Figma CEO Dylan Field on the future of design and AI

24:44And I also am just seeing some real talent war dynamics heat up for design. I'm not saying it's at the AI scale. Let's be real. We're not seeing headlines yet about the crazy 10 billion, whatever it is, pay package for a designer yet. But the talent wars are heating up. You do. I feel like people are begging for zero to one design founders right now. Yeah. Yeah. You see it on X constantly. It's needed. Yeah. Every founding team knows that need designer to stand out. And if they don't know it yet, like, you know, it might be a flag if you're investing. I'm not saying that every area requires it.

25:27If you're some deep infrastructure science company and you barely have a website, perhaps you don't need to have a designer in your founding team in the first five people. But if you're building software, and it's going to be user facing, that's something you should be considering deeply.

25:42Ellis:I feel like there's a temptation as a design founder, for better or for worse, to redesign the whole company from HR to sales. We're going to do a flat structure. We're going to do a vertical structure. We're going to do siloed. And I love that mindset, but I feel like it can also be distracting. as like a designer who sees problems and wants to fix them. Have you experienced that? It's like if you think about a brand and the journey a user goes on with a brand, you kind of think about a timeline and there's all these different brand touch points. It's like, what's the equivalent for a company?

26:16And some of those touch points, you know, feel really good for the user of Figma internally and employee. But what are the touch points that we can make like 10x better, 100x better? I think there's still a lot. And I really want to do that as well. So, yeah. Do you still spend time designing? Are you like cranking away and Figma yourself still? I would say my design contributions in comparison to the team are very low. But I think that my feedback is very large. I usually try to do a pattern where I am voicing feedback through customer voice. So the reality is if you're operating a platform at any scale, you can find people saying things that really represent any opinion about the product and they'll conflict.

27:10Ellis:But when I can find that this features the best, here's a tweet that it's the worst. Yeah, exactly. And our customers are amazing. Designers are incredible. They'll actually give very detailed feedback, as you might imagine. And so the chance to be able to funnel that towards our team is quite great. What do you feel like you've had to kind of pivot on in the last couple of years in terms of the feedback you've heard once you launch something? I think the biggest thing that I wouldn't say is a pivot. It was more just like it took time to build out was we had our first launch around AI, which I think was there.

27:55There are many things to reflect on for us. But basically what we did is we had a feature internally called First Draft. And this was a very basic AI feature in that it was essentially an LM assembling Lego pieces. And what happened with that was we shipped it as a feature called Make Design. And Make Design, we did not, in the rush to launch, do a full QA process on everything that were all those Lego pieces. And there were some situations, or at least one situation, where people could recreate UIs that shouldn't have been able to be recreated with Lego pieces that we had made. Like the Apple weather app, I think, was an example.

28:44Yeah, exactly. And so someone posted that. And I think if I hadn't felt it was a QA failure, if I felt it was, you know, a natural result of like a system, I might have acted differently and then fixed the system. But because it was a QA failure and because I felt like it didn't like sort of match up to my personal ethics of how I want to be operating this, we pulled it back and reintroduced it later. And to me, that was kind of a no-brainer. I think other people made a big deal about it. But then it also, I think, just started our track on, okay, how do we go build the system for real? And whether it be through frontier models, through first-party models, how do we make it so that we're able to do great work with our user, get them to design generation, whether it's zero to one or with a design system, make it so they can create edits, they can create new effects that are extremely interesting and eventually get to a working application, not just something in Figma design that's a stopping point, but go all the way.

29:55Putting that out there gave us the insights we needed that led to Figma make, that led to us and the work that we're about to start shipping in Figma design. And yeah, it's been a big learning experience overall to figure out how to work effectively with AI systems, but one that I'm really proud of the team for navigating.

