Dylan Field is the co-founder and CEO of Figma, which makes a very popular design tool that allows designers and their collaborators to all work together right in a web browser. You know how multiple people can edit together in Google Docs? Figma is that for design work. We just redesigned The Verge; we used Figma extensively throughout that process.
So for years, people have been waiting on the inevitable Figma vs. Adobe standoff since Figma was such a clear upstart competitor to Photoshop and Illustrator and the rest. Well, buckle up because in September, Adobe announced that it was buying Figma for $20 billion. Figma is going to remain independent inside Adobe, but you know, it’s a little weird.
So I wanted to talk to Dylan about the deal, why he’s doing it, how he made the decision to sell, and what things he can do as part of Adobe that he couldn’t do as an independent company.
Dylan’s also a pretty expansive thinker, so after we talked about his company getting the “fuck you” money from Adobe, we talked about making VR Figma for the metaverse and AGI, which is artificial general intelligence, or the kind of AI that can fully think for itself. This episode takes a turn. I think you’re going to like it.
Okay, Dylan Field, CEO of Figma. Here we go.
Dylan Field is the co-founder and CEO of Figma. Welcome to Decoder.
Thank you. I’m excited to be on Decoder. Let’s make it happen.
We may or may not have already been in talks with Adobe at that point. You were telling me about the redesign, and it hadn’t yet lit up the entire internet. I was very intrigued to see it at that point and saw it a little bit afterwards. You were telling me about the process of it. That was really cool to hear about.
Yeah, I was dangling. This is inside baseball of how we get people to come on the show. I was like, “I’ll show you the redesign, if you come on Decoder.” Then you announced your Adobe deal and went quiet, we launched the redesign, and now here we are — it finally happened. I’m very excited to talk to you. There’s a lot of Decoder stuff to talk about. Figma itself is a fascinating product with a lot of fascinating elements to how it works, what it runs on, all of that.
Let’s start at the very beginning. Tell people what Figma is in a way that they can explain to their friends who haven’t heard about it.
Listen to Decoder, a show hosted by The Verge’s Nilay Patel about big ideas — and other problems. Subscribe here!
Figma is a platform for visual communication design. We start off with FigJam, which is our brainstorming, whiteboarding, ideation tool. It’s a great place to help run meetings with your entire team and also to facilitate them, make them more fun and engaging, and bring more ideas forth. From there we have Figma Design, which is a great platform to be able to really work with your design team across the company and to express visual design, interface design, and more. Finally, we go into the design and production side. There is not a lot to talk about there today, but I think over the next year there will be a lot there to speak about.
When you say the design and production side, and how there will be a lot to speak about, do you mean actually generating the code that makes a website?
One thing people don’t realize is that about a third of the people that use Figma on a weekly basis are developers. I just think we could do a much better job of serving their needs. Right now we are so focused on designers and people that work very close to designers in that co-creation process. How do we make it so that there’s a better way to go from design to production, to do that handoff and to make it so that you’re able to align really well with your developer counterpart? I think there is a lot we could do there, so that is something we’re thinking a lot about right now.
So just to unpack that for people, let’s say you’re going to launch a new product or something on the internet. In FigJam, your client, your product manager, your designers, and other stakeholders all get together and have a meeting.
You might be in HR, sales, or really anywhere in the org. Anywhere that wants to design a better meeting, that’s FigJam. We’re seeing it all across the org now, which is pretty cool.
Okay. So from there, you go to Figma, where the actual designers design a product. Then from there, you have to go write some code to ship the product that’s been designed through this process.
Exactly, idea to design to production.
You’re saying you want to go into that last stage, but you’re not there yet.
So Figma is the main tool. It’s the thing that designers use and the reason the company is famous. I imagine it is the reason that Adobe wants to buy the company right now, because it’s the product that is taking market share away from them. It all seems very obvious right now, in the way that every successful company seems obvious at the end of one path. When you started this 10 years ago, people thought this was a bad idea. Talk about that journey, where you knew, “We want to bring design to the web in the way that Figma works on the web.” What was the moment that you knew it was going to work?
Totally. There’s a lot there. Before I get into it, I’ll also add that I think I probably had as much conversation and heard as much excitement from Adobe about FigJam as I have about Figma Design at this point. We’ll get back to that later.
Ten years ago, when I was an intern at LinkedIn, I heard its cofounder and then-CEO Reid Hoffman say, “Most good companies have a good why-now.” It’s interesting to think about the different why-nows that exist in software. For a long time, I couldn’t figure out Twitter’s why-now. I realized at some point that you had SMS, so people were used to short-form content exchange, and you also had this cultural change, which was AIM away messages. Remember away messages? You could update your away message — it was basically like Twitter before Twitter.
Yes, in a less democracy-altering way. I was a master of the away message.
Seriously, people were trained into this behavior, so it wasn’t new behavior when Twitter launched. It was something that enough people already had a critical mass around and knew how to do.
It can be social like that, or it can be a moment in time. “This thing was regulated; now it’s not regulated.” Or it could be a new technology. In our case, we were set on new technologies. We were looking at WebGL and drones. My co-founder was very convincing, saying that we shouldn’t do drones, so we went to WebGL instead and explored there.
It was like, “Okay, WebGL lets you use the GPU, your computer, and the browser. What can we do with that?” So we started proving it out. He had made a bunch of tech demos already, and we started to look at it in the context of professional-grade tools and eventually interface design.
