Matt Arbesfeld:
Hey, everyone, and welcome to PodRocket. My name is Matt Arbesfeld, and I'm the CEO and co-founder here at LogRocket. And today, we have a very special guest, Dylan Field, who's the CEO and co-founder of Figma. Dylan, I think I've probably known you for 10 years at this point. I remember seeing a very janky prototype as you were trying to recruit me to drop out of MIT, but maybe you could start off and share a bit of what Figma is for the audience.

Dylan Field:
Sure. Well, yes, I first met Matt, I think it was a good 10 years ago, like you said. And we were trying to recruit Matt first as an intern, and we're like, "Hey, Matt, you don't need to be an intern. You just come over to Figma, just drop out." He ended up taking the Thiel Fellowship, which I'd done as well and started an amazing company, which is now hosting a podcast that I get to be on. So I'm very glad to see Matt's success, but I do wish that we could have worked together at some point at Figma, not going to lie.
But yeah, Figma is a collaborative design platform, and we try to take people all the way from idea, with FigJam, to design, with Figma Design, making it so people can create mockups, websites, and more graphical materials, and then we go to production.
And I think that one of the big areas that we're trying to figure out is, how do we make it so that we're able to better serve the developer audience? And so I'm actually really excited to chat with you today because this is something that we've been thinking a lot about recently is how to be better for developers.

Matt Arbesfeld:
Awesome. And, yeah, really excited to hear more about that design to development process and action area. I've thought about a lot as well, and what can be done there. Maybe, I guess, would love to go back to your founding story. So you're from California, San Francisco area. It sounds like you learned to code in high school and then joined Brown to study computer science. I guess when did you first get into programming and design?

Dylan Field:
Yeah, I think the first time I ever wrote code was probably second, fourth grade. I skipped third, so that's why there's a null there. But I had a friend in fourth grade whose older brother kind of was able to show me some stuff. There was Scholastic, which I was able to use to get my hands on a basic compiler. I had at some point gotten a book on how to program JavaScript or like HTML dummies.
And so I'd use little things that could help me get access to it, but I didn't really have a compiler for a long time. Eventually started to get the chance to write a bit more C/C++, but didn't know the basics for a while. And so it was kind of a gradual thing from, yes, I was doing some stuff in elementary school, but until probably middle school, with robotics... And there was sort of a graphical user interface you could use that was actually, I think, really instructive for me to think about programming and different loop structures and conditionals.
But then, from there I found you can actually have a more powerful way to do it with a sort of C interface as well. And again, I really didn't understand all the basics at that point, but then going to high school, started to learn a lot more. I started to do a lot more web development in high school as well. I had been interested in design. A lot of my design exposure was coming from Neopets in their early days, which was a virtual pets website. But the cool thing about Neopets was that the virtual pets were kind of just the entry point. There was entire market economy, there was a website creation tool, it just kept going. They had everything you could possibly imagine, but people thought it was about virtual pets, which was kind of hilarious.
And through that website, I was doing web design for guilds and stuff. It was pretty bad, but it was fun. And then in high school, started doing more with PHP. I started making websites for friends and websites for my high school and stuff like that.

Matt Arbesfeld:
So you consider that your first business, you'd build Neopet's websites for Neopet's goods, like battle items and coins?

Dylan Field:
I don't think that I ever tried to charge people for it, so, no, I wouldn't call it a business. It was more just on Neopets you could join a guild. And the cool thing on Neopets was you could get avatars, or that was one of the cool things. And you get avatars by getting items. And so I was part of an avatar guild and we would basically trade items to get avatars, but we had to create a web presence for it. And so I think on my pet page, I did a lot of iterations to try to get my website to be really good.
MySpace was also fun. For what it's worth, I learned really fast that there are people that are better at web design than I was that time in CSS, one of which actually is now the co-founder of Kapwing. So I made a deal with him. I was like, "Hey, I really like CSS and design, but, actually, I think you're better at it than I am and you're way more creative on this stuff. If I give you server space, my VPS, will you help me redesign my MySpace page? Because I don't know how you're doing what you're doing, but it's amazing."
And so that is the co-founder of Kapwing, Eric, who coincidentally is my middle school and then we were buds. And so now he's making stuff in the creative tool space. It's kind of wild looking back, too, all the little intersection points and people that, even at a young age, I was lucky to be surrounded by,

Matt Arbesfeld:
You may have been the only person with a MySpace page that wasn't completely atrocious-looking and flashing text everywhere.