30:15Ellis:I think that also raises an interesting question about what the models actually include. And when there is so much out there that is kind of like the OS style, I feel like I'm hearing lots of fears that as much as we all prize design and craft, that a lot of the generations of AI tend to be generic. Do you worry about that for the future of software design? I think that's our opportunity. When you're generating things and you're getting this average result, that's not what people want. I think that we're in a world where with AI, good enough is not enough. Good enough is now mediocre. So if you really want to stand out, you have to be great.

31:01How can Figma help make you great? And what we're trying to do is allow you to, whether it's in Make or in Figma Design, iterate, explore, and really push the bounds of what's possible, both with AI and also through the human touch. And explore that entire option space of what's possible with design faster. Because that's, to me, the design process is you're looking and you're kind of seeing around corners and you're seeing here are all the paths I could go down. Well, can you identify the paths that are most promising faster? And can everybody from your team do that too? Can you make it so that more people can contribute ideas at the start?

31:41And then as a designer, you can guide the process and you can keep pushing further to make it so that what the end result is, is very differentiated. Do you feel like designers are still worried about losing their jobs though right now? It's interesting. I was at an event last night here in Hollywood and it was more Hollywood people, but people on the production side and just there's so much fear in the room about the tools and what they're doing and enabling. And I have to, I mean, I'm not a designer and I'm not in the world like you and Ellis, but I have to imagine there's a lot of fear and consternation among a lot of Figma users that like my job's going to go away because of this stuff.

32:19Yeah. I mean, I can't speak to the Hollywood side, but we actually did a survey around this recently and put out a report on it. And it was... I'm looking at the numbers right now. It was 17 % of our respondents that said, developments in tech tools and AI are a threat to my role. So that was the actual number that we got from the survey. Now, you could argue it's 17 % high, is it low? But I certainly don't think it's like everyone's going, Oh my gosh, what's going to go? What's going to happen here? And I think it's because designers kind of intuitively know this is an opportunity for us. With AI, Figma's point of view is going to lower the floor, make it so more people can get into the process, but also raise the ceiling and make it so that you're able to do even more with AI.

33:12And I think the more we can remove the drudgery of repetitive tasks, things that maybe are easy to conceptualize or describe, but right now they're tedious to execute. That's something that AI can help with. For example, one feature we shipped super early on was rename layers. Very basic feature. It's just you press the button and it looks at all your layers and it gives them a name. And you might go, okay, well, what's the big deal? Well, designers can take hours to rename the layers right now in a complex file. And if you're someone who likes to have the clean desk and be very structured, you might really care about your layers having all the perfect names.

34:02And so this is something that saved countless hours of time for designers already. You can now think forward into what else you can do that for and how else you can remove TDM and get people to the point where they're able to actually explore ideas further, explore more ideas. That's the promise to me.

34:20Ellis:Have you made anything good lately? I've made some internal tools for myself. I experiment a lot. I try to do a bunch of very random experiments that are a little bit more artsy. Some of the things that I've seen others make that are far more impressive than what I've created myself... I saw some really cool typography tools recently that just blew my mind as part of our Make Hackathon. I saw some people that made some pretty cool interactive experiences, games, and even things like complex synthesizers, poetry explorations. And not all of these are products per se, but just the creativity of our community is kind of always amazing to see.

35:19and sometimes they also went outside the field of software. You know, there's some users that were exploring ways to understand biology better and DNA, RNA sequences. So yeah, it's pretty exciting to me to see what you can do when you start to hook up these data workflows and make it so that you can really visually explore data in that way too.

35:49This episode is brought to you by Google Chrome. You think you know a browser, but Gemini and Chrome, that's new. It can help you with practically anything on the web, like restoring a vintage motorcycle from a 50-page restoration block, or finally break down that long article you've had open for weeks. Gemini and Chrome is here for it. Ready to make anything online make sense? There's no place like Chrome. Check responses set up required, compatibility and availability varies 18+. This episode is brought to you by Google Health. Stop chasing someone else's definition of health. What matters is what's healthy for you.