The more we built with WebGL, the more confident we were that this could be a technology we could use to go build a professional-grade interface design tool. But like you said, no one believed us. I kept trying to recruit people, and I found that if I didn’t show up and immediately open my laptop to show them the tool working, they just wouldn’t believe me. I couldn’t even get them to wait around in a meeting. They would do it, because they were polite, but people form their first impression quite quickly. If I waited until the end of the meeting to show the candidates the tool versus showing them the tool in the first two minutes, it was a way different conversion probability of whether they were interested in the next step. If you’re sitting there with me for an hour, you’re just like, “This guy is full of it,” versus, “Wow, this is maybe legit.” It was a very different proposition.
So people didn’t believe that you could run desktop-class design in a browser using WebGL.
Not at all, no. I think there had been too many examples of prior art, and not just in creative tools, but also in folks trying to build really good web applications. Outside of things like Google Docs and G Suite, there weren’t a lot of good examples yet. Some people in the know were seeing that things like Onshape were starting to come out. I think Aviary was a predecessor to Figma in some ways. It did good work, but in Flash, so it was kind of the wrong technology at the wrong time. It’s amazing how just a few years can make that difference.
I actually want to get into the web and WebGL and what that means for Figma as a company.
I think it’s really important but under-recognized. Figma is a web standards company in a very important way. But let’s come back to that. Let’s do the Decoder questions first. I think of these as the basic questions. How many people are at Figma right now after 10 years?
We’re nearing 1,000.
Have you scaled at a linear rate? Have you added 100 people a year for 10 years, or have there been peaks and valleys?
I would say we have roughly doubled year over year, most years.
Where do you double? How is Figma structured?
This last year we probably doubled across the org, with perhaps a bit more growth in sales.
Yes, in sales and go-to-market in general. We see the customer demand there, and we’re trying to scale to reach it. Most years we have doubled pretty much across the board, though there are little variations here and there.
What does sales look like at Figma? Do you have people cold-calling giant companies and being like, “Have you thought about designing on the web?” Or is it inbound, where you are answering questions and designing licenses? What does the sales process actually look like?
It is mostly inbound and outbound. The go-to-market is very influenced by bottoms-up. I think in general, enterprise companies are not rocket science. You have to find customers somehow. Maybe it’s through marketing, maybe it’s through sales development representatives that are, like you said, cold-calling or cold-emailing, or maybe it’s through your product qualified leads.
A lot of it is product qualified for us. People that are using Figma say, “Okay, wait a second. My needs are evolving. I’m trying to get more of my company on Figma.” They are raising their hands, saying, “How can I do that?” We are trying to help, so we try to make it very consultative, to make it so that people are able to have us as a trusted partner. Then we’re able to work with them to figure out the right way to expand that Figma usage.
Just to be clear, the bottoms-up sale is when people at a company just start using a tool without telling anyone, which is how I do everything. That’s the bottoms-up. You get the users because they are using a good tool, and the company is sort of forced to buy the tool for everybody. The other way around is to sell the product directly to the Chief Information Officer (CIO).
Well, you work with the company.
Right, you work with the company.
Usually there is some combination of people saying, “Hey, this is interesting,” and us trying to get the conversation going early enough so no one feels weird about it. We build that relationship and try to make it so people understand what Figma is. Why could it be valuable? How could it work uniquely with your company? So you work together with a design leader, core product leader, or IT — the champions in the organization who want to see this become the standard — and together you bring it into the organization.
I think Datadog has a really cool history of serving developers. It has some similarities, not entirely one-to-one, but somewhat similar to designers. They have a very heavy motion of expansion. Over time the account grows, and they have managed to figure out how to add more value through new products. How we build new products and new services for our customers is something that we care about too.
As CEO you have a lot of people reporting to you. Where are most of the people reporting from? Is it on the product side? Is it the sales side? Do you just have a lot of accountants?
Engineering product development and sales are probably the biggest areas.
Is that where all the recruiting focus is?
It’s across the board. You want to make sure that people have a great experience coming to Figma regardless. What I tell people is that it’s every manager’s responsibility to recruit. You can’t really depend on dedicated recruiters. Though hopefully, a recruiter will make it a lot easier and will make sure the experience is really good. Then it is HR’s responsibility to make sure once people are in the building, they have a really great experience of being a Figmate — that they are learning and developing, have great career options and good mentorship, and if they need feedback, they’re getting it.
The cumulative sum of all that is that our culture is the best it’s ever been, because we have invested and focused so much on that over the past few years. We’ve had days where our culture was not as good, mostly in the early days. So I have a comparator point there and feel really good about how Figma works. Hearing feedback from newer hires that Figma matches what they were sold has been great to hear. But there are always ways to improve too.
Actually, let me ask you about those early days a little bit. You have been very open in other interviews and talks that you’ve done that you had to learn how to be a leader on the job at Figma. You received some pretty hard feedback along the way. As you mentioned, your only other job before this was as an intern. So here is the Decoder question. You are 10 years on this journey and you just made a gigantic decision to sell this company. How do you make decisions?
Depends on the decision, always. I think if it comes to product engineering, design, and roadmap, there is always a tension in wanting to give people some guidance on how to be top-down in terms of the strategy. Where is the company going in broad strokes? What are the top three things we are doing that everyone needs to know? But it also needs to be empowering. You need to make sure at all times that people across the organization are able to understand, be on the ground, talking with users, and be really deep in the details and technical specifics of what is possible or what is changing locally for their product area.
You want to meet somewhere in the middle of, “Here is the local feedback and the ideas that we have,” and the top-down, “Here is a point of view on strategy and where we should take the company.” Somehow those have to meet, and you need to be able to resolve tensions and work through any conflicts that arise. Education and context-building are areas that can really help with these things.
