Making Startup Finances 10X Easier & More Powerful - Siqi Chen, Founder of Runway

Hey guys, welcome back. Today I am hanging out with my buddy Siki Chen, the founder of Runway and a bunch of other things in the past too. I mean, we go way back.

We do. Yeah, it's been 13— geez, how long? I want to say—

'09, '10?

'07. So that's 17 years.

Yeah.

Yeah, I knew him before he was famous.

Well, Siki, you are, uh, you know, sort of the head of one of the companies that had one of the best B2B SaaS launches, I think, in recent memory. Both like sort of first wave was on your website design on runway.com. And then now more recently, you sort of set X ablaze with a very sick swag launch. Maybe we start there. Yeah. I mean, what is Runway? And then I think one of the things that's really cool is the fact that you can actually still do a launch for something like B2B SaaS, but something that just gets a million people engaged and realizing, oh, this is a need that I actually have that I didn't even know I wanted.

So there's no rule that says you can't do consumer marketing with B2B SaaS. But yeah, so to start, what Runway is, is we think of it as a business tool for thought. And so we are in what's traditionally called the FP&A category. And so companies have these spreadsheets that they use to run their businesses, and we just make a very, very good version of that so that people inside and outside of finance can actually understand how the business works and collaborate on a good plan together. So the vision is a designer or a PM or an engineer can say, if we work on X instead of Y, here's how it affects the business financially, margin growth 2 years down the line. We've been very design forward, and one of our beliefs is that finance is not just about finance. It's about everything that happens in a company, everyone in it. And that informs the way we do marketing because the way we think about marketing is every person who works in a company, whether they're a finance executive, any other kind of executive, they're a person too, right? They buy shoes and they buy Apple products. And so why can't you market a software product that is for business the same way that every other consumer product does? And that's definitely a lot of the thinking behind how we did it.

Having run other businesses before, you know, I remember investing in your last startup in addition to this startup.

That's right, yeah.

You know, one of the things you realize as a founder, I think, is that money is actually sort of like blood in your body. Literally, it enables things. You actually have to make maybe thousands or tens of thousands of tiny decisions that add up to fairly big decisions about who do you hire, Where do you put resources? And then whether or not those investments go well or poorly determine whether or not the startup itself, the organism itself does well or poorly.

Yeah, I think that's a really good description of what operating business is. And I think what we found surprising is that the way people think about finance seems like something mysterious, right? It's siloed, it seems complicated, opaque, it's in these spreadsheets. But at the end of the day, that's what finance is about. What a spreadsheet model is, is if you think about it in video game terms, video games have a game engine and it simulates the workings of a park or a civilization depending on the game that you play. And that's at the end of the day what a model is. It simulates a business so you can understand the impact of the decisions you make today in the future. And right now there is a deep disconnection between the spreadsheets and the finance department with people who run businesses, like me as a CEO, or you as an investor, right? It's not super accessible, and we think you need better tools for thought to make it accessible, so that anyone can just play with the model, play with the plans, and say, we really should be investing more in this department instead of that department, because here's what's gonna happen.

Yeah. What's sort of the lifecycle prior to Runway? I feel like it's pretty haphazard. Yeah. You know, when you say operating model, it's initially just sort of a bunch of crude swags in an Excel spreadsheet someplace. Yeah. And then the weird thing is when you look at enterprises that get to, you know, 10, 100, several hundred million dollars a year in revenue, sometimes it never really progresses beyond a slightly less crude spreadsheet.

Yeah, it is shocking to learn what the state of the world is as I got into this space. But yeah, just going through the lifecycle of a company, when you start a company, you probably don't need to create a business model unless it's like very complicated or it's a low margin type of business. You're just trying to get to product market fit. The reason why most founders end up creating a model is because some investor who's not you is asking for it. And it usually doesn't map to any notion of reality. Where this becomes more useful is when you get to product market fit and you need to scale the company and you need to figure out, okay, how many people can I hire? How much money do I need to raise? How do I, how much do I spend on marketing? And like, what is my return on that marketing? That's when the models become a lot more useful so you can make these trade-offs. And usually everyone starts with some kind of spreadsheet, either Excel or Google Sheets. And what happens as a company grows is that becomes increasingly unsustainable. I actually think it's the right place to start because spreadsheets are amazing. They're like one of the best pieces of software ever made. What they don't do is they don't scale with companies for 3 reasons. The first reason is as a company gets bigger, you just have more data across more data sources. So you might start with like a QuickBooks or an HR system, Pretty soon you have analytics data, you have CRM data, you have sales data. And eventually the time that a team takes to update that model every month just becomes way too time-consuming. So you need something to automate that. The second thing that happens is the model itself becomes more complex. So you start with a single product, but then eventually you have multiple products, maybe in, or multiple countries, geographies, maybe you have multiple departments. And all of a sudden it's a highly dimensional problem and it's very difficult to model that thing in Excel. And the third thing that happens, and the thing that we help with the most, is you just have more people in your company. And when you have more people, you have more problems. And most of those problems are around communication. So you get more information from people, what their intents are, like what are your plans and how do you trade off the different departments and their plans against each other? And then you need to get more information out to more people, right? Here is a strategy of the company, here is why, here is the trade-offs. And all of those problems become very difficult on a spreadsheet. Like a simple example is, let's say you are a very progressive CFO or CEO and you're saying, okay, I want everyone to see the business model and plan how it works. You can't share your spreadsheet. Why is that the case? Because you can't even hide a single column that says salaries, right? So fundamentally you need very fine-grained permissions. You need to make sure you're not doing all your calculations on the front end for this to even work, which then creates pain and that creates an opportunity for something like Runway.