36:21Google Health offers a new kind of coach, built with Gemini for effortless tracking, sleep insights, and holistic coaching tailored to you. Visit googlestore.com to learn more and start a new relationship with your health. Requires Google account, Google Health app, internet, and Google Health premium subscription. Features subject to change. Availability and results vary. Not intended for medical purposes. Works independently of Gemini apps. Check responses for accuracy.

36:47MCP is fascinating to me for people who don't know. I think it's what model context protocol. Is that right? And you guys have this pretty hilarious video for your server that you put out this week that people should check out. Definitely everyone should watch this. I am, dude. They're awesome. It's good because it also explains it in a way that I think is more digestible because MCP is like a very heady thing to understand. Even if you're into tech, you have to be in it to get it. At its highest level, I think of it as like how apps use LLMs to talk to each other and to share data between each other.

37:19And it's in the like pre-Netscape era. There's not even like a browser to see all these servers, right? But like you have one, a bunch of other companies that are embracing AI have them. And I'm curious because, you know, when we were prepping for this, obviously a big concern with Figma is like what is competition in AI and all these companies that are popping up? going to do to Figma and like prompting and all this stuff. And I'm curious actually to back into that question, like the way you all are embracing MCP and working with other companies and really opening Figma up. What does that say about where you see Figma fitting in the value stack, so to speak, of like the AI world?

38:02Because there's this huge conversation, obviously, about like, are you just an app? Are you a foundational model provider? Are you trying to do both? Where did the app sit? And it seems like you have a POV on that and you're articulating it already with the MCP strategy. So I'd love for you to maybe like dive into that with us. Yeah. I mean, without going too much into all the strategy and feature ed map, I think that, first of all, the physics of the AI space, there's sort of a gravitational pull towards being more open. And I think it's important to recognize that and to follow it. I also think that the goal is to get you from idea to product.

38:44And yes, we want to give you all the best tools that you could possibly have in the Figma platform. but if for every reason you want to go elsewhere, how do we make it so that you can bring your context from Figma to other applications? And also the future, how do you come back? And how do we have a great loop? And again, I think that there's so much partnership we can do with the ecosystem here. And I think there's actually quite a lot of win-wins, even with tools that people think are our competitors. and so it's been nice to have just very open and friendly conversations with many of the tools in the space in order to make it so that we're able to figure out what can that foundation be so that your users can have better experience and also they can go iterate and get to the point where they are in a high craft place beyond what you might be able to do in another tool and Figma afterwards.

39:43And so, you know, lots to come, but ICS is a really core part of our strategy and what we're doing in the future. So theoretically, like Lovable, I think is a good example of a company that I think people from the outside or Figma investors would go. The rise of Lovable maybe suggests Figma has some competition to worry about. Are you suggesting that like, maybe you would partner with Lovable and you could bring Figma assets into something like Lovable? Maybe we don't have to say lovable as a specific example, but the rise of these no-code vibe coding tools like them, like you actually see a way to partner with them.

40:16Is that what you're saying? Exactly. Yeah. Yeah, I think we need to have remote MCP with the right conditions, of course. But if you've got a partner who is excited to work with you, then you're able to open that up to them and that they are able to use it to create better experiences for their users. And also, I think that can be beneficial to us. Again, it's like context is created in a lot of places. and how do you take that context as well and then use it to help them in Figma? And I think that in Figma Make,

40:56I guess the bigger question is how much do we believe in Figma Make? And our belief in Figma Make is very high. We think that what you can do on the Figma platform in a connected way and what we can offer users that is very differentiated and where we're headed, We're very excited about that. And we're just very high conviction. And so I think that also is a contributor to that whole strategy of being more open

41:26Ellis:is that conviction. I know a lot of companies get disrupted at the moment when they start expanding to different verticals and this and that. When you think about slides and buzz and make and all this other stuff, do you think of it as kind of serving the whole team in a way or serving some fundamental use case as you see it? Yeah. So basically, from the start, what we've done when we introduced new products is we look at Figma design and go, OK, what is a use case that we are seeing in Figma design over and over again at scale, but it's underserved? And with that use case, we could either dive into it and make Figma design much more complex, or we could create a new surface.