When it comes to making decisions in general, especially when you don’t have enough data, it’s not like there is one framework that is the magic framework. People talk about regret minimization, where you imagine yourself in the future and ask yourself which path you would regret less. That’s a really good framework, but it turns out it’s not the only one. You can come up with all sorts of frameworks for any decision you are trying to make.
An example is when we were deciding, “Okay, do we sell the company or not?” First of all, I was very lucky to have people around me, like my board, and once we eventually got to a certain point in the process, I was able to be more open to my exec team and talk with them as well. But ultimately, at the end of the day, it is my decision. I own the decision. I tried a multitude of frameworks, and each one led me to say, “This is a good call.” I won’t go into the details of all those frameworks.
What were the top two frameworks?
I won’t go into the details on all of them right now.
I didn’t ask for all them, I asked for two. What were the top two?
But I think what’s interesting is…
No, wait. What were the top two?
“Obviously monetary is a framework, right?”
Well, I mean obviously monetary is a framework, right? But it’s not the only one.
I knew it was the first one.
Impact. Impact is a framework. Will this lead to us having less impact? I think regret minimization is one that you could certainly think about. Anyway, I can list off 10 more, but again, I don’t want to go into the details there. Making sure that we were able to continue building a really important culture was definitely a consideration, so was attracting the type of people that we had been attracting. I think it would’ve been spooky to me if we didn’t think that was the case. I built a lot of conviction that it was. Thinking through all these different frameworks, not only did it highlight where the gaps were in my thinking and what things I had to figure out, it also highlighted important conversations I needed to have with the Adobe team.
The more I went into those conversations and brought up my concerns, the more I gained conviction that this was a really good decision for not just the company, but for our users and our community. It also built our trust and our relationship with the team. There was never a moment where I was like, “Oh, I have this framework that I’m trying,” that led me to a decision of, “We shouldn’t do this.” That was a sign. I think if you do get to a point where you have different frameworks that lead to different answers, you have to find some meta frameworks or a way to sort through all of them, which is maybe a little tougher.
All right. Let me give you a framework where it doesn’t work out so well. It’s shocking to me that you are out doing media in the middle of this acquisition. Here is a framework that the government is going to use: The Federal Trade Commission or European regulators are going to say Adobe is a gigantic corporation with dominant market share amongst designers in the world. Figma represented real and meaningful competition to Adobe, and was taking share away from Adobe in a variety of contexts. Adobe buying Figma reduces competition in the market. I have heard this from people who do not care about antitrust law, “Crap, this thing that was competing with Adobe that I was excited about is going to get swallowed by the board.” I have heard this from regulators who care a lot about competition from the market, and I have heard it from your local podcast hosts who think that tech companies need a lot more competition, especially big tech companies.
That is the framework that I think comes out as a “no,” right? You just listed off a lot of people you obviously care about, who are on your team and in control of the destiny of this company. You are going to cede some of their control over the future of Figma to a very large company with its own needs and aspirations, and might snuff out the thing that makes Figma such a good competitor. Whether or not you think that is the right policy choice, the right financial choice, or whatever, the thing that is exciting about tech is that there is always a disruptive challenger that shows up and knocks these giants off their pedestals. Figma is that thing. Isn’t that the framework that says, “No, you shouldn’t do this. You should see if you can actually knock Adobe off the map?”
I think for the user community, that was a critical component of the decision. That was probably the first audience that I thought of as we were thinking through it. As I thought through that decision and what we could do for the community, it actually led me to becoming more convinced that we should do this versus less so. Part of that is because there are a lot of things we don’t do in Figma today that are needed in the product development process.
If you think about product development, it’s not just pixels on the screen. You have a range of media types you are going to bring into the product development process in today’s world. Think about the Verge redesign for example. It’s not like you just use static imagery; you use video, and in the future you might use 3D content.
But those tools exist. We can pay for those tools too. Why does Figma need to be a part of those tools? It’s not like JPEGs and MP4s are difficult to move around.
I think that there are really good advantages that you can have as someone using a set of tools if the tools actually connect really well in one process. It’s like what we were talking about before. Figma is going from idea to design to production, and we are really trying to make the product development process as good as possible. What else can we do if we are able to bring more mediums into that? I think there are a lot of things that become possible there.
Name a thing. Name a specific thing. Adobe is a vast company. I think most people listening to this do not know that Adobe has a gigantic ads business. Many, many things can plug into whatever Figma is doing. Give me a concrete example of why being owned by Adobe is better than a partnership with Adobe.
Photo editing is a great example of something that I think we could do a much better job of in the product design process. There are a lot of times where you are using Figma plus other tools for photo editing, many of which might be owned by Adobe right now.
Why does the ownership of the tools matter?
It is extremely helpful to be able to go seamlessly into your workflow with a new modality, versus bringing assets and doing a complex import-export process into another tool. At the end of the day, our vision for Figma is to make design accessible to all, to make it so you are able to get things out of your head and onto the canvas faster. How do you do that while making it so that you are able to transition across these modalities quicker?
So you’re saying you’re opening Figma? I am just trying to be very concrete about this. You are opening Figma, you have designed some interface, you have a photo element in it, you click on it, and you get the full set of Photoshop tools right there in front of you, instead of having to open Photoshop and create a file.
I think there is a huge opportunity to bring these capabilities from Adobe Creative Suite and Creative Cloud into Figma and utilize them more, and to make it so that you are able to go and somehow transition across these different creative modalities to have a more seamless way of working.