One of the things that's funny as an analogy is that, you know, when Figma first started, people sort of didn't really even understand what it would do to design. Because again, it, you know, instead of an Excel spreadsheet, it was like a PSD file. That's right. Like literally people were still using— like when we came up, we're all, you know, it's gonna make us sound super old, but believe it or not, people did like what I remember doing this at Palantir when I was employee number 10, like literally sitting in Photoshop. It was like a photo editing piece of software. Wow, yeah. I used that for pixel-perfect mock renditions of my bank. It wasn't that long ago. Yeah.

That was the state of the art. Yeah. You know, that reminds me of like one conversation I had with another founder. I think it's really difficult to see the possibilities of how tools affect behavior, right? There was, I think, the default stereotype is that, oh, why would finance want something where They can be more collaborative. This is what actually what the most progressive finance people, they need because as the world becomes more competitive and fast moving, they are under pressure to be more strategic and to be more collaborative and the tools aren't keeping up. So Figma's a great example because, you know, a story I like to tell, and you'll remember in the first pitch I described this as a Figma for finance. Sure. What I like about Figma is people forget that before Figma, it wasn't just like a slightly better design tool in the web. it was considered kind of rude to look over the shoulder of your designer as they were working. But now everyone is in Figma, whether you're a PM or engineer or a business person, you're collaborating together. It's made design better. Amplitude is an analytics platform, which I was an early investor in, and it is obviously a great analytics tool, but people forget that before tools like Mixpanel and Amplitude, you had to go through the data team to get access to this data around retention and growth, and it was considered semi-sensitive, but now we just take it for granted that everyone can be in a company and understand what retention is, and conversion rate is, and growth is. What is that equivalent for the entire company, right? For the company as a business, that is a thing that doesn't exist, not for lack of trying, but we think the thing that will enable that is a better tool for thought, which hasn't been built yet.

So I wanna talk about even before this sort of launch with this really sweet jacket you sent a bunch of us, and there was the website runway.ai, Yeah. That you put out. And what I hear now is in the Valley, people now have 2 websites that they like to copy. One was gonna be always linear. Linear, of course, yeah. And then now Runway is like— That has happened. A brand new, like, it's a different design language. Yeah. It's, you know, sort of instead of minimal, it's like almost maximalist in some ways.

Yeah, I mean, so I think so much of what people do in business is what other people do. And so the reason, our default assumption is that the reason why enterprise website looks the way it does is because it works better. I don't think that's the case. I think one, it's most likely because, well, it's what everyone else is doing and you assume that it works. And 2 is precisely because everyone else is doing the same thing, that's where there's an interesting opportunity to do something different. And so that's definitely part of our motivation is we're not afraid to do something new because we think that has not just a 10% better outcome, I think it has a 100 or 1,000x better outcome if you do something truly new. The other thing that informed our website was what we were about as a company, which is we think what is required to build a disruptive product in this space is a consumer experience. And if we want to have a consumer experience, then our marketing and website needs to reflect that, that this is a product for people. Because people work inside companies no matter what your job title is. So those are the 2 big sources of inspiration, and then we just spend way too much time on it.

I guess, what would you say to founders who want to do that type of launch? Obviously, there's committing some amount of money to it, like being willing to pay more than what you would pay Fiverr. Sure, sure, sure. There's that, but then there's also taste. Walk us through that process.

I mean, I think the thing, the mistake to make is to do it because someone else did it. Mm-hmm. It's not going to hit the same way the second time, right? Just like everyone knows about a Linear website, nobody knows about the 200 Linear websites that look exactly like Linear. So it does require taste. I don't know how you develop it, but it is somewhat a requirement. But like taste is true to a single person, right? I think like one of my favorite websites is, Patrick Collison's website or Nat Friedman's website. It's just plain text. The browser company had a very different design language as the SF Compute company. They just have plain text websites. It's interesting the first time you do it. And so I think the thing to think about is what is true to who you are, what you're doing as a product, and try to do something new and do not be afraid. A lot of great ideas that are new don't get done because, well, No one else has done it. And yes, no one else has done it, but if you do the exact same thing everyone else has done, you're going to get the exact same results.