Dylan Field:
Eric also had a really cool MySpace page. I think eventually I took it down though, his version, because it was sort of a form over function thing. He had made it so you couldn't actually get to any of the links on the page in the header. He had managed to figure out a way to get rid of all of the MySpace stuff and just basically make it a web page. And so, at some point I was like, "Okay, this is really gorgeous, but I think I need to now go back and make this again myself."

Matt Arbesfeld:
Wow.

Dylan Field:
I also definitely studied- I was trying to learn about computer security a lot. I was really interested in cryptography and also just how to do database stuff. I was fascinated with some of the MySpace forums. And then also just doing a lot around robotics, programming for FIRST Robotics or trying to learn it, but honestly probably not that helpful to my team, unfortunately.

Matt Arbesfeld:
Cool. So then you went to Brown, I guess, two-and-a-half years and you dropped out, took the Thiel Fellowship to pursue Figma full-time maybe? I'm guessing you had a lot of projects throughout the years. How did you decide now is the time to drop out and pursue this, or what was that experience like?

Dylan Field:
Yeah, and before Brown I actually had another really good experience, which was working at O'Reilly Media. O'Reilly makes all sorts of stuff, but at the time they were really known for their conferences, computer textbooks with animals on the front. And honestly, it was just amazing working there because I got such access to the warehouse. And they had a used books rack where I could just go and... It's not used; it was like damaged books, but the damages were pretty low. So I could just get these books and learn about these different CS topics, which was awesome.
But yeah, went to Brown, studied computer science there, realized how much I didn't know, which was very humbling. And I also had a chance, like you said, I worked on a lot of other projects. One of them was this website CourseKick with my friend Devin, who's now the CEO of OpenSea, and it was sort of like a social course browser with recommendations. We convinced Brown's IT team to get us access to all the course pair data, so you can make recommendations for different classes you should take. And it was super fun, but at some point I was like, "Okay, this is not a business. Maybe we can make something social around this, where you're able to go talk with other students in your class and it's like a hub for each class you're in," but it just didn't seem like there was a real path there.
I tried to create something that was a little Discord-esque as well at Brown, it didn't make it off the ground. I had plenty of other projects along the way, but I think for Figma I kind of realized... My junior Fall was my last semester at Brown. I was going back to Flipboard for a second six-month internship. I'd already interned that point at LinkedIn, InDinero, and Flipboard a first time, as well as been a research assistant at Microsoft Research, and I was like, "Okay, if I want to go do a company, which I am very interested in doing, I got to be all in. I can't try to do it while I'm at school. Maybe some people can do that, I can't." And so, yeah, I decided, "Okay, I'm going to go all in."
I had the chance to work with my co-founder, Evan. And he was my TA at Brown, he was the smartest guy I'd ever met, and so that was a no-brainer. And I had a chance to work with some really amazing people at LinkedIn and Flipboard too. And it was just, Evan was so next level. So yeah, I was like, "Okay, Evan, are you down?" And he was. And thankfully we got the Thiel Fellowship, which provided a source of funds, and hit the ground running.

Matt Arbesfeld:
And at that point, had you decided on WebGL as of the technology you wanted to build around? I know there was drones in there. Was it just, "Let's just do something," or had you had a idea or hypothesis at that point?

Dylan Field:
Yeah, I think it's probably an anti-pattern. So I think what's good about our approach was that there should always be a "why now" for a company. And if you don't have something about the world that's changing that benefits you, that's not a good sign. And if your intuition is that there's something that you should build and you can't find that "why now", it might be you still have to keep looking.
A lot of times I think it takes a while to figure it out and to articulate it. For example, Twitter, I think for a while I couldn't articulate the "why now" of Twitter. Why could Twitter exist one year but not the previous year? And I think one of the reasons why was that in the early days of Twitter people were texting their updates. That's why you had 140 characters, because that was the text limit at the time. And that sort of SMS and adoption of SMS as a way to communicate with the internet, that was important. And then I think another reason for Twitter was everyone had away messages on the AIM. And so people had been trained with this idea of, "I have an away message. I updated the away message," and that was a behavior that translated into tweeting.
So I wasn't able to articulate that though for four or five years after Twitter launched. So I think it's sometimes hard to say it in the moment. Now, for us, there were all sorts of why now's for Figma, but the technology why now is WebGL. And I think we were kind of going with that framework of, "Okay, what other technologies are there out there?" Drones is another one, like you mentioned. We thought, "Okay, drones, the regulatory stuff is just not going to be good so we should not go into drones."
And I think it was a hard thing, for example, Zipline, to do. They had to start in Africa before they were able to come to the States. And I really admire that team's perseverance. And ironic, by the way, because Zipline was another name we considered for Figma, but I digress. Anyway, so yeah, I think not the right thing to do if you're founding a new company, but ended up benefiting us because WebGL was such an important technology.