42:14and sometimes with uh for example fig jam uh that ends up being not just a way to create a simpler experience for people to engage in whiteboarding brainstorming diagramming like we saw that spike as a need during covet along with this need of i want to be in the same place with my teammates and open up conversation and make it so that we're all able to contribute to that conversation more. And by making FigJam its own space, we're able to not only prevent complexity from occurring to figment design, but also it became a space where we're able to explore new concepts like, how do you just make it really fun?

42:59And it was actually kind of an interesting story because, about a month ahead of FigJam launching. We had this very strategic conversation, kind of serious, ironically. And the conversation with my board and with the team was like, okay, which way do we want to differentiate for FigJam? And my push and what we ultimately ended up doing was let's have the differentiator be fun. And the board was like, all right, let's go. The team was doing... Boards love fun. My board likes fun. And my team, I think, was a little surprised. They're like, this is our strategy. Okay, let's see how this goes. We did a design sprint one day.

43:43And out of that day, I think 20 ideas came up, many of which made it into the product as fundamental features of FigJam, and many of which have spread other places as well, like Cursor Chat came out of that day. And so then as we go and expand the platform even further, just something like slides or buzz. These are examples of, okay, people are already doing Figma design. We want to make it so that if you're in a product process, if you're on a product team, how do we give you one connected experience across all Figma? And how do you go further through the workflows that you're already doing? And these are opportunities for us to really support those workflows, those personas with these products.

44:28Ellis:consider this a formal feature request to add a hamburger sticker in fig jam please send that i think you can find already but noted i'll i'll see if we can make it more prominent really oh okay i'm gonna jump in we have a fig jam feedback channel on slack i'll be putting that in right after this at the very least that's the ultimate advantage of having a podcast right you get to ask for features i mean at the very least you can generate one so you've been talking about building for the kind of latent demands for product builders. I know when I was at Config, I don't know what percent of the people were like illustrators, graphic designers, brand designers.

45:03Ellis:I feel like those needs are kind of diverging, right? You know, the vector people versus the product and code people. No, you're shaking your head. I think that, yeah, I have a point of view here. I think that you look back at the Flash era and the web was a more dynamic more wild place i'm not saying the quality was always like as high as it could be it was a place that crashed your browser as well and maybe there was some browser crashing but also sort of the range of what was explored the dynamism uh and the visual styles that people were going for it was wide ranging in a way that we don't see today as much because there was a wild swing back first to skeuomorphism and then eventually to really Swiss minimalism.

45:56We've kind of stayed there for a while. And for a few years now, it's felt like, okay, I think we're about to swing the opposite direction again as a reaction. And I think now if you think about AI and this kind of like average output that you get and the need to differentiate through design, through craft, I actually think that's something like illustration, that's one example of a way to do that. And there's many more. But I actually think that the illustration needs for brand, for product, they're only going up right now.

46:28Ellis:This is my favorite conversation because Silicon Valley is not usually known for its fashion sense. I don't know if you've put together your iconic Dylan outfit. Okay. So yeah, I also have my ultra blingy config badge here today. I don't know if the investors approve this much bling. But yeah, Silicon Valley, not usually known for its fashion. Yeah, there we go. And a lot of times you see people on Twitter, they're like, oh my God, skeuomorphism is back. It's like, wait, no, it's flat design. And it's like, guys, it's just like fashion. It's going to be cyclical. That's how it works. There's minimalism, maximalism, serif fonts, no more serifs.

47:11I mean, first of all, I have talked about our swag. You can find our swag store online. Just search for Figma swag store. Plug it! And I think, honestly, our brand team is exceptional and has made amazing, amazing stuff. So definitely check it out. But yeah, I think that right now we're in a world where there's sort of more micro communities and more micro sort of expressions of culture than before. And this is sort of a perhaps result of the For You, I'll Go Feed experiences that we're seeing on social media right now. Because you get to these little filter bubbles and the way that people then engage with them and kind of develop in-group memes and whatnot.