Think from first principles about figuring it out.

First principles is key. Yeah, you want that in a product, and I think you want that throughout everything that you do, from your marketing to your culture. We have actually only one advisor to the company, and it's Wolfgang, who we both know.

Oh yeah, Wolfgang Hammer.

Yeah, Wolfgang Hammer. And It's been magical for us. And I think taking a page from Lulu Cheng, this idea of knowing what your story is and having that same story be the story that you tell investors, as your team, as your customers, as the wider world, is so important. If you start with that story, everything else follows.

I guess that really comes out in the launch video that felt more like, I mean, almost like a movie trailer or something. It just immersed you into Oregon Trail. I mean, a lot of your customers are probably our age, our generation. So, you know, sort of hitting the nostalgia button was a really good move there.

We spent a lot of time refining the runway narrative of Wolfgang. Like, I mean, hours and hours over the course of almost half a year, I think.

And he came over from the film world.

Yeah, so Wolfgang Hammer is the executive producer of House of Cards and Hunger Games. and was a studio head of Lionsgate and Miramax and CBS. So extremely knowledgeable person. And we were working on that story for quite a while. And one of the outcomes of the story is we have a really good idea of what the default world is. With every story, you want a really great villain. And we spent most of the time, and the most difficult thing is coming up with a very believable villain. Because as a startup person, You're working on what you're working on and it's just better. And so whenever you tell your story, your default is like, well, the world sucks and our product is way better because of X, Y, and Z, but you don't end up with a believable villain. What's really hard is being able to tell the story of what already exists in a way where it sounds amazing. That's perfectly reasonable. It's like, I want that. And if you have that and then you can point out the subtle lies in that story, your own story becomes so much more engaging and believable. So spend so much time doing that.

I guess what jumps out at me is you sort of personified maybe the buyer, like the CFO, the C-level finance person, as like a character in the video game. Yeah. And then there was a sort of game board that was like scrolling across, and then crazier and crazier things were sort of, you know, happening to that game character.

Yeah. As Wolfgang put it, I mean, one of the things I think about, he has so many quotable quotes. One of those quotes is like, all stories are about one person. And in the startup world, that story probably should be about your customer. And so we really tried to figure out what is the story that makes this character a hero, and how can we tell a believable story about the world as it is and the world that it could be? So we spent a lot of time figuring that out. And one of the analogies that felt right has ties to my background. My background is a little bit in gaming, and I do think of Runway in some ways as a game engine, right? Like a company, and this model is in some ways a software simulation of business. Just like if you're playing RollerCoaster Tycoon, there's an underlying simulation of an actual park, right? And it allows you to simulate the future and push buttons and see what happens. Running a company doesn't feel like that today. And it really should. And if you can build good enough software that maps closer to how you think about your business, connects all your data, we think that's possible. And that's what we think Runway is. And that's the story we try to tell. And that's why we chose the video game motif for the video.

Yeah. Let's jump back to product for a second. Sure. Because one of the things it was cool to see over the course of years building this thing, really early on, you spent a lot of time trying to figure out what the correct nouns were Yeah. As in, like, what are the— what— how do you model what you're doing in a way that's, you know, low bar so anyone can start using it, but high ceiling so that, you know, it's as expressive as a spreadsheet and, you know, in software, which is pretty— like, actually, those 2 things are sort of like the enduring really, really hard thing to do in UX. Yes. So for actually the software to be something that people can use and then use over the long haul.

