S8E1: Cristóbal Valenzuela Brings Us into the Future of Creative Editing with Runway and their Generative AI-powered Suite

The best way to think about Runway is to think about it as a set of tools or a suite of tools. We have around 35 different tools. Most of them are AI-powered. Baking video sometimes just takes you a lot of hours, and we have algorithms that will help you do that. But we also have tools and algorithms and services and offerings that help you generate new content. If you don't have videos, you don't have content, don't worry about it. You can just actually generate it.

Welcome back to season 8. Clauds, how we doing?

So well. Can't believe it's 8 seasons of The Room Podcast.

It feels like just yesterday we were getting started in the middle of COVID and had this vision for sharing the future of founding stories to let other founders be in the room where it happened. And so much has changed for us since August 2020 when we first had this inception. You're now a founder and co-CEO of Priv, which is a startup unlocking and disrupting recurring revenue for e-commerce brands.

And Madison, you are crushing it in venture. You're now a partner overseeing investments at Defy VC, an early-stage venture firm in the Bay Area, and of course an investor in Priv. We're just two women navigating our careers and asking the people who inspire us to shed light on their stories. Unlocking access starts with a conversation, context, and As a reminder, we open the door to moments in technology history that traditionally happen behind closed doors. With our guests, we unpack the experiences that led to their success and look towards the future in their verticals.

And we hit an incredible milestone at the end of last season with Spotify sharing that we were a top 5% global shared podcast in the technology sector. And so we want to thank you for being a loyal listener and helping us achieve this milestone of amplifying more equitable voices in the room where it happens. As you look towards season 8, get excited. We're excited. We have some incredible founders from the founder of Superhuman, the founder of Incredible Health, Webflow, Ty Haney with her new business TYB, and more. So join us every Tuesday where we sit down with leaders, founders, and funders who are changing the way it's always been done. You could expect themes across, of course, generative AI, navigating your startup in a downturn, the creator economy, and more. So be sure to hit subscribe, like, and follow so you don't miss any Tuesdays in the near future. And one quick note, since season 3 in 2021, we have been delighted to work with our partners Kooli and Silicon Valley Bank. This season is no different, although the events of the past few weeks have brought Silicon Valley Bank into the forefront of our ecosystems conversations. Now known as Silicon Valley Bridge Bank, SVB. Thank you. And now a short message from our sponsors. Following message was recorded prior to 3/10/23. Support for The Room comes from Silicon Valley Bank.

What's next? What if?

Now what? Silicon Valley Bank understands these questions can keep founders up at night like Claudia. For over 35 years, Silicon Valley Bank has helped high-growth companies through scalable financial solutions, plus insights and expertise that many other banks just can't, which could be why 50% of US-based venture-backed tech and life science companies bank with SVB. Learn more at svb.com/next. Silicon Valley Bank, built for what's next. Cooley is a global law firm built around startups and venture capital. Since forming the first institutional venture fund in Silicon Valley, Cooley has formed more venture capital funds than any other law firm in the world. The firm has 60+ years working with VCs, helping form managed funds, make investments, and to handle the myriad of issues that arise through a fund's lifetime. Learn more about the firm at cooley.com and also at cooleygo.com, Cooley's award-winning free legal resource for startups. We're starting season 8 by jumping straight into the modern era of video and audio editing with Cristóbal. Chris Valenzuela, the CEO and co-founder of Runway ML. Runway makes the future of editing accessible to everyone by providing tools to make your editing journey quicker and easier, from ideation all the way through to post-production. Chris himself grew up in Santiago, Chile, where he went to a local university before ultimately moving to the US to attend graduate school at NYU, studying interactive telecommunications. Chris and his co-founders, Anastasios Dourmandis and Alejandro Matama Ortiz, put it all on the line to stay in this country after graduation to start Runway ML. Chris speaks to the nerves he felt taking that risk, but the reward that came a few short months later when Runway found a believer and their initial investor base. Fast forward from 2018, the term generative AI is the hot topic everyone is talking about, and large language models, or LLM, is the new acronym of choice. Runway is at the heart of innovation in this space, having recently released their Gen 1 foundational model, as well as continuing to advance their application layer product for creators, just like Claudia and me. With their recent Series C round done at a $500 million valuation, Chris and the Runway team are indeed in a strong position to continue moving storytelling forward. Themes in this episode include the intersection of art and science in storytelling, the magic of the word-of-mouth go-to-market, and the future of large language models. Let's open the door. Chris, thank you so much for joining us today in the room.

Thank you for having me here. Super happy to be chatting with you.

We're really thrilled to have you on. With all of our guests that we have joining us in the room, we love to start it off with the beginning, the true beginning. Where did you grow up and how has that shaped your view of the world?