Matt Arbesfeld:
Yeah, and I think, especially for technical folks, it can be a good framework. We started LogRocket because React and frontend was emerging, and seeing there could be a lot of opportunity there. So that resonates.

Dylan Field:
Yeah, I think it's a good source of hypothesis generation. I'm just trying to say, you should be willing to consider ideas that are outside of your "why now" framework too. If you get two married to one technology, that can be a bad thing because then you might be having a solution to problems that don't exist, and I've seen that happen with a lot of developers, but also it can lead to really great outcomes as well. So it's a sort of double-edged sword maybe.

Matt Arbesfeld:
Yeah. And maybe we'll get to crypto, probably not, but I think a lot of that in crypto.

Dylan Field:
Right. I think that might be a different podcast.

Matt Arbesfeld:
How did you decide for vector editing to be the first product that Figma launched? How did you get to that place?

Dylan Field:
Yeah, I think the insight was, okay, interface design and screen design is going to be really important. And it already is, but the number of people doing screen design in the world is going to be going up. And I think we started to realize that as we were talking with different companies that were differentiating through design.
And we'd just come from a decade where we had really changed from on-prem servers to cloud, with AWS, from box software to the app stores. Developer tools were getting better. Not as good as today; we can talk about that in a bit maybe. But I think that all these things were making it so that design was becoming the competitive differentiator. And companies were starting to really up their hiring of designers because users expected better design, they were coming from places like using Gmail or using Facebook. And with all that, they were suddenly requesting that design not be terrible, which was awesome.
And so with that, I think we started to go, "Okay, what can we do to make design tools better?" Fireworks had been killed. We felt like there wasn't a great design tool in the market. Sketch was so nascent, incredibly buggy. It seemed kind of like an indie thing, which was a total miscalculation on my part, but that's what we believed at the time. And so we went head first into that, and it took a while to build. It took a few years, different twists and turns along the way. We thought for a while we should de-risk it on the web by doing this kind of crazy cross-platform framework. And Evan is amazing and built this really cool framework, but it actually ended up slowing us down. I drug it all out later on and ended up just focusing on the web because the web was good enough.
And over the course of it, we also saw things like Emscripten and WASM sort of develop and mature, as well as WebGL become more prolific and more important, and seeing people invest more in that. But yeah, I think without WebGL, but also without other web technologies as well, Figma wouldn't be here.

Matt Arbesfeld:
 I've been surprised not to see, say, Photoshop in the browser or video editing tools in the browser. Has there just not been an equivalent product that's Figma's quality, or do you just think there's something about this designing for application space that...

Dylan Field:
Yeah, I think you should look harder. I think there's a lot of really good tools out there for all sorts of different use cases. And the browser is awesome because it's an easy way to implement things, share proof of concepts, and make it so that you're able to also democratize access to things, and so I think that it's the future of all tooling. But for Figma, in the early days, all of our competitors were not in the browser, and so that was something that gave us a leg up.

Matt Arbesfeld:
And when you were building, the first few years, and you'd show it to people, was there excitement? How much conviction did you have when you'd show people, or was it a lot of questioning?

Dylan Field:
There was less comparisons to their products and more going, "Hey, is this actually ever going to work?" We'd user test, and I would bring a bottle of wine with me because I knew that entering text took so long. "You want to write a paragraph? Let's take 15 minutes to do that. I'll pour you a few glasses of wine." And so it was kind of brutal in the early days from a performance standpoint, but we knew we could get it there. We had a lot of conviction, but we knew it was going to take a while to do that as an engineering challenge.