Totally. So I think this was a key part of our product strategy from day one. So I often say Runway really wouldn't exist if Notion and Figma didn't exist. I think those products have really proven what you can do if you have the right nouns. We call them, I think for a time we called them primitives or abstractions, but what is the fundamental unit of the product? And so if you look at a product like Notion, it's all blocks and it's very expressive and you can rearrange them and it ends up being an incredible tool for thought. Figma is a very similar concept because one of the core ideas behind Figma is the idea of components. You have components and master components, you can group them and you can nest them. And what these things do is they expose the power of software engineering principles like inheritance and polymorphism visually to the end user, which is incredibly powerful. And so we spend a lot of time trying to figure out what are the right abstractions for Runway. And there are some existing art around this particular problem. So if you look at a a product like Airtable. What that product really is about is instead of having the primitive being a cell from a spreadsheet, it made a primitive a row. And on top of that, you can build a very interesting platform, very powerful workflows. And so for the business planning use case, we had to have at least like 2 or 3 more primitives. So we have the idea of drivers, right? A concept of price or runway or headcount. All of these are rows in a spreadsheet. That's one of our primitives. The other primitives I mentioned, right? So what are different departments, different product SKUs that you have? And a third one we have to invent from whole cloth. And this is one of the key ideas behind Runway, which is when you think about you as an operator, how you think about your business, it's not in the form of a wall of numbers in a spreadsheet. It is in the form of a product roadmap that's going to improve your product in some way, or a marketing plan. Or are you going to have certain campaigns and how much are you going to spend? Or a fundraising plan. All of these things usually live outside of a spreadsheet. It lives in places like Notion or Google Docs. The way it's represented on a spreadsheet is let's say you have a row called price, and the formula is like whatever it was last month, that's my price. And the way you would model a change in price is let's say you want to change your price into New Year's Day next year, you would just edit that cell and the price would go from $500 to $1,000. Say, that's a discontinuity. That change is a thing, right? And if you group a set of those changes together, they become roadmaps and projects. And so that change is a native object in Runway. So now you can connect things like your Jira to Runway and say, if we work on these tickets, here's how it affects the KPIs, the conversion rate, and how it affects our margin down the line. So having the abstractions right is so important. It's a foundation, I think, of a tool for thought. And if you get them right, that's how you make something both accessible and flexible and powerful.

I think that that's something sort of underreported or underappreciated in startup land. Everything is so task-oriented as opposed to sort of getting the right modeling allows all of the tasks to happen. And if you choose the wrong abstractions or literally the wrong models in the lower level, you just literally can't do a bunch of things.

If you don't think about it right, and I think you'll definitely appreciate this, you're a designer. This is an expression of the continued undervaluation of design as a discipline. The iPhone was out there. You could have built that, but no one thought of this. This problem is like, if you can design new patterns of interaction, you can build something very powerful. I think the same is generally true in SaaS software. Everything is like workflow-driven. What is a customer problem? Okay, let's build a thing that solves that one problem. But we don't think about like, what are the components of this product such that you can solve like many problems in a way that's easy to use? And that's very much a design discipline.

So in terms of what you've done in the past, I mean, there sometimes isn't a connection. You know, it's sort of hard to go from gaming to, you know, one of the more serious use cases in business, period.

Yeah.

But it's also something that a bunch of our, you know, our friends have also done. Like, Justin Kan has done this. Imad? Yeah, yeah.

From Mercury? Exactly.

Yeah. How did that come about? I mean, what was— what were startups and sort of the idea maze? What did that look like in, you know, 2007? And then—

2007, man. I was a year out of college, right? I just wanted to build something and build a startup. So funny story, I was actually Y Combinator's first dropout. if I'm not mistaken. I don't know if you want to cut that out or whatever.

Oh, it's on there. The spicier the better.

Yeah. But yeah, I mean, I have been and remain an amazing Paul Graham fan. Like, his essays have changed lives, mine included. And I applied to Y Combinator winter of '07, and I got in, and I dropped out, which we'll get into because that's related to First Company.

Was Chatteris in there with Wilkie? I think I was Chatteris.

Yeah,

yeah, yeah. I think— You got to know them later, right?

No, I think we knew them. I knew them through mutual friends, but actually with Wilkins, yeah, I mean, his family and my family would go skiing every Christmas. Yeah,

totally.

But so at the time, the Facebook platform just came out. This is mid-'07. I just moved here for a job at a company called PowerSet, which is a search engine startup. And the side project of a Facebook app blew up. And the first thing that blew up was, If you played Werewolf or Mafia, it's that but for Facebook. That's my first world's app. Oh, that's fun. And it got to like 100,000 DAUs. I was the only person maintaining it. I became known internally at Powerset as, my nickname was FB Mills for FB Millionaire, 'cause I was making $3,000 a day from ads.

Nice.

On that app.

That's so funny.

So that's how I met my co-founder, Alex. So he's my co-founder of my first company. He eventually ran product at Reddit shortly before the IPO. And he was like, let's make something together. So I thought, okay, well, let's build something on Facebook. Then we built an app called Friends for Sale. And what that is, is today it would be called BitClout, it'd be crypto-driven. But the idea was you turn a social network into a market economy. You could trade relationships. And that became at one point the second largest app on the platform. It was many times bigger than Twitter. And that was the first company. And simultaneously with that, I got into the winter batch of YC with a different team, with a different product, and this one's taking off. And in the background, I talked to Alex and I talked to my team and was like, hey, maybe we should just merge the companies. And Alex didn't want to do it. And I get it because we were blowing up and it's like, why would you join YC? He wasn't much of a Paul Graham fan at I was. So the bad thing is I just stopped showing up 2 weeks after Intermatch, and I started working on this. We were raising the Series A for Serious Business, which was what the company was called. So we ran that company for about 2 years. We built a bunch more games. It was a profitable company, about 40 employees, but the 800-pound gorilla in the space was Zynga.