I'm from Santiago, which is the capital of Chile. Chile is a relatively small country. We're like 15, 16 million people. I always compare it to like one state in the US. It's the entire country, right? I grew up in Santiago, have 3 brothers, and so we're very close to my mom who raised us. And from the early stages of my more professional career, when I was deciding, Chile, what you have to decide is a very interesting kind of like college situation where you go and study like different subjects. When you're 18, you need to define your career. It's like you go and like you study one thing and that's it. You can't change it. And so everyone, when they're like 16, you're trying to figure out like, what do you do? Right? You need to figure out the rest of your life in the next 2 years. And what I've tried to always cultivate at that time when I was trying to figure out what to do with, what to study, what to work on, always try to get open-minded and curiosity-driven approach to my education. And I think that has shaped a lot of my professional life after. I think a lot of it has to do with, I would say, my mom raising us with that curiosity component. Back in Chile, it's a very conservative country. It's a small country. It's very traditional. So it's always hard to find different points of views and different things to do that are outside of the norm. But I've always personally tried to be very open-minded and try to find new things to do. And I think that on its own has been something I've tried to embrace and has changed the way I see the world.

I love that, and I love that thread of curiosity, which no doubt has pulled into your founder journey today with Runway and having that mindset of really the world is your oyster. What could we build and how could it change? But curious about that vein of early identity, did you, in that world of curiosity, think you would become a founder?

No, not at all. I've always like building things and doing things with my hands or computers. And I came to start a company because I thought, and I realized that this is the most efficient way of building things. I think being a founder is like the subsequent effect of building something. From the very beginning, actually, we're 3 founders and day 1, when we started the company, we never actually thought of ourselves as co-founders or like CEOs and CTOs. We're just building something. That's it. You're trying to prove as a company and as a team that you can actually execute on an idea and actually learn and continue iterating on it as much as possible. I think later on the journey, you get the chance to call yourself like a co-founder, the titles that people love having, which I really like never felt very deeply passionate about. I still just like building stuff.

And I think that really determines how I like to think about my role and also the So with that builder-first mindset, we're gonna touch on your co-founders in a second here and the journey that you guys decided to embark on together. But for you, taking the curiosity builder mindset from undergrad in Chile where you went, and then ultimately finding yourself at NYU in a graduate program, could you share with our listeners just a little bit more about what you were curious about to continue studying and learning and building around?

Now that I think about it in retrospect, all of what I've been working on or researching on and studying has come to a perfect like conclusion right now with Runway and all of my interests in art, design, software, and business have kind of like— Berkeley find its way for Runway to happen and to be what it is today. I studied econ and business in Chile, and then I studied design. While I was doing design and working in a consulting firm, I was also working as an artist and I was helping other artists in Chile work on different like pieces. And all of those things happened at the same time, which was At the time it was like, to be honest, like strange because I was on my day job, I was having like more of a business kind of formal like job, but at day night I was like making art. I've kind of embraced that, to be honest. It always felt conflicting because there were two worlds, two separate worlds, but for me they were like one. There was always just one. You're working on one single thing. And the change for me and the realization of how can I take this further, combining these disciplines into one single thing. Was when I came across ITP, which is this technology program. It's part of the art school at NYU where you can mix technology in very unexpected ways. It's basically, it's been running for 40, now 50 years, I think. The best way to describe it is art school for engineers or engineering school for artists. You just go and explore your curiosity from a technical perspective, from a research perspective, from an artistic perspective. And that for me was like the inflection point of, hey, I've been just doing a bunch of random things. Perhaps I should try to combine all of them. I don't know in what shape or what form, it just, I will follow up that curiosity, that stepping stone to understand what's next. And now that I come back to it and I think about it in retrospect, building the type of tools that we're building on Runway, reimagining the creative suite and thinking about very expressive tools requires you to really connect the science world and the business world with the arts, they need to speak the same language. You're building core research on generative models, on AI for image and video generation. And if you want to apply that in filmmaking, in storytelling, in art making, you really need to understand how artists work, how filmmakers actually make movies, right? And being able to speak both languages, what I thought was like strange 10 years ago, is like now very obvious and very useful. It's allowed me personally to really have empathy for both ends. And at the same time, it has allowed us to build a company with that culture as well. Most of the folks in Runway right now also have very similar backgrounds and we tend to hire folks that come from similar backgrounds. So we have our first hire actually ever, he has an arts degree and now is our senior backend engineer. It's a very interesting combination. If you study art, but you're a backend engineer, that's perfect because you're who are we going to empathize really well with our customers, with the folks who are actually using the tool? And so that combination of like elements and very unique approach has been a core DNA and ethos of Runway.