Matt Arbesfeld:
Got it. So it was like if this worked and it could be built, then you knew the market was going to need it. It was mostly technical risk and, were people going to care?

Dylan Field:
Yeah. I think, looking back, there was also some design risk in that, in the early days, I would show people Figma and a lot of them said, "I just don't trust it." I was like, "Oh shit, is it because it's in the browser? Why don't you trust it?" And they just couldn't articulate it.
And then one day we did a refresh on our visual design, and suddenly that complaint went away.

Matt Arbesfeld:
Wow.

Dylan Field:
And it was like, "Okay, I guess part of why you didn't trust it was because it looked like it was more... We were focusing more on the UX than the visual experience", and I think that that ended up being a really big inflection point for us. Turns out designers really care about visual design. Who could have guessed?

Matt Arbesfeld:
Who could have guessed? Yeah. And I guess maybe it gave you more conviction even that it was important to have good tools so you could design good product.

Dylan Field:
Exactly. The multiplayer, I think we waited a long time to build that, and that was something that we really had to take a leap of faith on. But we also had been using it, and it just was clear that you needed multiplayer for the web feel. The form factor of the web, it was like that medium needed multiplayer to work. It felt wrong without it.

Matt Arbesfeld:
Yeah, people are not used to saving files in the browser and asynchronous work.

Dylan Field:
Yep, exactly.

Matt Arbesfeld:
So moving on to Figma, I guess it was six years until you launched the second product or so, FigJam. Is that six or seven years? Maybe kind of talk through that, the thought process behind when to launch a new product versus invest all your resources in the core product, and how you thought through the calculus on that decision.

Dylan Field:
Yeah. I mean, we were watching the use cases of Figma. And we saw when the pandemic started that diagramming whiteboarding was exploding. We were getting so many questions about this means use case as well as lighter weight use cases for Figma, and people were starting to treat Figma as a space to hang out in.
And so it was something that we'd been considering for a long time, is there a lighter weight version of Figma to make? Is there some way to really emphasize diagramming whiteboarding, brainstorming as a use case? But yeah, we started to really accelerate those plans once we saw that feedback at the start of the pandemic and went all in on FigJam. And it was awesome.
The team really committed. We built FigJam in about eight months and shipped it to go as beta, and it's now a really meaningful part of our business. We started monetizing earlier this year the product, and it's been amazing just to see how many teams around the world now are using this as the way they whiteboard, the way they brainstorm.
And I think we're also really excited about thinking about it as a visual platform, and how do we make it so that we're able to bring more visual content and connect more applications into FigJam too? Because that can be useful when you're brainstorming to have all that information at your fingertips.
So yeah, I'm quite excited about by the future there, and also to complete the vision, from idea to design to production, because I think that that entire loop and that process is the one that, that's the box we're in for Figma. And we're trying to figure out, how do we make this entire software development life cycle better? And it starts with an idea, but it ends with code, and I think that Figma is still not as good as it could be at going from design to code.

Matt Arbesfeld:
Yeah, so let's talk about that. I know our development team, they're in Figma all the time looking, talking to designers, getting sizes, CSS colors. What do you see next in that space of design/development collaboration? What are the challenges?

Dylan Field:
First of all, about a third of our users for Figma are developers, and so we have a lot of developer users at Figma. And when we look at their satisfaction for Figma compared to designers, it's much lower. And I think there are core problems about, first of all, how do I select things? How do I use Figma as a product? Some of them are very foundational, but also, how do I use this alongside the tools that I love? How do I use this in the context of my design system? How do I use this to go to code and speed up my workflow? And we're trying to take a very holistic view of all the challenges I just mentioned plus more to figure out a new way for developers to interact with Figma. And yeah, we're spending a lot of time on that right now, so pretty excited.
I think that it's just really important to have really good developer ergonomics, not just when you're coding, but also when you're going from implementing to design. And I think if we can bring those richer primitives to product design, that can help streamline design to production.

Matt Arbesfeld:
And are you envisioning something like you can take a exporter component directly to a React component, or tie them one to one, or still thinking through what the main problems are there?