Oh yeah.

Yeah, and Mark Pincus is an extremely experienced entrepreneur, Knew what he was doing, very aggressive, and he ended up acquiring a serious business. And I joined as first a director of product, eventually ran all the product for Zynga until the IPO. And I learned so much from that experience. So that's how I sort of accidentally got into games. I never thought of the thing that we made as actually a game. I thought it was thought of as a social network. But Jeremy Liu funded the Series A in that company. And he said, yeah, there's a new category of games. I call it social gaming. I think it's a big category. I passed on Zynga, and I think you guys are gonna make it. That was not a good choice. But yeah, so that's how I kind of got exposure with games. But that was never something that I thought I was good at or wanted to do.

Yeah, I mean, for the current generation, all they know is sort of the AI hype, and they're like, can I trust this? Should I work on this? Like, did I miss it already? Oh my God. And it's like, no, dude, you didn't miss it at all. Like, it's just getting started. I

think we're missing it. I wish, like, yeah, like, this is 2007, right? Like, there's never been a better time to be a founder.

How do you think about AI being a part of this? The interesting thing about what you have is you actually have the right abstractions. And then the wild thing about large language models is they are sort of exhibiting emergent behavior around reasoning and logic. but only in very small, sort of constrained little sandboxes of a constrained context window and a constrained prompt in a workflow.

Yeah, I think it'll get there. The models will get better and they'll be able to reason and use tools better. I think the current batch of, like, agentic type of companies, they're not quite there. I think they had the right idea. You just, like, replace it with a better model. It'll get smarter. It'll be able to self-reflect. So I'm, generally bullish on that direction. I think there is a tremendous amount of opportunity around the design and expression of AI in products today, given the current capabilities of LLMs. And so I think the default way to think about that you've seen products use AI is to treat AI as basically a creature external to the user, like an independent entity. So Agents in general are examples of that. I think agents by and large with current model capabilities is still pretty unreliable. It's very difficult to get productive use out of them. And the other expression is chat. And inherently chat, it's like you're talking to an external entity, right? But if you look at the most economically productive use expression of AI today, it's something like a coding copilot. And that's very different. What that is, is it's a better tool for thought. You're doing work, and it's just like helping you in the background, autocompleting things, inferring what your intent is. And a lot of what Apple released in iOS 18 reflects those same ideals, right? Like you have a notepad, you're just writing, and it's like, oh, well, we can do the calculations for you. That's how we think about AI Runway.

Yeah.

Less as this agent and more as like, you're doing work already, we know so much about you. And so we try to not make it, you know, a sparkling emoji and like this character that you talk to, and far more like it's embedded into every character and every pixel is how we think about this.

When I really reflect on it, like, I hate when people call it generative AI because it's just the wrong mental model for what these things are. Yeah, true. Like, I think the better model is, you know, I spent time with Jake Heller from Casetext. I'm sort of like obsessed with, them as an example because they were sort of the Instagram-esque, like, $600 million giant exit, like, right at the beginning of the boom. Genius. And he got access to GPT-4 first out of, you know, pretty much everyone else in the space. And then he, you know, had to pivot the company into, like, putting all the chips on that. And what he figured out was that you could, like, tailor time and motion study what any, like in his case, a lawyer would do. Yeah. Basically replicate it in the workflow itself. So when you say generative, people think, oh, like give it a good question and have it generate some sort of answer. And then it sort of misunderstands like the power of like even a single request. Like if you put in the right amount of context and ask a very specific question, you can get ranking back that is like, is it relevant to— I am looking into— I want to make a chronology, and I'm interested in person X who did this deposition. If you feed in specific parts of that deposition, you can actually rank all of the things sort of chunk by chunk, and then it's the equivalent of a human being highlighting the actual specific text. Yes. So what I think Jake figured out that I think the rest, it's been more than a year, a year and a half since you figured this out, but I don't think people have figured it out really yet. We're starting to see it. That's right. There's a company in the current YC batch that's going to sort of replace all people working on accounts receivable. They already have a customer that had 12 people working on it, and then those 11 people are working on other stuff in the company because they only need one one person to sort of run this thing that does the work of 12 people.

Yeah. I'm reminded of Joel Spolsky. He worked on Microsoft Excel back in the day, very famous blog, right? I think he's the founder of Stack Overflow. He had this really interesting observation way back in the day. He described talking to Google engineers and how it's different from talking to Microsoft engineers. He said, Google uses Bayesian filtering the same way Microsoft uses the if statement. And that's the way I think you want to think about AI is like, what if you use LLMs the same way you use an if statement? You can make unstructured data structured. You can rank fundamentally unstructured data. So there's some companies that I've seen where let's say you get 10,000 applicants and you have certain qualification criteria. How do you filter them? Well, you can actually rank them now. And so one of the expressions we have at Runway is Runway powered by AI is going to sound as ridiculous as Runway powered by AJAX. pretty soon, if not immediately. So we don't want to think about it as something external. It's just something that you do all the time. It's just built into software nowadays. It should not actually be, it won't be a differentiator very quickly. It's just part of the baseline expectation of how software works.