Thank you so much for sharing that. I'm smiling because I did a joint degree in art history and computer science. And I also thought at the time that I was a little crazy trying to mix the two, but certainly in this day and age where tech is just disrupting every industry and people are building tools for users that touch anything from painting to deep tech. It's really important to harness those dualities of interests in order to build products that are better for different groups. And so I'd love to just double-click into the founding journey, the early days of Runway, and how you and your co-founders met. So in 2018, you set out to build Runway with 3 co-founders, Anastasios and Alejandro. You all met at NYU. Tell us a little bit more about how the three of you entered the journey as co-founders together.

We met at NYU. We were all studying at this program called ITP, but we immediately connected. Alejandro, one of my co-founders, is also from Chile. We didn't know each other in Chile, which is strange because Chile is super small. We never know each other. We just met here. We immediately became best friends. And then Anastasios, who's my other co-founder, we immediately— I was just struck by his like technical abilities and his way of thinking and his perception of art and how he's thinking about doing work in general. Same with Alejandro. And so you learn a lot in ITP by doing. You just build, constantly build. It's a program where you can just execute whatever idea you have and just, you do it. And so when Runway was an idea, was forming, and we were like experimenting with how do you take this really disruptive technology and put it in the hands of artists, a lot of questions emerge around, hey, how do you make this work? What are the technical challenges of making a tool like this? For whom? For what? And I think like for us, again, I go back to like building mode. We're less concerned about starting a company, more concerned about just doing it. We're passionate about just doing things. We actually looked at very different like funding modes. We're 3 immigrants as well. So we had to figure something out to just be able to stay in the country. We had to like had a job. It wasn't like, oh, I'm just gonna take 2 years off and figure something out and then see if it works. No. You graduated, you have 4 months to figure something out. We had a conversation with investors and one thing led to the other. Anastasios actually had a full-time job after school and he quit and it was like, great, we're gonna do this full-time. We need to all be pretty much doing this in a full-time basis. And Alejandro and I were research residents at school after we graduated. So we also had to quit. So that was an inflection point for us when we was like, okay, we're gonna like try really going further with this, we just need to do it. But the story is always, again, goes back to just that mentality of building more than anything else. And we come to find ourselves as founders almost by accident, because that's the most efficient way of building a company.

We've had many guests on this podcast who have identified themselves as accidental entrepreneurs. So it's really cool to hear that you guys were just passionate about something and you just kept working towards this idea that you were excited about, and then it resulted in something that started scaling. There might not be a specific moment, but if there was a bit of an aha moment where you realized, hey, we have something that's going to change the world, tell us a little bit more about that juncture in time.

One moment that I always remember was a striking moment for me. It was very early. We were still early on the fundraising path and trying to figure out possible ways of making the product go further. And I've happened to connect with this very famous founder. He invited me to his office. So we had coffee and I was telling him, hey, I'm thinking about starting this company. I want to build this product. I'm thinking this could be huge, specifically for creatives and arts. Every time you say creatives and arts in businesses, you're like, great, that's not a good idea. It's not going to work. But it's true. This is where I'm coming from. Like, I want to build tools for artists and I feel like this could be very big. Beyond what we already consider as filmmakers and artists. I think this technology will change everything and we should pay attention to it. And he was like, yeah, sure. It's like small. I don't think it's too big. And like, maybe I would have realized is that a demo always means more than any presentation or any combination of words. A demo is worth like a thousand meetings. And so I was like, let me show you something. Let me just show you what we're working on and just open the computer. And I think body language really matters a lot when you're speaking. And his body language when we were speaking, he was like, oh yeah, whatever. Like, sure, maybe it's gonna work. And then I showed him the thing and then his whole posture just changed immediately. Like, in a second, it was so obvious. The biggest impact for me when he changed his body posture and he started paying a lot of attention to what I was showing him, it was like, whoa, okay. Seems like I'm into something. And so for me, that was a moment perhaps of realizing that, hey, what we're doing is gonna be huge and we are building it from the inside and we think it's gonna be important, but getting that validation and getting that exposure and that it's just a body language component of, hey, this is okay, this is important for me, was like, great, where we're heading into is gonna be massive.

Thank you for sharing that, that anecdote. Today Runway is the next gen content creation suite, and there's many different aspects to your platform, but maybe for our listeners, tell us a little bit around what your core product is and what were some of the early instantiations of it?