Dylan Field:
I think that there's a lot of frameworks right now that are really interesting. I think that React is one of them certainly, but I actually think that there's a bunch of frameworks that are starting to come out that will provide a challenge to React, which has, I think, been needed for a long time. And it's great to see the space get more innovative. And so I hope that whatever we do and solve, we can solve in a general way, and hopefully do it in partnership with our ecosystem and developers to figure out how to make those bridges.
But yeah, I feel like, looking ahead, there's so much we can do to make developer ergonomics better, to make developer productivity better. And I think that that's the next wave of innovation around web frameworks is, how do you have faster build systems? How do you innovate, explore faster? One thing we're talking a lot about internally right now are frameworks like Qwik and Remix, and it's cool to see just the pace of development that's happening in the webbing system now.

Matt Arbesfeld:
Where do you think developers are, I guess, wasting time or there's opportunity to save them? Is it in the component building itself, the testing? What's the low-hanging fruit for where people are wasting time?

Dylan Field:
Yeah. I mean, I think that all the things you just mentioned are definitely part of it. And I think there's the web development part and there's like the Figma to code part. I don't think Figma should ever be a one-click code generator probably, the needs of our users are too complex for that, unless the technology gets so good that you can start to do that and save people a ton of time, but I do think that Figma should be able to help with that journey.
And then as I think about developer productivity outside of Figma, I just think that web apps have gotten really complicated and development's gotten slower. And so whether it's through new build systems, like Evan, my co-founder, he's no longer at Figma; now, he's working on something called esbuild, which is pretty amazing, and probably a lot of people that are listening to this have heard about it.
But I also think that there's sort of a chance to reset that bar of what's good enough from a user experience standpoint too. For example, I believe that the entire web should be fast and real-time, and secure. And also, probably one thing that we're not great at at Figma yet, but one thing that the web should be going forward is better at offline. And I think that as you sort layer those requirements on and think about what you need as a system holistically for development, you think more about what's on the edge, you think more about declarative approaches. I think that there's a lot that we can see over the coming years in terms of improvements in web development.

Matt Arbesfeld:
Do you see a world, and there's been kind of an explosion of these low-code, no-code type tools, where designers themselves could actually get code into production and developers can work on, say, some of the harder problems, or do you think that professional developers will be needed for every type of application?

Dylan Field:
Good question. I think that always is a long time. And I think that if you were to teleport back 30 years ago and try to explain the world of web development today to someone in '92, that would be a very hard task, right? So what does the web look like in 30 years? Gosh, I think we can only start to imagine. I think it's maybe a little bit easier to imagine what it looks like in 10 years.
I think that developers and designers, the edges between them are becoming more blurry. I think more designers are learning to write code. I think more developers are learning to design. I think that both professions need to figure out how to work more closely with each other. And yeah, if there's a way to automate part of the work, I think that hopefully that can be done to enhance the end product, the work product, rather than replace people. I think that there's a path to both those futures that I could see, but I'd definitely want to work towards one where automation serves designers and developers become better craftspeople and to express themselves in new ways in collaboration with technology versus trying to get rid of designers or get rid of developers or something like that.

Matt Arbesfeld:
Yeah, I remember using Dreamweaver back in 2004, where you would code and you'd see an update on... You'd see the visuals, and you could drag the visual to change the code. I guess do you see a future like that for Figma, where you're generating code in real- time and it's all synced to an actual code base, or is that you see those as clearly separate concepts, your design from your code base?

Dylan Field:
So we've explored that, and a lot of our customers that are more on the product side, I don't think they want that yet. I think that their needs are more sophisticated and they need things that are more on the component level. I do think that that's the promise that everyone gets very excited about is like, "Aha, but what if this could work?"
And I think that it could also be great for learning, by the way. Yes, Dreamweaver used some really rough implementations of design to table. I think I learned more about the table element than anything else from Dreamweaver, but I also learned about other things too. And I definitely, by studying the code output of Dreamweaver, that was a way that I learned about some aspects of web design and HTML and CSS.
We've also had people on the Figma team that were part of building Dreamweaver. And I think that one thing that Sho, our VP of Product, who was on that effort and is now working at Figma, really highlighted for me is there are different thinking modalities that you can be in when you're designing. One is your more freeform and you're thinking without structure, the other is you're being more structured in your thinking from the start, but they really are different ways to go about a problem. And I think that Figma is more freeform and you add structure later on. And I think that if you're trying to go more in Dreamweaver direction, you really need to start with more structure. And I don't think it's where we're at currently.
Could we get there? Certainly. Do we want to get there right now? That's not our immediate focus, but I do think that there's a lot we can do before going all the way from press the button and you get code. Before you go do that, I think there's a lot more you can do to make the design to development handoff better in Figma, and make it so designers and developers can work better together. So yeah, that's what I'm excited about for the year ahead.