By the way, we aren’t just talking about product design at the end of the day. I think that there is an opportunity to scale the impact of Figma much more broadly. We can start to make it so that you take the web-based tech, the different methods of collaboration, and the platform we have developed, and apply that to many other creative areas as well.
Finally, there is a huge opportunity to think about how creativity and productivity come together here. Like I mentioned at the start of our conversation, FigJam is not just a tool for designers. Isn’t this being used across the entire organization? It turns out that visual canvases are actually really helpful for tons of people to use. It’s a way for teams to come together, and it’s a way for people to run meetings.
A lot of people are trying to get into the creative productivity game. It’s heating up in competition a lot.
We are in a world where a lot of people are trying to get into the creative productivity game. It’s an area that is heating up in competition a lot. For example, Microsoft Designer just came out.
I mean, I’m rolling my eyes because another thing that just happened is that Adobe announced it was going to be purchasing Figma, and that it was going to sunset Adobe XD, which is the competitor to Figma.
I don’t think they announced that.
You can add and you can subtract from the market, and it is net zero that Microsoft is going to bundle some more stuff into the Office Suite. At the end of the day, Adobe’s competitor to Figma is going to go away because the company is going to own Figma. I hear you, in that there are ways to make things more seamless, but I’m still not clear on why Adobe has to own Figma to make these things happen.
I’m trying to highlight that there are a lot of things we can do here to make the customer experience of using Figma way better as a result of this. It’s really good for our community, really good for our user base, and really good for the designers in the Verge team.
We did our redesign on Figma. It was fun to use.
There is stuff in this structure that is really interesting to me. Everyone has been very clear that Figma will be an independent division inside of Adobe when this is all said and done — not a subsidiary, but a division of Adobe into itself. You’re going to report to an Adobe executive. Who are you going to report to?
What is his role at Adobe?
He’s the chief business officer of Creative Cloud and Digital Media.
This is just a really small question. Did you get to pick? I have never been acquired. Did you look at a sheet with your kitchen cabinet of people and say, “All right, I’m going to report to this person,” or did they tell you?
I mean, look, I have had a great relationship with Scott [Belsky, Adobe’s chief product officer] since the start of Figma.
I think he reports to the same person, right?
Yes. I actually tried to convince Scott to be an investor early on. Turns out he was getting acquired, so I didn’t get him as an investor. We have continued to be friends and build that relationship over the last decade. In this case, it was David who was really bringing Figma in, and so I just assumed that was the structure the entire way.
So there was never a conversation. You never said, “I actually want to report to the CEO of Adobe”?
How much do you stand to personally make from this deal?
Depends on the stock price on any given day. Look, that has not been my focus.
You’ve done the math. What’s the low and what’s the high?
I don’t have the spreadsheets in front of me. It becomes very complex.
Are you going to be rich when it’s done? Like private jet rich?
I feel very well-off already. I have nothing to complain about, and I feel super blessed and lucky. I couldn’t be more thankful to be where I am at 30, regardless of this deal.
How much will your employees get?
I assure you, every one of your employees who has equity in this company has done the math. Is it significant? Are you going to mint a bunch of millionaires?
I hope that we’re able to make it so the team is able to do whatever they want to do, and to move on if they want to. Hopefully they will stay at Figma a long time, but they can also go make a great impact in whatever community they’re in. I think we have an incredible, super creative group of people at Figma. I’m really excited to see them be even more empowered.
The reason I ask all those questions in that way is because I think you are going to stay at Figma. My feeling is that you are very committed to this thing you started when you were a very young man. But when people get the “fuck you” money, they might leave. They don’t have to stay after the deal closes. You have to retain great talent that is going to enter the Adobe ecosystem, be able to leave, and even just do things inside of Adobe. It has to come into a much larger corporate structure that is going to have all the attendant Byzantine problems that come with it, as well as the resources. I don’t want to say there are no pros to that, but there are some cons.
The reason people join startups like Figma is the promise of this exit payday, which you will not be able to offer folks anymore. Have you thought about that life cycle? “Okay, I’m going to graduate out a bunch of people who are going to get the money — they’re going to get the bag and walk. Then my recruiting pipeline is going to dramatically change because the pot of gold at the end of the rainbow will be gone.”
This is wild, you don’t have to believe me if you don’t want to, but the number of people that started to apply to Figma after we announced this went up tremendously. Which is counterintuitive, because a lot of people have that mental model you just described. That is just the fact. I don’t know exactly what the percentage lift was and how sustained it was. I mean, it probably went down after the acquisition announcement a bit, but we have been getting people that are really excited about doing this at the combined company.
Well, the deal hasn’t closed yet, right? If you get a job at Figma today, you still get Figma equity that might get paid out.
It’s pinned to Adobe equity.
If it ends up going through, then it becomes Adobe equity. If not, it’s Figma equity. I think that the chance of going and building this context is something that people are actually really excited about. Just looking at our team, we very deliberately tried to hire people that are super missionary throughout the company’s life.
We haven’t always been the highest-paying company. And you can always find a way to stretch a band and not be fair in your compensation, but we have been very deliberate about making sure that our compensation bands are fair and equitable. Through that, I think we have a lot of people who really believe in what we’re doing, not just people who are trying to make the most money. It turns out that this has become a good outcome and people are going to do really well in aggregate, which I’m super happy about. Our employees are in it because they really love creativity, design, and building things, and they love doing it for other designers and creatives. I think they’re stoked.