48:02I think it ends up creating paralyzed fashion rather than there's one mainstream. And everyone in New York is wearing this brand this month, but next month you wear it, you're not hip anymore. Sorry. Instead, it's more like, okay, well, do you know what's happening in this corner of the internet right now? And if not, where were you? So I suspect that in terms of visual style, we're going to see a range of expression. And just in general, people are going to be trying to explore a lot of different things. And I think that's super cool, especially coming out of this era of supposed to be Muslim, where not as much as it was explored.

48:51Ellis:So Dylan, we know you're an avid investor. You just talked about finding these little niches online? What are you seeing that's exciting you most right now? I think there's just so much exciting stuff happening in the world right now. And yeah, I think it's really important for me to keep focused on Figma, of course. And so, I mean, 150%, 200 % of my time is Figma. But yeah, I'm too curious to not always be trying to learn. And so I do invest as well, but I try to keep it pretty focused at this point. Dylan, I think most full-time VCs I know would give their left arm for about 5 % of your deal flow.

49:30So the fact that you're doing this part-time is amazing. There was actually a moment a few weeks ago where browser company, Ellis' former employer, sold to Atlassian. Congrats to Ellis. Yeah. He said before the recording, but congrats again. Yeah. Congrats, Ellis. And then OpenAI bought Statsig as well. And you were like a seed investor in both. And I was seeing jokes on X of like, man, finally, that poor Dylan guy, finally, he like catches a break. And I'm hearing about you in interesting AI startups all the time. Like I'll frequently meet with a founder and I'll be like, who did your seed? And it was like either Dylan or Sam Allman.

50:10And you're actually, you led the A for one of our sponsors of the show, Warp. Really awesome. It's amazing. AI coding platform. And obviously, you are running Figma. That's your full-time job. But to me, that suggests you are very keen to see around the corner and see what's coming at the early stage. And I imagine as the founder of a now public company, that gives you kind of a unique edge of where especially AI is headed. And so I'm curious, like when you're doing this investing, I imagine it actually accrues to Figma as well in terms of like your ability to see ahead. So I'm curious, like, how does that, yeah, how does that impact how you run Figma, the bets you make?

50:54And is that a thing that's just kind of naturally happened or was that a strategy? Oh, no, I mean, it's a natural result of, I think, my curiosity and just the amazing friends I've gotten to meet along the way, as well as folks I'm getting to meet now. But yeah, I think that, maybe it's two categories that help value accrue to Figma. One is not looking ahead, but almost reminding me of lessons that have already been learned. You know, there's a lot of times where I'm talking with a company and I'm giving advice and I'm reminded of a lesson that we've already learned at Figma. And then something clicks and I'm like, wait a second, why am I not applying this to this problem that we have today?

51:44You know, so it's kind of wild, but whether it's that or it's like literally learning from, you know, live case studies that are going on in companies that are very different from Figma, those are very helpful for me. And then I think that the category of for Figma, how to think about around corners and understand what's happening. I mean, there are patterns to the challenges that AI companies have. and the ways that they are pushing forward and trying to solve problems that are hard for everyone. And I think that definitely, as I talk with founders, that's helpful to understand. But also, just in general, I just think about AI right now and I think about design and it's so early.

52:38like we're all kind of hyping it up like this is you know it's the future of everything and that's not to say it's not but I think if you think about the future of interfaces you think about what interfaces will look like and what the distribution of activity will be between humans and agents and how agents will show up in interfaces like I'm a big believer that it'll be more of the model where you're working alongside agents collaboratively and you're understanding what they're doing because the trust is not there around with the current models at least and I think even models that are in the future they're going to execute sort of your desire perfectly and also humans for better for worse we hold AI to a higher standard than we do other humans.