The best way to think about Runway is to think about it as a set of tools or a suite of tools. We have around 35 different tools. Most of them are AI-powered, so there's algorithms and research powering the core components of those tools. And with those tools, you can do a combination of different things. You can edit video, you can automate traditional video editing processes that are very time-consuming and expensive, which is something that A lot of post-professional folks in the film industry use Runway for, for example, I don't know if you've seen Everything Everywhere All at Once, a beautiful movie, it's been nominated for 10, 11 Academy Awards. They've used Runway for a few scenes. We have The Late Show with Colbert, the whole team in their video department using Runway and countless like VFX studios and post-production teams. So they use Runway to automate and simplify a lot of the tedious aspects of making video, right? Making video sometimes just takes you a lot of hours in these very boring, like time-consuming aspects and workflows. And we have algorithms that will help you do that. But we also have tools and algorithms and services and offerings that help you generate new content, right? And so this is where the generative AI component comes in. If you don't have videos, you don't have content, don't worry about it. You can just actually generate it and then take it back to the editing component and do the editing part of it. That's why the best way to think of Runway is like a whole creative, like, suite and experience where you can do pre-production, which is like getting the content or generating the content, and then post-production, which is can you edit the content? So you can do both. All happens in the web, so you can access all these tools directly from your browser and in a very collaborative environment as well. So we have teams and companies and departments and different organizations sharing assets and sharing their projects with one another.

You mentioned some of the AI technology that underpins Runway's products. And I'd love to double-click there a little bit. Could you tell us a little bit more about working with early large language models and how that really shaped the trajectory of the product suite?

A unique aspect of Runway is that we think of ourselves as a full-stack company. What I mean by full-stack is that we just don't do research. We do research on core fundamental foundational models and techniques for both image and video creation, but also multimodalities. And we publish multiple times. We have a core research team that drives a lot of the product forward. We also have an infrastructure team that deploys these models and builds robust systems to make sure that we not only can train them and improve them consistently, but also deploy them in safe ways. And the third component is the product part, which is we build applications and UIs and services that allow you to use those models and our stack in a very simple way. So When I was mentioning CBS or The Late Show or Everything Everywhere All at Once and the countless filmmakers and editors, they're not using the models on a raw basis. They're just going into an interface, clicking on the video side, and the magic happens behind the scenes, which is fantastic. Now, to enable that to happen, we do a lot of research on language models, as you were saying. We do a lot of interesting work on code generation, but code generation not for the sake of exposing that code generation. It's actually for driving product. The way people interact with the product could be or might be driven by some of those language models behind the scenes in the same way that the image models are driving the image generation process or the video generation process. So our stack is unique because we don't only focus on developing those research models, but making sure that we integrate them directly into the product side of things. And language models, have been one of those many kind of like research endeavors that we've adventured into.

That's super fascinating to hear. It seems like a lot of companies are either on like the research foundational technology building side, and then you have a suite of companies that are around like application layer. Super cool that Runway touches the entire scope from the research to making sure that what you learn from the research is incredibly helpful for your end users. As an early stage founder, one topic that I've been spending a lot of time on is figuring out a go-to-market motion. And so I'd love to chat a little bit around Runway's journey there. You've mentioned some incredible TV shows, Academy Award-winning movies leveraging Runway. At the same time, today's product can be for anyone starting at $12 a month with the individual creator package. Was this a bottoms-up motion from day one? Tell us a little bit about how your go-to-market has evolved over time.

I think every founder is always trying to figure out their go-to-market that never ends. From day one, we were, I wouldn't sound like a broken record, but very product and building focused, right? Let's build, let's make sure that we can not just imagine a world and a vision, but actually execute on it. And so product for us has meant most of our focus for the last 3 and a half years. And that actually has driven most of our user acquisition. We Couple of months ago, we didn't have a marketing team. The whole marketing team was actually just a product. The product was driving organic signups and in a very word of mouth driven approach, we were growing within different companies and teams and organizations. Actually, it's funny that the story about how CBS got into Runway, I think perfectly exemplifies I would say the momentum and the go-to-market motion that we've seen in repetitive ways. One of the editors of The Late Show just came across Runway in a YouTube video, and it was a YouTube video, an organic video that just someone made because they were using the product and they were just fascinated with it. They thought, there's no way to break to feel grateful for this product that I'll share with my YouTube channel. And he did that. And then this other person just found the video. And he was like, okay, this can't be true. He used the product and it was true and it worked really well. And he used it for one video and he used it in a way where his boss was like, hey, you need to do this workflow. I need you to edit this video. It's going to take you like 6 hours. So just go ahead and reply back or whenever you're ready next day, send it to me. And in 5 minutes, he was like, here's the video. And he kept that secret for a week where he was just superhuman. And then after a week, he's like, okay guys, I'm going to Everyone in team, I'm going to show you my secret and here's the thing I've been using. And then we saw 6 subscriptions coming up like right after each other, like da da da da. And like, it's great. Thank you for being our head of sales. And now we have the whole team with Runway. And I think that motion has been very similar in the way we've grown and see the product being adopted by companies mostly. Yeah.