I think you probably know this, but Tony Fadell is a friend of the show and a friend of The Verge, and he was on Decoder recently. He has a book out called Build. It’s uplifting in many ways, except for the chapters about the Nest acquisition at Google, of which Tony has nothing but unreserved scorn. Nest was a company that had its own culture. He was building it, Google bought it, they dumped a bunch of money into this company, they turbocharged investment, Google culture seeped into the company, and things went totally sideways. I don’t know how familiar you are with that story, but it is a pretty common story. You do the acquisitions so you can turbocharge growth with other people’s cash.
Oh, there are really positive stories too. GitHub and LinkedIn are some examples. Pixar is a really good example.
Pixar is a different direction. Pixar ate Walt Disney Animation Studios.
The culture of Pixar took over Walt Disney. Here’s my question for you. What are the safeguards against the negative outcome? The negative outcome is well-known. The positive outcome is well-known. Do you have a commitment in writing that Figma will be independent? How does that work? What is that conversation like?
There has been a conversation with all levels of Adobe and everyone related to the deal about the autonomy of Figma, our goals, and how we’re going to execute against them. Honestly, the open-mindedness of the Adobe team to think about this in a very unique way was part of what got us comfortable. They were amazing with it. The proof is in the pudding.
Do you have a contract that says Figma will be independent?
We have an operating model doc where we talk about autonomy within Adobe. That is not legally binding, but it is a plan.
Not in those words. We have an operating model doc where we talk about autonomy. That is not legally binding, but it is a plan. It’s important to think together about how we will make this combined company really amazing in the long term. Adobe is really wanting to set this up for success. Not just the management of Adobe, but having just been to Adobe Max last week, it’s the greater Adobe team as well.
I was really heartened by that reaction. Getting to meet them and spend time with them, we felt very welcomed. I’m really excited to spend that time with their team to learn and figure out how we go from here. How do we keep building and do the best thing for the customer and for that vision of making design accessible to all? And not just interface design, but more globally, all types of design, creativity, and productivity. There’s a lot to do there.
I have to ask you about the tweet, then I promise we’re going to move off of this and talk about WebGL. You know what tweet I’m going to ask you about. In January 2021, somebody said, “Give it 15 years and Figma will replace Adobe.” You replied, “The goal is to be Figma and not Adobe.” I have to point out to you that you’re about to become Adobe.
“I still stand by that tweet. I don’t take it back.”
Again, we are operating autonomously. We are going to have our own offices and our own culture. We are definitely going to have a lot of people that are coming from Adobe into Figma. We are going to interview people, so it’s not like an automatic rubber-stamp “yes.” We are also going to be interviewing people from outside of Adobe and Figma to come in. We are going to continue to have our values.
It turns out our values are actually very similar to Adobe. That is one thing I learned through this process. Our mission is very similar to Adobe. Literally, you put the values side by side and it’s like, “Wow.” Turns out, design-oriented creative people are attracted to these companies. Having met people at Adobe Max, a lot of them feel like Figmates already. There is probably a lot I didn’t know about Adobe at that point in terms of how close they felt to us in demographic employee makeup and psychological makeup. At the same time, we definitely want to preserve the best of Figma. I still stand by that tweet. I don’t take it back.
It’s a pretty bad market right now, if you look at the economy. This is not the market for companies to go public in. This is not the market to go raise money in. In all of your decision-making frameworks, did that factor into it? “Hey, it’s going to be really hard to get more money. We’re kicking the ass of our competitor Adobe a little bit. They’re ready to give us a bunch of money to accelerate this.” Was that one of the decisions here?
We were on a path to continue to be independent as well. We’re cash flow positive, and we’re doubling revenue year over year. We were in as good of a state as possible. It was really about the merits of a combined entity, what we could do together, how we could be useful, and how we could make this product better for our audience. That was really the thing that was weighed the most heavily.
Figma is a web product built on WebGL. You mentioned at the top of the show that WebGL’s technology lets you do 3D rendering on the web. The web is kind of under attack. On top of everything else, this primary surface that you distribute your product on, has maybe been perpetually under attack. First Microsoft tried to kill it, now maybe Apple’s trying to kill it, and who knows what Google’s trying to do with it, but it always feels a little unsteady. I know a bunch of designers who use Figma on Chromebooks. They don’t have Macs or Windows PCs, they just have nicer Chromebooks. You can use Figma on Chromebooks. The web is fundamentally Google’s revenue platform. Very few other companies make money on the web writ large. You make money selling seat licenses and enterprise software. You distribute on the web, but you don’t make money on the web. Google is one of the few companies that makes money on the web.
That means the Apples and the Microsofts of the world are constantly trying to push back against this and capture more value on their platforms directly. When I say a perceived threat, specifically what I mean is that Apple continues to encroach on the web on iOS. The Safari browser is a big deal, and there is a standards battle between Apple and Google in particular that gets pretty heated quite often. You need to build on those standards. Do you watch that from afar? Do you just hope that the web persists, grows, and innovates in a way that lets you build new features? Or are you like, “Screw it, we better have some native apps in the background”?
What are examples of the standards battles that you’re referring to?
WebGL itself. It needs to continue innovating, and it needs new standards-level capabilities that leverage new hardware capabilities and different computers over time. Apple in particular probably does not want to open up the entire graphics pipeline of the M Series chip to the web. They want to save that for their own native apps because a lot of designers run Macs. How do you get through that as a company?
Well, it turns out that you can make our software really efficient, even on WebGL, by mapping to low-end chips from phones that were from five years ago. Can we make it even faster with even more access to the hardware and even better standards? Absolutely. At the same time, if you’re efficient about your code and about the way you write your shaders and manage memory, it turns out it’s pretty incredible what you can do.