It's funny because there's also the reverse belief out there. If you talk to Parker Conrad, he's like, we're just gonna focus on building the best software we can. They don't have specific things that are like bet the company on large language models yet.

I think that's the right way to think about it, really. I mean, and I think that's the way we think about Runway too. The temptation is like go straight to, okay, the AI is going to do all the work with existing tools, but I think you get tremendous leverage by building better tools. It's better for AI and better for people too. So a real example is we have a an abstraction, a noun called plans in Runway. And this allows you to connect things like strategy and intent with your numbers for the first time. Without this abstraction, you can't actually put intent into a spreadsheet. But with the right abstractions, you now have a tool to reason about things, manipulate them, and it can hold a lot of context. All of these make it way more useful for AI to do productive work. So I think figuring out the right product that can contain context and workflows such that the AI is more powerful is actually not a bad approach. Yeah.

I think the moat is actually getting the model right and then actually being embedded in real customers. Yes. With real data.

That's exactly it. That's exactly it. Yeah. And it's just hard to skip right to the end. AI isn't gonna solve that for you. You still have to build good products and good tools for thought.

Yeah, it's funny. Yeah, basically all the AI-first companies, still have to beat the incumbents.

Yep.

And then sometimes, you know, I guess that's, you know, the cautionary tale that when people do like the pure AI-first thing is they're able to get the first customer, and the first customer might be paying them $10,000 to $100,000 a year, and they might even get half a dozen people, and then they'll go and raise a great seed round. But then, uh, the incumbent sees it, realizes, you know, hey, that startup put that together in about 6 months. So, Skunk Works project with better model, better abstraction, better UX, and actually embedded into how people already work. And they just sort of take back all the customers. I think that's right.

I mean, the game fundamentally hasn't changed. You still have to like build something really valuable, new, disruptive to the customer, right? And I think a lot of companies, because AI is hot, it's like, well, we have AI, what can we do with it? As opposed to going in the opposite direction, which I think is more fruitful and sustainable.

So, Sikhi, this is gonna make us feel old.

Oh, God.

But one of the things that's kind of funny is I think that you and I are about the age Mark Pincus was back when both of us were just getting started. It's been a decade and a half of, you know, going on 2. What advice do you have for sort of basically us who are like 22 now? I mean, That's the wild thing. We just did Startup School on Saturday. Wow. And it was all like, you know, 18, 22, 24-year-olds asking like, actually just stuff that I didn't know at the time. And so the big thing I think is like, it's not different. It's sort of the same. It's actually better.

Man. Well, overall, I think my career could have turned out much better if I was way smarter, but I also really value the sort of twists and turns that got me To where I am today, I think the number one thing is get started. I think that's what startup school does, and that's the one thing I did right. I think one of the philosophies I have is from Scott Adams. He told a story about, it's very difficult for you to be the best writer in the world, to be the best anything in the world. It's very difficult for you to be the best artist in the world. And he looked at himself and he's like, I'm not the best artist in the world or the best writer in the world. I'm okay at both, which means I can be the best comic strip artist in the world. Right? So the intersection of skills, finding a thing that you can be really, really good at is interesting. And so one of the things I did in my career, which I wouldn't necessarily recommend to everyone, but it worked for me, is I always looked for experiences where I will be able to add a new circle to that Venn diagram of my skillset. So the thing that led directly to Runway, as you know, is I joined a company called Postmates who acquired my second company. And I joined that company because I thought, okay, this is interesting. We're moving atoms around. I've never dealt with atoms before in any of the companies I worked on.

Keeping track of atoms. Keeping track of atoms. Is even more important because you have unlimited electrons, People try and take your atoms all the time, obviously.

Yeah, that's a bit of the laws of physics, right? So that led to me joining a company called Sandbox VR. And that company was interesting is because not only was it about VR, but it's in the real world and you have to construct and operate retail locations. I had no idea what I was gonna do with any of that, but I just thought, how many people with my background will also have that experience? And that actually led directly to Runway, because as it turns out, if you run a company like that, that's very operationally intensive. You're hiring part-time workers, you're constructing stores, you're building hardware and content. You have to really be able to plan and manage your operations extremely carefully. So we put together a financial model, and a lot of those learnings, that experience led directly into Runway. So having a diversity of experiences, at least for me, served me really well. And maybe the final thing is just as a leader, being comfortable with the role of being a leader has been one of the big breakthroughs in my growth in the past couple of months, if not years. For a long time, I felt like an individual contributor in the body of a CEO or a founder, and that didn't serve me well. I think Having very high self-awareness, ultimately, the operational capability of your companies, it's capped by how self-aware you are. Indexing extremely high on that, meditating, getting coaching, is extraordinarily high ROI in terms of time investment and money investment.