Matt Arbesfeld:
I guess for folks who are... Probably many people here are in Figma and they're maybe at a 101 level, like they're creating designs and they're sort of... Designers, developers go in, they look at things, they try to recreate it. Are there things in Figma that you recommend people start using, like the design systems, or, I guess, how do you see leveling up one's usage of Figma to start improving that collaboration process today?

Dylan Field:
That's a great question. I think first is just visual fluency. I think it's exactly what you said of making things that you're seeing, or go and try to build your favorite mobile app, for example. Take a screenshot and recreate it. And then once you have that visual fluency of how to use Figma, I think there's all sorts of optimizations you can do. Yes, design systems is one, auto layout's another, but there's just so much that's already there and so much coming too. Our YouTube channel, I think, is a good place to start, but also there's a lot of stuff on YouTube. And if you start to search for functionality and Figma, I think you'll be surprised at how much you can find. I think also the help center at Figma is something that we try to do a lot to always have that very updated, and so hopefully that's helpful too.

Matt Arbesfeld:
Maybe switching gears a bit, most of the audience is developers and we all have ideas of companies we want to start, and there's a lot of new stuff now coming out, like AI systems, crypto, new things, like all the frameworks you mentioned. How would you help someone think through, "Do I start a company? Do I quit my job and start a company?" In this, when do you suggest someone do that versus stay and build something as part of a bigger team?

Dylan Field:
I think most people that start companies are obsessive about the idea that they have. And for us, it was a sequence of ideas, but I guarantee you, and I think you saw some of it, we were obsessed about each one along the way. So it was kind of like a nagging thought in your head of, "One day maybe I should start a company. It could be fun," but you're not waking up every day going, "Oh, man, there's this thing I want to build," or, "There's this technology I want to explore and the only way I can do it is starting a company," or, "This person I want to work with." I think probably hold off until that is the case because it's such a huge commitment that you're making.
I think that if you're doing it, expect to do it for a while. I'm very thankful for the Thiel Fellowship that we both were able to enjoy, which gives you two years of funding, $100,000 over two years. It's dispensed monthly because people that get it are typically under 22 and you don't want to give those folks a big lump check probably, for a lot of 20 to 22-year-olds. Some of them I think would be fine, but others not so much. They're like, "Oh, what if we put it all in Bitcoin?" Which maybe some Thiel Fellows have done along the way, but anyway, I digress.
But I think you want to have that time to build a company. I mean, if we had thrown in the towel six months in and said, "In six months is our evaluation point. Is it working or not? If not, let's go and do something else," for me, go back to school, for Evan, go accept any of the 1,000 jobs that he was being offered. I think if we had evaluated that six months in, we would've certainly not ever gotten to where Figma is today, because we just had not much to show at six months. We had a lot of prototypes, but nothing that was like, "Here's our clear direction," and was a harder point.
I mean, we didn't start charging until five years in. And by the way, I hope that people don't take away from this conversation that you should wait five years to charge or wait three years to launch. All the canonical advice about getting out there, iterating quickly, and moving fast, I think that's the right advice. And when I see companies that are taking three years to do what Figma did, I really get worried about them. I worry about the team's morale, I worry about their ability to actually get a read on the market. I think we're very much the exception, and I worry people sometimes think that we're a new rule or something, and we're really not.
But anyway, yeah, so I think you want to make sure you have the money to support your path and to try to iterate over time. And I think I would suggest a partner. There's some people that can go it alone. I don't know if I would've been one of them and I definitely could not have built Figma by myself. But even if I was doing another company, myself in 2012 when I started Figma, I don't think I could have done it from an emotional maturity standpoint. Even having been through the journey now, if I was to start a company today on my own, I think it would be very tough.
My wife is a solo founder, and I'm just so impressed by her, the fact that she can do that. So it can be done; just, it's really rough.

Matt Arbesfeld:
Yeah. Especially to your... So at LogRocket, it took us two, two-and-a-half years to really launch and get conviction around it, and during that time you just always have that existential, "what am I doing?" and having the partner helps a lot.