It’s a story that persists across computing for a long time. Look at what people were doing in the demo scene in the ‘90s. It was amazing. One thing that seems to happen with the way that programs and software evolve over time, is that you start to become less disciplined about the way you build software. There are more abstraction layers that exist over time.
As more abstraction layers exist in the software you’re building, each one incrementally slows down a little bit year over year — especially as they become a sort of black box that’s not actually introspected into and people rely on them in almost faith-based way. “Of course that abstraction level is good. We don’t touch it anymore. The person who worked on it is no longer here. Don’t go there. The code is really messy.” Then of course it degrades a little bit year over year.
At Figma, we try to be very disciplined about how we structure the code base, making sure that we continue to refactor our code year over year, that we clean up our tech debt, and that we take out old systems that are no longer needed. By doing so, we have been able to keep things really fast and efficient.
At the same time, maybe every month or two, something happens. I see some uptick on reports from our support channels or through Twitter, or I’m talking to someone and I realize, “Oh shit. With some chipset or model of computer somewhere, or after some recent update was launched, we introduced some regression. Figma slowed down in some way, and someone is unhappy.” Our team has a really good culture of identifying these issues, diving into them and talking with users, and figuring out what’s going on and fixing it.
We also do a lot of good performance testing to make sure that we are able to continually monitor and understand if we are still on-track or if we are off-track somehow. Did we ship something that’s bad? By having that really high standard from the start and maintaining it over time, that hasn’t been an issue for us. Of course, we are always looking for the web to be better, but I think the fact that we have our render based in WebGL gives us a lot of control over what we actually write to the canvas. That means that we are sort of on this weird, independent path that is a little bit different than having to run every single thing in web standards.
Now, that said, browsers do occasionally ship something where they’re trying to fix something over here or in a different spot, and there is some kind of inefficiency that’s added or an issue that accumulates. We have worked really hard since the start of the company to maintain and build relationships with browsers so that if that happens, we’re seeing in a canary or beta build that there is an issue. We are then able to flag it to a browser vendor way ahead of time, tell them that issue has occurred, show them the reproducible steps, and help them patch the issue. Doing that over time, having those relationships, and hiring people that worked on browsers in the past, who have a deep understanding of the browser stack, has also helped.
Do you think that you need to ship native apps on Mac, Windows, iOS, and Android? I mean there are native apps there, but again, the primary canvas here for you is the web.
There’s definitely an opportunity to do more natively in the future for Figma. For example, there’s pen and touch input. Even if web standards improve a ton, there is probably some better feeling of lower latency if you do more of the handling of those gestures natively than on the web. Then of course, there is always that next thing of, “Well, maybe we could do this thing natively instead of on the web.”
It turns out that “on the web” is good enough most of the time. However, I think we are constantly trying to figure out, “Okay, what’s the next thing that we can start to do natively? What’s the progression of that over time?” I do think that eventually a tablet version of Figma will be really useful.
One thing we’re excited to announce very soon — and I’ll give you a little sneak peek — is our mobile comments that are coming out. Making sure we could add that in a performant way was really important to us. If you think about it, there is a lot that we can do with comments on mobile that is unique to the mobile form factor over time. We’re super stoked about that.
We also have a bunch we’re doing with FigJam on tablets, and we’re adding more functionality. For example, we’re making it so you can play music while you’re in FigJam now. We’re making it so that you can do better integrations with Teams, GoMeet, and Zoom, as well as add sections, tables, and better text formatting. People have asked for these things for a long time. We are able to add this stuff even though we’re not having a ton of native surface area with FigJam, but we are trying to improve that native surface area over time.
It’s hard to go get this stat, but my instinct is that Figma is one of the biggest and most lucrative WebGL clients out there. There aren’t a lot of others, so it seems like a safe bet. Figma is a big and lucrative company. The product is excellent, a lot of people use it, and there’s just not a lot else. You are built on this thing that you depend on, maybe in a way that the other users of WebGL don’t depend on — they’re mostly games, from what I can tell.
I think there’s going to be a lot more people that use WebGL over time. I’m excited to see so many people betting on the browser as a platform. I get emails every day of people that are building Figma for X (you fill in the blank). People are doing it and they’re betting on the web.
If that was the only thing that we looked back on in 10 years and we said, “Wow, because of Figma existing there are more people on the web,” that would be something that would make me feel really proud. I think and hope there will be a lot more too. It’s really cool to see so many people betting on the web now, whereas before it was something that people just didn’t believe in.
I mean, I’m a huge proponent of the web. I think more people should bet on the web every day. That comes and goes in waves. Five years ago, if you weren’t shipping a native phone app, you were basically not in the conversation to get funding. Why do you think the web is back in vogue?
I think that the power of the URL is one where you’re able to easily share content. I can pass content from me to you so easily with a URL. Having that URL as this global address that people can find information off of forces choices for a medium.
So for example, if you don’t have multiplayer editing on the web, it just feels wrong. That doesn’t mean you have to be on the web to have multiplayer editing. You can do it natively too. But they go together really well. It also goes together really well when you have a URL on top and can share it with people. There’s this cluster of features and functionality I think that starts to merge, where it just feels better to be able to navigate from links that you’re encountering on email or Slack onto a website, where you’re able to have this rich content experience and then in the browser, work.
The web really captures the ethos of my generation... transparency, access, and collaboration.
It really captures the ethos of my generation at least. I can’t speak for my generation…
Be the voice of a generation, Dylan, you can do it. You have the name for it. Do it.