Your existing customers being your salespeople is probably the best way to acquire new customers. So that's a very cool story.

Could I just double-click on that really quick, Chris, as well? Because as Claudia just alluded to, right, you had this super power user who got their whole team on it, at what point did you then go into CBS and say, hi, we have an enterprise license, so you're not paying $12 per person per month? Like, did you switch the flip on them at some point?

Yeah, we, we talked like, hey, what's up? All these folks coming up and signing up for the product, we should chat. And it's like, yeah, let's chat. And we of course were able to onboard them on more parts of the product and help them achieve more with it. And that for us has been crucial, but it's really interesting where, when you are so product focused that you don't have to sell the product because the product can sell on its own, it's great, but it's super hard to do because you need to do 10,000 things to get to the point where people are going to like the product. And also I think it's important to make sure you sustain that. So it's important that you can build it and people will love it, but sure, can you make sure that continues to happen to be the case as you do more stuff. And I think our motion in terms of go-to-market has always been driven by that organic adoption. And like today, we've never run a marketing campaign. We have never paid for an ad. And so all, every single customer that has come to Runway has come because they heard about it somewhere else. I think that's a stronger motion personally, I think, but it's also harder to do because it takes more time. And if you don't have time to do it, it's probably not the best option. So it depends a lot on the type of company that you're building.

I'd love to ask, the entrepreneurial journey and building a company, even with those incredible stories and high highs, is not always a straight line. Could you share a moment where things weren't going as planned and how you really got over that? What advice would you give to other founders?

There's been so many. Um, the world is saying no and you're persistently trying to say yes. I mean, you're just very optimistic about making sure things will be figured out and you figure out something. There are a few moments where I think like it was particularly challenging when we're thinking about starting the company and like just getting things ready. Everyone told us you couldn't do it. You just can't because you guys are 3 immigrants. Like you don't have green cards, so you can't start a company. And I just remember like speaking with like previous founders and they're like, yeah, you need to go through like this 6-year-long process or you need to first work at a tech company and then do all these things. And it's like, I don't have time. I have 3 months. That's it. And I think we clearly realized that it wasn't that— at first I was like very frustrated. It was like, I want to build this, but it seems like I can't because it just like really can't stay here. But I quickly realized that it's not that I can't, it's that it's never been done before, but no one has ever tried to do it before. Because if you fail, there's no plan B. And I think for us, it was like, great, let's go all in. And the thing is that you can do it. You can start a company and like you can build what we build. They think it's super hard because if it doesn't work, if you don't get funded, like sure, there's no plan B, you're out. And so a challenge for us has always been navigating and trying to figure out a lot of those things, especially as like immigrants that are relatively new to the ecosystem. But it's now become a strength of us where the mode of Runway is just figure it out. Everything can be figured out. And if someone says like, hey, that's not feasible, that's not possible, is it? Or is it just because they've never actually tried this themselves? I've realized that's just the journey of a founder. Like, there's a lot of times where like, things aren't gonna work as you expected, but you just need to figure it out. Just figure it out. And if you figure it out, a lot of things will happen in a good way.

That definitely resonates, the idea of the reasons behind the no. And if you understand the why behind the no, sometimes you can understand why that specific person said no and actually uncover why for you. That might not be a no. And in this instance around your willingness to go all in despite the downside risk of maybe getting deported really was saying, hey, this is a challenge like I am willing to take on and I want to be the one to get to a yes here. And I have conviction myself for turning this no into a yes.

Yeah, exactly. And I think for me that's just first principles thinking. You should really go to the root and understand like I mean, I had to read things and understand the law and understand like immigration systems and understand like processes that were like very foreign for me, but just figure it out. Just figure it out. If you're smart enough, just go and like learn it and you'll figure it out and you can do it. And now on the immigration side, that's been a strength of Runway because we've been able to attract with, I think like half of the team in New York are immigrants. So we're able to support visas. We're able to bring people from other countries to work here. And we know how to do it because we've been doing it since we started the company. And so it was a challenging thing at first, but now it's become a strength of ours.

I love that idea of turning a challenge into a differentiation point for your company and leveraging that to scale what is now Runway. And speaking of yeses, you've raised multiple rounds of financing from top-tier investors. And most recently announced your Series C at a $500 million valuation with investments from Felicis, Medrona, Lux, and others. First of all, congratulations. That is incredible. And it's just wonderful to see the success for you and for Runway. But we're curious, talking about some of the uphill battles, who was the first person to say yes to Runway?