53:34And so you actually really want to see and audit what an agent's doing. And I think that whether it's, you know, the sort of like, how do I understand what this agent did behind the scenes, or it's watching agent real-time work, or it's, you know, you had in your last episode, Mark Khan, talking about the new glasses and the neural band had a chance to see those before launch a few times and i mean just the design challenges these new services have are so exciting uh and i think it's it unlocks like so much creativity um and possibility and i have a lot of respect for that team for taking that huge bet and executing against it for so many years um but then you go to auto sector and the way that we're We're just going to see innovation there because everyone's trying to push forward on digital and how that shows up in auto.

54:35It's all to me just extremely exciting about what the role of a designer is and how you create context and systems across all these different environments that are consistent and are understandable to a user.

54:49Ellis:I think we're almost out of time. We have to deliver you back to your company. But a couple of final questions. One very short and stupid. one more thoughtful first one i'm big believer that you got to treat yourself a little bit you know to mark a milestone when you get a win did you treat yourself after the ipo at all what colors the lambo what can you tell us what did you get yourself uh a good night just some fro yo good night sleep good that's fair enough last question um everybody and their mom now wants a line as good as design is everyone's business which you did for the ipo so thanks so much for that.

55:27Ellis:How did you all come up with that? Where did that come from? I mean, I think it's really been our core message all along, you know, whether it's idea to reality, make design accessible to all. This idea that we can lower the floor, raise the ceiling. Like these are ideas around how do you get more people in the design process and how do you also elevate the role of design? And I think that in this world where design is the differential, how you win or lose, Like design is everyone's business. And I also was really honored with the IPO. I think this is the thing that stood out the most to me. This opportunity and privilege to, I don't want to say represent, but to show up as part of the design community and have a moment where design was going public.

56:20I mean, I think back to the start of Figma, where, you know, we're looking at the Bureau of Labor Statistics data. And so there was like 250K designers. And the conversation in the design community was like, does design have a seat at the table? How do we get a design seat at the table? It's like, okay, I think we're at a place now. We're not universally, but in a lot of places, design has that seat. And we need to elevate the voice design more. we need to show companies how they can win or lose with design because if they don't lean in they will lose and so if you're working in a company as a designer like your job is to speak up and if you're working in a company that is in value design either find another company or figure out how to make that change but uh to be there and to you know be able to advocate for design that was really meaningful to me i know ellis said that was the last question this is actually the last question.

57:15Oh, I see how it is, Alex. Your Twitter handle, we meant to ask this. What does Zoink mean? Is that how you say it? Zoink? Zoink. What does that mean? How did you come up with that? I tried to start or create a product in university. It's kind of like, I'd say Discord-y somewhat, but not like with a video game aspect. Some blend between Discord or Slack. It probably didn't have like clear enough product point of view. And that was the name for that product. And then that didn't really go anywhere. And so then I just co-opted it and that became a Twitter handle. Zoink. Follow Zoink. Z-O-I-N-K.

58:02Ellis:Thank you so much for having me. It's great to see you. Thank you for joining us. This was so much fun. My first time hanging with Zoink, one of the only Twitter handles that truly competes with mine. And there aren't many. You can find all my other work at sources.news. And Ellis, people find you on Twitter, right? At Hamburger. You can find me at Hamburger and at meaning.company. Though my email is on there. Only send high quality pitches, please. I'm back in media, but I don't want to overwhelm the inbox. Thank you. A reminder that we are a new show. So please follow, like, subscribe wherever you get your podcasts.

58:37We could really use your reviews on Apple and Spotify. We are in video also on YouTube. So please subscribe there. That will help the show a lot. And if you like this, please just share it with a friend or a colleague and let them know about access and what we're doing. We're excited to keep building this with everyone.

58:54Ellis:Access is produced in partnership with the Vox Media Podcast Network. We will see you next week. Bye.