Yeah, I can agree with that. I am astonished how bad a leader I was earlier in my career. Same. Especially the first time I had to manage someone.

Same.

I think the first designer I ever hired was After one of my crit sessions with him, he was like, that was the most disrespected professionally I have ever felt. And I was like, whoa, okay, whoops. It's not my intention.

I bet he learned a lot from that, 'cause I've been in meetings with Mark Pincus, and product managers would literally break down crying in a meeting. I think maybe that is another thing that I think served me well is I feel like I got as much out of tough situations as I got out of really good situations. Totally. You learn just as much. And so having that outlook is really helpful.

Yeah.

Right? Nobody wants to be the complainer, 'cause you don't really get much out of that.

Yeah, I guess what I've tried to learn over the course of years, and it took a lot of work, and actually working with a coach, even role-playing difficult conflicts, I was really disrespectful in that meeting because I actually allowed myself to sort of blow my top. It was like— Yeah. I allowed, like, too many things to sort of just happen. I sort of enabled too many bad decisions in a row. And then at some point, it just exploded. And then that's actually, A, being a bad boss. And then B, even if I wasn't their boss, I was like a terrible teammate because I allowed us to just, like, ship, like, things that I didn't agree with.

The wild thing is the opposite of that is also bad though.

Right.

Right? Like, at least like you were angry enough to like tell people how you feel, which is like worse than maybe a lot of other founders can be conflict avoidant.

Oh yeah.

Right? That happens quite a bit, especially with younger founders, and you feel a thing, but you're not direct about it. One of the frameworks that helped me mentally is when you feel these emotions, like one of the difficult things about being a founder is that you feel these things and you think you're supposed to feel them. What I mean by that is, let's say a customer is unhappy. What do you feel? You feel really unhappy, you feel guilty, you feel angry. And you also simultaneously think, well, I am the founder and CEO. It's good for me to feel these things. And what I realised, what really helped me is thinking about why and for whom am I actually feeling this? And what I realised is that I actually was not feeling it for the customer, or for the company, for the team. I was feeling it for me. Hmm. Because that's how I thought I was supposed to feel. And if you ask the teams, what is most helpful for us to solve the problem? It's not that. And so changing that, it's like, well, what is the right thing for me to feel and do on behalf of the team in this moment? Made me feel different things.

Can you think of a way that you've changed to sort of approach that in the right way?

A lot of what I did before is I felt frustrated about a thing, I would express that frustration. Oh, actually, real example. This is actually not too recently ago. So a customer was experiencing performance issues with their instance and our head of engineering expressed guilt and responsibility. It's like, man, I feel really responsible. I feel like I'm holding the company back. And before, the way I would react to that is like, good. Okay, I trust you. You're a very responsible leader and this is good. You're taking it very personally. That's not wrong, but now I look at it a little differently, and I realize that, well, what is this person feeling this thing for? And I don't think it's behalf of the team, right? He's feeling it for him because he feels bad personally, and there's a part of his identity and ego that is mismatching with reality. But what does a team need from this leader? A team needs support, clarity, Not frustration. And so I told him, hey, I get you. Think about this though. I'm the CEO and I feel guilty and frustrated and all these things. How does that help you? He's like, no, it doesn't help me.

Right.

Exactly. So how do you think you feeling these things help your team? He's like, got it.

Yeah.

Right? So it affects how I coach people, which is most of what I do now in terms of managing a team.

Sort of reorienting to, you know, what sets us up for success as opposed to, you know, what we just might be feeling on the inside.

Yeah, when we, like, go down inside and make it all about us, that's when bad things happen.

Interesting. I guess that's the match pair to companies are actually sports teams and not families. So in a family, like, unconditional love, charity, like, forgiveness. In, like, you know, a workplace setting, We need to be championship warriors. We need to be Lakers or Celtics in their prime. What's going to happen if you miss the layup? Yeah, that didn't help us win.

Right. I think it is a little bit of both, right? I think even Netflix, they said, especially, we're not a family, we're a team. And they had to dial a lot of that back. People didn't feel safe, right? You look at the newest version of the culture deck, they'll say—

Did they get rid of that? Oh, they got rid of that. Fascinating.

And some of it, yes, has to do with scale, right? And they say like, yeah, we had to dial down the language a little bit around freedom and responsibility because like we're at a scale now where it doesn't quite make the same sense.