Some of the things that I feel are really important, and I think a lot of my peers feel are important, are transparency, access, collaboration, and working together. I think the web really embodies that ethos. A lot of desktop software didn’t traditionally embody that ethos, it was more siloed. It’s not like there is one thing that makes the web so powerful, but being able to have that cluster of functionality of values ends up really having an amazing effect.
Do you make strategic decisions at Figma around the market shares of different browser engines on desktop? Chromium is really dominant. Microsoft was forced to effectively concede and move Edge to Chromium on mobile Safari. If you don’t address mobile Safari, you don’t get a whole bunch of people, because there are no other browser engines on the iPhone. Is that something that you have to think about as you make technical and business decisions?
Not really. If that conversation is happening, it’s not rising to me. Like I said before, when there are issues, we have to pay attention to them. We have to be in communication with browser vendors. It’s probably the case that there are small bugs or issues that exist, and depending on market share, that impacts the priority at which we attack those and push browser renders to try to fix those things. That’s probably the extent of it.
Does anything on that level come up to you? Do you ever get a phone call that’s like, “I need you to call Sundar Pichai and get him to fix Chrome”? Does that happen to you?
I’m laughing at the Sundar thing, because I literally have an email thread from a long time ago — I forget exactly what it was about at the time, but this was like 2015 — where I was trying to get some issue fixed with Chrome, and I got an intro to Sundar who was not yet CEO of Google. It went all the way to Sundar, then it went through this chain of product managers, and eventually it ended up with my friend, who already knew about it because I hung out with her all the time. She was like, “Oh, hi Dylan.” I felt like a total fool for escalating that one.
How often do you escalate things? Does it happen, or is that mostly on the technical side of the company?
It hasn’t happened much recently, because we now have such good relationships with browsers and they have been so helpful. In the earlier days, we just didn’t have the attention of these teams, partially because we weren’t at scale yet. If you have more people that are being impacted by an issue, that impacts the time in which that issue is dealt with, which makes sense, right?
Yeah. You mentioned bottoms-up. The reason I’m asking is because I wonder, does it get easier when half the company is using your tool and they are running into problems?
Totally. Oh, absolutely yes.
I have to imagine that that is the case at Microsoft, Google, Apple, and all the rest.
For sure. When Google adopted Figma, I feel like they definitely cared a lot more about Figma running well, which also makes sense.
I just spotted my review unit of the Meta Quest Pro, it’s just sitting right over there, so I need to ask you about VR. Meta thinks that we’re all going to be hanging out in the metaverse, winking at each other and connecting emotionally because we can see each other’s faces. They have a little bit of a whiteboarding tool. It’s very early. Is that something that you see being competitive with you? Is that something that you want to participate in? Is that just a distraction?
I think it would be great participating in that one day. Right now, we are still early in terms of how many people are using VR, but it seems like an exponential graph, which is really exciting as a new platform. But we’re at the early stages, so I think we’ll probably spend more calories thinking about that when that exponential curve goes up more.
You’ll believe it when you see it is basically what you just said.
No, no. I mean, I like VR. I was on paternity leave early this year, and I spent a lot of time in VR. It’s really good. I’m excited to try the new Quest 2, by the way. I think as a gaming platform, it’s excellent; I think as a productivity platform, it’s emerging. There is still stuff that has to be solved around how much time you can spend in VR without getting a headache, motion sickness, or fatigue. I personally don’t get motion sickness or a headache, but I know people that do. I still feel some fatigue after half an hour to an hour in VR.
Now I haven’t tried the new headsets yet, so it’s possible that it has been fixed already. When I tried VR in 2016 or 2015, it was for like five minutes. After that I felt like I had a crazy night out drinking. Now it’s such an improved experience, and I feel like it is going to keep improving, just like there are going to be more people adopting it. It might be exponential. As it gets to that point where you can spend a few hours in VR and take the headset off and say, “I feel fine,” I think the possibilities unlock a lot more. I’m actually quite bullish on VR, and I’m really impressed with Meta’s dedication to the platform. I think it’s quite visionary and bold. It’s a great example of someone trying to will a new technology into existence.
How many engineers are you paying to build VR Figma right now?
Right. I love my Quest 2. I use it every day, but I use it as a fitness platform to work out. If I try to use it for collaboration, I would tell you that being on a Zoom meeting with FigJam open on a tablet, just so I can draw at people, is immediately more evocative to me than all of us being cartoons in the metaverse. I’m just wondering where you are calibrated on that spectrum.
When you take these paradigms and you try to retrofit them into a new space, it’s not always right. It’s a good starting point, but I think that the real opportunity is probably that task of having a meeting, of brainstorming, of ideating, of trying to collaborate in a space. Those are the things people are trying to do in FigJam. What does it mean to do that in VR? I think that’s a much more interesting question than, “How do you take this 2D primitive and put it on a plane in a 3D space in VR?” When we get to the point where we see the traction of the platform and, like I said, the ergonomics are a little better, I think it’s going to be really exciting to explore this.
By the way, we are just now building tablet for FigJam and for Figma, and tablets have been around for a long time. It’s not like this is a new thing. It takes a lot to remove people from the surface that they’re already used to and put them on a new one. You have to demonstrate some benefit. That said, as a product person and as someone who cares for design, I love the challenge of being able to work in a new medium, and to think about the ways to solve that product challenge in that new medium. That’s really exciting.
By the way, a lot of what you’re saying around VR, “Oh yeah, I use it for fitness.” I’m like, “I use it for gaming.” It reminds me of crypto in some ways.