Dylan Field:
And also though, Matt, I remember when I visited you in Cambridge, and you had within two to three months of starting a demo that you were showing people and getting feedback on.

Matt Arbesfeld:
That's true. Yeah, getting feedback early and... For us, it was, though. People would say, "Oh," to your point, "I don't know. Can you actually do this? Will this work? I use so and so." You still have to kind of push through some of that skepticism, I think, if it's a new idea.

Dylan Field:
Yep, totally.

Matt Arbesfeld:
I guess on the counter side, now, funding's getting tighter. There may even folks who raise the seed, even series A, and have been at it two, three years. When do you, say, wave the white flag and join another company versus keep persevering? How do you help people think through that decision?

Dylan Field:
Well, if you're an amazing founder with an amazing team and you're ready to wave that white flag and join Figma, call us. We're acquiring companies if the team's amazing, and doing the acquire route. But no, I mean, I think that the best founders persevere a lot of the times, but also they're very honest about where things are at, where their team is at, where morale is at. And I don't think there's any one answer here. There's no one framework to make that decision. It's a huge decision for anyone to make, just like it is a huge decision to drop out or start a company. And so I was trying to give some guiding principles earlier, but they're not holistic, and I think there's so many ways to do a startup and the rules are... There's always an exception, and the exceptions are the ones that drive the biggest outcomes often.
So I think that for funding right now, yes, it's harder. If you're in a seed stage, it's very different than later stage. If you're a seed stage, there are still things that are getting funded. If you have traction and you're growing and your retention's good on consumer or more software, that's a really good sign and people will definitely be taking your calls. Warm intros help a lot, especially warm intros from founders. And I think getting the right connections to the right VCs is hard from a networking standpoint, but very possible. Once you're kind of in and know people that are connected to VC, it only takes a few folks to get those warm intros and build a raise around. If you do raise around, do it in parallel, time box it, don't let it drag on forever; otherwise, it can be a huge distraction for you and your team. But also I think have a point of view of how much dilution you want to do.
And I think it's really tough for founders to figure the right thing out, but there's sort of a spectrum of, on one side, raising a lot of money so you feel safe and you have that psychological safety to go and be creative and innovative and do it for a long time, another side, raising just enough to get to the next round. And some people are more comfortable playing it close to the vest. For Figma, we always raised a little bit more money than we needed, but expensed that as dilution. And so you're going to have to decide for yourself what's right for all that.

Matt Arbesfeld:
I think a lot of successes have been teams. That first idea doesn't work, second idea doesn't work, third idea, even at scale. I was at Meteor that pivoted to Apollo and is now doing really well, and then we all... Segment started as a whole different thing, Slack.

Dylan Field:
I remember crashing a... Segment was a bunch of MIT people plus one RISD, and I was friends with them when they were starting Segment when it was Class Metric. They were trying to do an education thing, get metrics on how you're performing as a professor or whatever, and it was brutal.
And then at some point they showed me, "Hey, we built this thing over the weekend. We're calling it Segment. What do you think? We launched it in Hacker News and people seemed to really like it." I was like, "Oh, cool. Good for you guys." But I was so surprised when, after all the Class Metric days, it just took off. Not because I didn't believe in the team, I always believed in them, but just because it's like a weekend thing, and then, boom, it took off.
And I think it's really hard to break that stuff, but they were committed, they were all in, and they're all obsessed of working together and figuring it out together as a group. And that's really rare, I think. And again, I went back to the earlier part of the conversation about once you start something, part of that I think could be you just have someone you want to work with. And that was definitely the case for me and Evan too. I wanted to work with Evan and I thought he was a genius and I felt so lucky to get to spend time with him even.

Matt Arbesfeld:
It takes time to wander and your judgment just improves. Maybe final area to touch, and I've always found you to be years ahead of the trends. So for folks starting now, say we just started a company now, how like WebGL was in 2013 or you had a few other things, what would you say are some interesting categories or areas that you think they may be new technologies that could have problems to be solved or could solve problems in uniquely new ways?