Yeah, when you're small, you can still do that. Yes. I guess you can be literally in the room to like keep it from going off the rails.

Right. Like one of, I think one of the mental models I'm still trying to crystallize is, I used to think that it is a zero-sum trade-off between giving the team autonomy and being a strong leader and giving clarity. And so you have these examples where people who are strong leaders, like, this is the way it's gotta be. And we celebrate that. Steve Jobs is stereotypically like that. And where do you find the right balance is some of the art and how I thought about this. I think about it a little differently now. At Runway, I feel I've never felt a stronger leader than I have today, and the team has never had more autonomy. I think when you're able to have strong clarity over what is important, what the company is about, what the story is, where you're going, the team can maximize their autonomy. So it's not anymore a zero-sum trade-off. Actually, they can go hand in hand. The better you are as a leader, the more autonomy the team can have. And I think frameworks that are simple don't capture that nuance. I think of a company and a team as a company as a team. It is not exactly the same as a sports team, nor is it exactly the same as a family. People go to work, they get paid, they expect to have careers here, and we are a company that has really high standards with a clear mission and a clear culture, and both are true.

Well, these are all things that we wish we knew.

Things that we wish we knew. Yeah, I mean, The great thing is you're still learning. I'm still learning. I feel like half the things I'm saying I probably learned in the past year or so. And that's a wonderful thing about Silicon Valley. You never stop learning. You never stop growing. If I could look back when I was younger and you were younger, it's like, that's okay. There's a lot of shame and insecurity attached with not knowing a lot when you're younger, but that's just normal.

Yeah. Okay, I have an interesting one because on Twitter someone was like, They're, you know, week 1 or it's like week 3 of YC. And they're like, okay, my day-to-day is I go to— I'm at home and then I go to the gym and then I go to YC and I go home and code. And then, you know, how do I take advantage of San Francisco? And I was like, don't do anything else. Like, maybe you should also go and talk to your customers or go to places where your users are. Yeah. But like, don't do anything else. And the reason is Literally, the median startup fails, even in SF. Like, the median startup is a total frickin' failure. And then I think the thing that is a little unsaid or unseen is that, like, all the builders end up getting to know each other, mainly through their own creations, I guess.

That's right.

Because it's like, I think you use Postgres, and I mean, a bunch— basically, it's like a bunch of my best buddies today, like, I met them by answering their support requests or something. That's right.

Yeah, yeah, yeah.

It seems like all the Gs end up meeting each other at some point in the narrative.

That is so underrated and so difficult to see when you're in a moment, right? When I met you, you just left Posturas, or you were still working on it for a little bit. And Gary was just another YC founder and a great friend. Anish, who we both know, is now a general partner at Andreessen. The peers that you have around you who are building things, who are smart, are going to be the ones who are going to be in chairs like Gary's and everyone else in 10, 20 years. So careers are really long and these relationships matter. And they matter not because you went to a party with them, but because you built something interesting and you're working together. Those are relationships that really lasted.

Yeah, I guess the harsh thing to realize was like, between the two of us, we probably met like tens of thousands of people in the startup world. Oh my

God, yeah.

I don't think this was that conscious per se at the time. Maybe working backwards, that's how it worked. It was like you would meet someone and you'd be like, oh, that person's a G, you know, that person's a gangster. Like, you know, they, they made something like fucking cool, or like they're a good designer, or like there was like some output to the thing. And then it was just like a game recognized game thing, and I just didn't There are tons of people who you'd meet and it's like, man, what's going on over there? And then you just wouldn't, like it or not, you're just so busy. So you're like, all right, I'm just not gonna spend time with that person.

Yeah, I think that's one thing that's more difficult today maybe, because in 2007, YC wasn't what it is today. And so the people in YC and Jason in YC, They were different. No one was writing a billion-dollar checks into .ai anything at a time. So you really had to—

It's noisier now. Yeah,

it's noisier. If you're in YC, you're in there for a very specific reason. You believed in this philosophy of Paul Graham that you can compress your working life into very few years and build something that people want. It's more difficult to suss out today, but I think that's what you want to look for. That's also how we hire. You want people who are doing it because they really care about the thing they're doing, not because of like, now YC is like 100 times getting into Harvard, right? To have that validation is very different. So that's what I would focus on. Why are people doing what they're doing? They have to do it. It's part of who they are.

Good stuff, man. So call to action.

runway.com/launch. Check it out.

All right. And everyone, who can use it?

So we target companies from 50 employees to about 1,000. Or so. So some of our companies include AngelList, Superhuman, Lambda AI, ConvertKit. And if you're smaller, once you get bigger, Runway's here for you.

All right. Thanks for hanging out, man.

Thanks for having me. Good times. It's been awesome.