it was such a big deal that it was in the front page of the New York Times. We had a like a whole article if people want to search for it. It's like everyone was like amazed that you could render this like incredibly low res, incredibly bad by today's standards like video output. But it was it was it was new. It was totally new. Like no one thought you could do it. And so there we are. And today we're able to output like incredibly hyperrealistic 4K footage with a level of realism and cinematography that you'll like have never imagined by just two years ago. I thought yeah I think the industry and the field has gone really far and the idea for us will be like well hopefully in like three more years we'll look back at today and the models we have today will feel like the model we had two years ago and I think we're heading towards that trajectory very soon. From a research perspective we do a lot of research interesting research. We publish a couple things. We don't tend to publish all of it for mostly competitive reasons. We do have research in the video domain and image domain. We publish also things in like in biasing. So how do you make sure that the models maybe that's the one you're referring to. It's how do you make sure that the models can become a bit more self-aware of how they're depicting things in the real world? Because like the models are trained in such a way that they may reflect how humans see the world. And so if you want to had a more diverse sort of outputs of a model, then you need to improve the way you train the model. And that's something we've spent time and spend like significant amount of resources to make sure the models are from both from a product but also from a created standpoint and diverse standpoint like interesting enough for people to actually use.
Zooming out a little bit. So Runway has become you know go-to for filmmakers, video editors. You've raised hundreds of millions in funding. pitchbook says somewhere close to 860 million at a 5.3 post money valuation. Your last round I remember covering that and it was with the angle that you're you're trying to go deeper on world models. Can you explain why there's this push we're seeing just generally from video to world models? Like how do those two meet? Because Google trains theirs separately. They train VO separately than Genie. So I'm kind of curious how you see them moving together.
Yeah. To the point of like not having like a necessarily like a owner on destiny that I was speaking before like we spoke about this idea of general world models back on 2022 2023 and so relatively early within the like lifetime of what we now consider like a common trope of AI which is how do you build this general world models or world models and I think we build our own intuition as to what that actually means and how we think about them. Maybe an observation is I think generally world models have become like a cultural term to describe anything. I saw a legal company recently describe their work as an model that I'm like okay you're kind of stretching it a little bit.
So how do you define a world model?
So for us it's always been around this idea that you could train models that can understand the physical world in perhaps a similar way that we humans do. And there's an argument to be said that the best path for that is you need to show the model or models like enough pixel or video data. And if the models can understand how videos should be or how videos are created in the real world, they're naturally understanding or becoming aware of things that occur and happen in the real world. Physics, gravity, cohesive reflection, cause and effect. those are things that we don't tell our models to learn from. We don't tell a model hey when you throw an object like the object should move at this velocity and with this set of vectors and this is why the model just implicitly learns about it. And so in a way video has become the most obvious representation of how you can train AI systems that can understand act and behave in the world in similar ways that we do. And that's ultimately I would say a lot of the holy grail over time is you don't want models that can describe reality. So you think about language models as models that can describe reality because they train on text and text is a representation like an abstraction of like reality that we humans have created. Training on video alone is a bit more pure because you can simulate reality. You don't have to describe it. You can just simulate it and learn from that simulation. So that I think for us has been a much better path and philosophy of research than just fully focusing on language alone.
Are there certain things that LLMs do today that you think could kind of be replaced with just with video?
Yeah, I mean like visual reasoning and like visual planning. So that's why some of the most interesting compelling use cases of world models is in the robotic side. Having a system understand how to navigate a very particularly complex action in the world. The way you can think about is I'm going to like have a language model like just do like vision and try to reason about it with language. Another one could be more like we do like when when you're tasked to do things in the real world, you're not you don't have an internal monologue all the time about how things work. Like sometimes you drive and you don't even remember how you drive. you just like you drove, you know, but you have an internal implicit understanding of the world and intuition, right? So, but because you you've seen enough times, you've drove it enough times, you've seen it that your brain is kind of hardwired to do it and there's an argument to be said that there's a similar way to train artificial networks and systems that can do that same kind of behavior, but now you fit it videos and instead of language alone.
Have you started training single models on multiple modes like voice, text, video all at the same time?
Yes, we we train like what we call omni models. Omni models are models that have kind of the combination of those things together. Yes.
Okay. So for world models the you know near-term use cases that everyone's talking about. I mean others in the field you know I mentioned Google before there's world labs but you know Fei-Fei's company. They're pretty focused on world models. It seems like the trajectory is entertainment use cases first, right? Like gaming, interactive entertainment and then robotics and I know Runway has started doing a robotics push as well. Is there something beyond that?