Dylan Field:
Yeah, I think there's so much. And honestly, a lot of the technologies I'm getting excited about now, which I probably won't go into this time, we can talk later about it, are actually beyond software. I think so many technologies are outside the world of software development right now, and some of them have the chance of doing things like curing cancer. I mean, it's pretty incredible what we could do for our immune systems and in the bio realm. But also, am I an expert on them? Absolutely not. So take everything I say about that kind of thing with a grain of salt.
But I think in the world of software, I mean, I'm very excited about zero-knowledge proofs and have been for a while. My wife is working in that area, and I think it's pretty magical what you can do with your knowledge proofs and how far those systems can get you, and how they can enhance privacy and enable more trustless computing.
I think the obvious one everyone's talking about is AI. I still don't know what's going to happen with AI. I think there's a chance that it's very competitive and very deflationary, but there's also a chance that it'll end up enabling so many new things. And by enabling so many new things, there'll be entirely new industries that sprout up from it that are just really hard to predict. And so I can't tell you what it's going to look like in two years with a technology like that. Do I find it fascinating and really enjoy watching the latest developments? Absolutely. But the game theory of how it all plays out is a little bit hard for me.

Matt Arbesfeld:
Yeah, and where value goes. Does it go to the big companies that already have workflows or are there new companies that emerge?

Dylan Field:
Or does it dissipate entirely? I don't know. Again, I think it's potentially very deflationary.

Matt Arbesfeld:
What do you mean by that? In that it automates certain parts of the economy, and therefore makes...

Dylan Field:
Well, there's that risk too, but I meant more about like can you... So how much does it cost to train a model? Once the model is trained, does everyone end up getting those weights eventually? I think that the cost of training models is going down over time, so even the ones that are capital constrained now might not be forever. Does the advantage go to the infrastructure layer? Does it go to the application layer? My guess is it goes to the application layer, not the infrastructure layer, but I could be wrong, and it could be that there are real modes. And maybe once you get to the point where there's one model that everyone's using for something, there are sort of data feedback loops that end up being really important there.

Matt Arbesfeld:
One thing we've done at LogRocket is we have workflows our customers do again and again, where we can detect those and automate those workflows. Are there things like that for Figma that are a potential, or is there a place for AI at Figma?

Dylan Field:
I think there's so much potential in trying to make it so that people are able to improve their productivity in Figma and also just explore more variance of what they're trying to do. When it comes to that efficiency, we're not just thinking about AI, we're thinking about also, how do you improve our tool? And we're not done yet. We got a lot of improvements coming for Figma over the next few years, and so I'm really excited to see how people continue to evolve their uses of the tool. And the challenge for us is, how do we keep it simple? We have so many ideas and so much power that we can pack into Figma, but how do we make it so that people are able to use it really easily and that you can take that first step into Figma and not be intimidated too. And so I think there's also a huge UI challenge in, what's that interface when you start using Figma? And how do you make it so that it's totally approachable for everyone, whether they're in fifth grade or they're a veteran designer?

Matt Arbesfeld:
And you mentioned biotech. So will you be spending your evenings learning about genomes and RNA? What are you learning about, I guess, these days?

Dylan Field:
Oh, man. There's so much to learn about, man. Right now, I'm learning about babies because I just had a kid and he is endlessly fascinating. So my current obsession is just cognitive development and probably more cog psy and how do you help kids develop and be a good parent? So I'm trying my best at that.

Matt Arbesfeld:
Do you think your experience running Figma has helped you with a child, or is it a whole different .. ?

Dylan Field:
I don't know, ask me in 10, 20 years. Early on in Figma, I remember talking with a famous VC who said over dinner conversation, where I think they were kind of trying to fund me, and this VC and I are now friends, but I won't say their name, but they said, "I don't think you can be an effective manager until you've had kids."

Matt Arbesfeld:
Hmm.

Dylan Field:
And I was like 20 and had no plans to have children, and so I was like, "Well, shit. I guess I will always be a shitty manager then." So I think that hopefully something about dealing with people and learning how to navigate relationships will help me with children. But also, already it's been a challenge beyond what I've encountered with Figma. And my child is quite young, so I don't know, we'll see.

Matt Arbesfeld:
Yeah. Having also grown up as I've been growing the company, the parallels of the personal life and the company each inform each other a lot. So yeah, excited to see what you learn in the next few years doing those. Well, this is great. Yes, really interesting conversation. Super excited to hear more about what you guys are doing to help developers get more value out of Figma.

Dylan Field:
Thank you. And thanks for having me. And it's great to see you, as always. Be well, man.

Matt Arbesfeld:
All right, take care.