My mom, I guess. It's like, I'm gonna, I'm gonna start this company. She's like, yes. Do it, just be very good at it. It's like, okay. The first investor, uh, there were a few. I mean, our seed round was led by Lux, Compound, and Amplify. We had a few other term sheets over the table, but we decided to work with them because I think for me, it's really important to work with investors that really understand what you're building and gonna be helpful for the long journey. It's a relationship and you need to make sure you both really understand the scope and what needs to be done in a product sense and in a go-to-market sense as well. Compound, Mike Dempsey specifically, and Sarah, Kat and Sarah from Amplify were the first two folks who were like, okay, we're gonna, we're gonna do it. I think that for me was very important that like they believed in, in, in us. Same with Tavinder from, from Luxe. He now started his own fund Dimension. Shoutouts to the Dimension crew. They were like the three folks that when I showed them, I didn't have a pitch deck at the time. Again, I just had a demo, the same demo I showed that founder I was telling you about. So every time I met a VC early on, I was like, look, here's what I'm building. Do you want to build it with me? That's it. And those 3 folks were like, yes, we want. And that yes, I mean, I got 100 nos. And like most of them were like nos. Like kid from like art schools, like trying to build like, it doesn't make any sense. But these 3 people were like, oh yeah, this is going to be huge. And to be honest, they were actually, Amplify, for example, was looking for us. They actually reached out to us first before we even started thinking about raising because they were like, they knew something was coming in the space. So it was very early. But yeah, I think like all you need is a yes among the hundreds of nos.

That's incredible. Just love the vote of confidence people have early on. And then of course, to see how that carries through with that thread of confidence and conviction in you all. It's an interesting time, right? These are some early believers. And today we're sitting in 2023, we're kind of at the end of last year. Gen AI became the term and OpenAI released ChatGPT. And you're probably, of course, sitting here being like, we've been talking about this for a while. Welcome of the party to the rest of the world. And just for some fun, we thought we'd ask ChatGPT to know what we might actually ask you because perhaps someday that's how we'll create all of our questions. But in that vein, this is a question that ChatGPT wanted to know about what to ask Kris from Runway, which is if you can explain to us the difference between generative and discriminative models in the context of AI and machine learning.

Sure, I'll try to do my best. When we started the company 4 years ago, we were like, hey, we're building this generative AI company. And I was like, Chris, that's crazy. Don't do that. And everyone was like, no, oh, generative AI is going to be so huge. And it is. It's just great to see more people realizing that. I'm happy that the field has been moving forward in very positive ways. In a broader sense, generative models and generative AI are basically techniques or systems that can generate new data out of existing patterns in previous data. So You feed in an algorithm with existing data points and the task is basically come up with net new data points from a distribution. And so what you're trying to do is not duplicate or replicate the existing data. You're trying to understand what makes the data really unique and then come up with new aspects, new data points. And in language, those are just tokens, which could be like characters of a word or a sentence. In the context of images and videos, those are just pixels. You're just trying to predict and come up with the sequence of pixels in an image, for example. That's different from other traditional methods or techniques in AI where the task or the problem at hand is to try to understand the world, for example, right? You're trying to understand the semantics and the meaning of words, right? There's natural language understanding. You're trying to understand what words mean, but you're not trying to generate new words. So generative AI just tries to take a different approach, which is just generate new data. And that data happens to be, in this case, images, videos, or text.

Amazing. Well, an image and video really are worth a thousand words. And right before we hopped on, I had the pleasure of watching your guys's new release of the Gen-1 model. Absolutely huge. Congratulations on that incredible announcement today when we're recording this in early February. Can you just maybe give us a little bit of a first look at what this Gen-1 model means for Runway users and potential Runway users?

Gen-1 represents a very big step forward in generative AI. It's been something we've been working on now for some months, and it builds upon our internal research on video generation and image generation that we've been working on for some years now. Video generation is a task that is relatively new in the space, I would say. Being able to generate consistent and coherent and temporally consistent video has proven to be very challenging, and now Deficient models have proven to be a good candidate to solve that problem. But for us, when we think about video generation, we don't think about video generation as a one-off solution where you type in something and you get a video. Because I'll go back to what I mentioned before, you really want to understand how people are going to use this in the context of storytelling. Try to create a 10-second or a minute-long video in the way you envision in your head using text. It's really hard because there's camera angles, there's blur, there's motions, there's occlusions. Perhaps you want to start with a camera on the lower side and there's so many like details on how you're going to craft an idea. And so for us, it's like, great, if you want to solve the problem of video generation, there could be different approaches. If you want to solve the problem of storytelling and making storytelling accessible, that's a different problem. And so when we take that insight, we quickly realize that there's a better way to drive the generation process of video, which is using other video. So think about taking your phone and thinking about an idea on a camera and a scene, and you just shoot that with anything you have around, the glass, and you put it on your table, and you're like, great, now I want that glass to be a building, and my idea is I want a camera that's gonna be moving across a city. You can just do that and just type in like, hey, turn that video that I just recorded of a camera moving with a glass on my table to a view of an helicopter moving across the city. And so Gen-1 basically does that. It takes your prompt and applies the prompt to your source video and transfers the structural elements of the video that you've recorded to this new video. It's a very unique, novel approach driven again, mostly by our insights in how filmmakers actually work and how stories are actually made and helping drive that process forward. And you can also drive it with images. So you feed in a video, an original video source, and you say, here's the style, the structure that I want to have applied to the video, and then the model will basically take that the same way it does it with text prompts and then apply it to all the frames in the video. And the most consistent, the important aspect is that it's maintaining the proportions, the relation, and the structural cohesiveness of the objects as they move across the different frames. Because our eyes, due to thousands of years of evolution, have managed to really detect the slightest change in a frame. So if there's a frame and there's a flickering, you'll detect it. And it's like, you won't believe the magic of video because there's a frame that didn't match. And so making sure models are smooth and cohesive is really challenging. And so this is like taking that step forward by trying to solve that.

Gosh, I'm just struck by the depth of understanding you have, the more thousand-foot level than you did when you got started, because the team has grown so much around the persona of the user and understanding the depth around what their needs are and how to craft a really very sophisticated technology product around a very specific use case for the entire creative community.

It's about focusing, obsessing about use cases and customers. I think that's really important. Never lose sight of that.

And so Runway has really solidified itself as the platform and tool for creatives, and I'm so excited to see where this model, and I'm sure many more to come, take your incredible business. The space itself is rapidly evolving. There are a ton of macro conversations about open-source models versus proprietary algorithms such as the one that you just released. I'm curious, just generally, how do you see things evolving between cloud services and edge apps over time?

I think we're very early still. I think there's a lot of things that need to be defined and developed and experimented with. I think there's definitely going to be a lot of things changing in the cloud side of things, given mostly that you need a lot of compute power to run models in a real-time basis. But I think it's very early still to define how that's going to play out. I think we're like in the early stages of understanding changes in a stack. I always thought about it as like there's the first wave. There's been multiple waves dating back to the '40s, I guess, on the AI hype cycle. But roughly I would say the last 10 years, from perhaps 2015 to 2022, was like wave 1, which was like models are getting good, infrastructure is getting better, compute is getting more accessible, and datasets are also getting more accessible. And then the second wave, which is I think 2023 onwards, it's just like we're beginning on it. It's great. Next stage is to build everything we need to build as an industry to make these models really reliable and professional and quality, the quality that they require to be used in the context that we're thinking about using them. And think there's going to be a lot of like massive changes on infrastructure, on services, on applications. Products. And so I'm not 100% sure. I think we're really early on, on that journey still.

I'm excited to see where it goes as well. From the investor side of things, if you could put yourself in our shoes, speaking as an investor, how would you think about differentiating some of these companies who are really getting off the ground right now, building on top of some of the models that have been released? Like, is there a heuristic or any advice you would have to investors listening to this podcast on how to think about it?

Definitely, I'm not an investor, so definitely don't take my advice.

This is not investment advice from Gus.

Yes, exactly. Yeah, I should say that, like, whatever banks say at the beginning, this is not investment advice. I personally prefer just builders, like folks who can not only have a vision and an idea of the world, but can actually build it. And I think AI is not different in any shape or form. Just understanding your strength and understanding how quickly you can iterate. On the thesis system and the assumptions of your product. And so one thing I try to look a lot for in early teams for building in this space is who are you building for and what's your like learning rate? How much have you learned about how customers use your thing and how fast are you adapting to it? The learning rate, I think it's just unique to requiring just startups. Like you just need to continue to be learning all the time, but specifically in the field of AI right now, things are moving so fast. That if you build preconceptions around, let's say, one specific model, most probably that model will become obsolete and replaced with another technique. And so if you build your whole service and offering around that specific thing, then when you get replaced by another thing, like if you're not moving fast enough, it doesn't matter. And so speed of execution and like learning rate for me will be probably the most important factors to look for.

Okay. A fun question for you, really wearing our own hat of creators, as we ourselves are very interested in how Runway is scaling seamless editing for video and audio and all of the methods of creation. You know, we actually didn't even realize we've been using GenAI here at the pod, but we've been using Descript for a number of years now, and their audio-to-text has just been huge for our energy and mental bandwidth. So we're just curious, what do you think about the evolution of the creator economy and what does Runway think about just empowering creators of all shapes and sizes?

That's a good question. I, we are, we're a company devoted to storytelling. We do core research and we develop infrastructure and build software, but deeply at the core, we're a company devoted to moving storytelling forward and allowing more people to tell more stories. To do that, we, we develop technology. That's the most efficient and effective way of doing that. I think what the future holds for this space, it's something I'm really excited about, and I think about this often, which is being able to translate ideas into final outcomes in the shortest amount of time possible becomes possible, feasible now. Just with Gen-1, like if I, for example, if I have an idea about something and I wanna record that and I wanna see it live, technology basically allows me to do that really fast. 10, 100 years ago, it was very hard because you just had to actually like either paint it or like hire a team. Write a book about it, right? That's the way you're translating your view of the world, your ideas into like literary paper, right? 50 years ago, it was a bit better because you had a camera, but it was still like funky because you had to go and reveal the camera. It was an expensive process and you're limited about the amount of times you can do it. Now things have gotten better, right? You're just faster. You can create more. I mean, anyone can host a YouTube channel, can start a podcast, can reach millions of people, right? Without having to go through like an agency or a production studio or A manager, you just do it. You have the tools, distribution. You have to have good ideas, right? And you have to be able to execute those ideas really fast and really quickly and be good at it. And this podcast, it's a perfect example of that. You have a great idea, a great set of like speakers and people you're inviting, and you can just do it. Doing this 20 years ago would require you to need an agent and we need like a professional studio and like we need a radio show and like we need money. It's like, okay, I think the next step is there's continuing that trend of AI models and these systems will get progressively better at being able to transfer your ideas into reality. And if your idea happens to be a movie, great, you'll start to see generated movies. And if your idea is to create music, then great, I'll help you create that music. I think that for me is a really exciting part where the human component becomes fundamental in the creative process and these tools are just helping you augment and speed up your ideas.

Well, Chris, we've chatted a lot around the future for AI, Runway, all of the incredible things that you were doing to move storytelling forward. You know, we have a few minutes left, so I'd love to shift gears and focus a little bit more on you personally as a founder. A lot of aspiring founders are listening to this podcast and a lot of them are probably first gen as well. What advice would you give to other first gen founders that you wish you knew when you were starting out?

There's a lot of things I've learned, but I always think about, it's hard to give advice because you always give advice to yourself in the past, and everyone has different like stories and like context. And so I always take advice with a grain of salt. Yeah, that's a very Meta thing to say, but I think just figuring it out, I really feel that's the important DNA to have when building a company. You're, as a founder, you're gonna have to do everything. Like it doesn't matter. I've done every single part of the company and it doesn't matter. You just have to do it. And if you haven't done it, just figure it out. That's it. I think that for me is an important takeaway and it's something hopefully with the conversation you have since have seen has been reflected in how we build a company. And every time you start something, there's always like this risk. There's like, oh, maybe like I shouldn't, it's too risky. I haven't done it. What about X? What about Z? What about W? Like, it doesn't matter. Like, you'll figure it out.

Thank you for sharing that. It certainly resonates. We have the last question of our podcast and it's a question we ask all of our guests. As an aside, we do ask that the answer not be your mom, but someone else that you've encountered in the professional world. And that question is, who is a woman that has had a profound impact on yourself and your career?

I mean, there are two women that have a very profound impact in my life and my career. I guess I couldn't name my mom, but for sure my mom will be number one, breaking the rules here. And my wife. I think my wife has been like a very instrumental piece in everything I've done and I couldn't think of doing anything I've done without having their support. And so nothing gets built in a vacuum and without other people's support. And I think like for me, that's just been fundamental on an emotional level, on a professional level, on an intellectual level. Just having that level of support has been really critical to the success of Runway.

Thank you so much for sharing that, Chris. And thank you for taking the time to be on the podcast. It was awesome learning more and being able to chat with you today.

Of course. Thank you for having me here. And again, this is not investment advice, so Madison McIlwain.

You have a standard disclaimer at the end of the podcast.

Yeah, disclaimer. Exactly. No, but thank you for having me here. It's been great chatting with you.

Thank you so much for joining us at The Room Podcast. If you want more from The Room every week, subscribe to our newsletter at theroompodcast.com/newsletter. We'll be back next week with a new episode Tuesday, 10:00 AM Eastern, 7:00 AM Pacific. See you in the room.

All opinions expressed by Claudia and Madison and podcast guests are solely their own opinions and do not reflect the opinion of The Five EC. This podcast is for informational purposes only and should not be relied upon as a basis for investment decisions.