Oh God. This is another hour. You have another hour for this metaphor?
I have time. Do you have time?
Let’s do it.
With crypto, for a long time it was like, “Yeah, people are just gambling on this thing. What’s it good for?” It was like, “Well, maybe there’s a store of value.” Then it was, “Well, maybe there’s this DeFi thing.” Now it’s like, “How about NFTs?” Perhaps next there’s gaming.
I’m not saying that the crypto industry is without sin. There are a lot of things that are pretty fucked up in crypto. But I’m also saying that what you see when you have a new technology or platform emerge is that people discover the use cases of it one by one. At first it seems like it’s nascent, that it’s actually worthless, et cetera. Then over time you realize, “Holy crap, there is a lot going on here, and there’s so much that we can do on this new platform,” like machine learning.
I remember when we were starting Figma, and Chris Olah was in my Thiel Fellowship class. He later went on to work on Google Brain, and was responsible for all sorts of interesting research into how neural nets actually work. Then he was at OpenAI, and now Anthropic. He was showing me these very hacked-together demos in his terminal, where he was shoving into AWS, running on GPUs, very basic neural nets. He’s like, “This is amazing.” I was like, “Yeah, it’s cool. You can classify some numbers. I’ve seen the same demo before.”
What I was missing was the exponential curve of that technology, and the fact that it was actually a paradigm shift from more structured machine learning models and people trying to use hardcore math to make AI into this new representation and new way to actually solve problems. Then you get something like transformers on top of that, and suddenly so many more applications become possible. We are at this phase now, and not just with Generative Adversarial Networks (GANs) and diffusion models, where people are actually taking some real, interesting tech and applying it to entirely new industries and new areas. The use cases are proliferating.
Let me push back on this just a little bit, because this is fascinating. AI, machine learning, and now generative image models are all a very cool party trick, right? I know some newsletter authors who are like, “I’m going to illustrate my newsletter with DALL-E.” That’s cool, it’s very neat. There’s some stuff that might get built on top of it, and there are some machine learning applications that are indeed very useful to large industries all over the place. They are capabilities that were layered into things people are already doing, or capabilities that were layered in to make something cheaper. Image generation is actually just going to make something much, much cheaper for people.
In VR and crypto, they are applications in search of a market. I don’t think DeFi is an actual application yet. It’s a conceptual market that might exist if crypto works the way we want it to. I mean, I have the VR headset; I use it for games and fitness. It’s searching for its big consumer application, and they keep firing things at the wall. I think that’s a real difference. I could come to you and say, “Okay, in five years I’m going to type ‘music player app’ into Figma, and Figma is going to call up DALL-E and fire out a sketch of a music player app that I can then further manipulate.” I can see that sitting from here, whereas I couldn’t tell you what NFT capability you should add to Figma today.
First of all, it’s pretty amazing. A year or two ago, I think, if we were having the same conversation about AI/ML, you could very easily have said, “Hey, this is research in search of a problem. This is just a bunch of people who are doing cool math, and it’s not clear that there’s actually something here yet. What are these models actually good for? You got GPT-3 sure, but what recent advancements have we seen? In GPT-3, large language models have kind of been tapped out. What are people building on it?”
It turns out, even on GPT-3, let alone the new improvements we’re seeing to large language models, we’re actually seeing incredible things being built on those now, and they’re impacting a lot of different industries. The timing of the question ends up mattering.
There is something to having scarce digital goods. Do I think it was overhyped and there’s a bubble? Absolutely, I do. I think that’s the curve that most technology is on. In fact, I think that we’re probably going to see an even bigger bubble around a lot of things that are called AI as a collection of things. If I’m seeing the VC investment patterns that are going on, AI feels a lot like crypto right now.
From before the crash, just to be clear. Every technology has a hype cycle, and with VR, at the end of the day it’s like, “Okay, what are the behaviors that are new that you’re doing?” Did you increase your fitness spend when you got an Oculus Quest?
Yeah, I really did. It’s a real thing that happened. You should listen our episode with Chris Milk from Supernatural. We talked about the whole thing.
I will. I’m sorry I missed it.
It’s the thing that drives VR adoption. It’s a completely under-reported phenomenon that the thing that drives VR adoption is fitness. I think it’s fascinating. Here is this thing that’s supposed to disembody you into the internet, and it makes people consider their bodies in a much different way. It’s great. I totally buy it.
It is fascinating. It’s also interesting that the GDP has gone up too. There has actually been a net value creation there. Crypto is interesting because it’s a little bit different when it comes to value. There are a lot of things in the crypto ecosystem that I wouldn’t call Ponzi, but they have these Ponzi characteristics. The more money that flows in the system, the more the asset price goes up; the more money comes out of the system, the more the asset price goes down. It’s wrong to say that collectibles markets or art markets don’t have value. At the same time, there’s something about that characteristic that people realize feels a little weird. Then AI could be very deflationary, by the way.
“All technology is deflationary, but depending on how much more efficient AI makes things, it could cut into the time that people are being paid for.”
Well, all technology is deflationary, but depending on how much more efficient AI makes things, it could cut into the time that people are being paid for.
I see what you mean. So let me ask you this directly. I mean, this is your business, right? You make people more productive and you make designers more collaborative. Do you foresee a world in which Figma has generative tools in it, so that I really can type in “new Verge website” and Figma pops out a mock for me to then iterate on?
I think that there are a lot of ways to do that. What I’m excited about is how you make it so designers are able to explore more of an option space, and then bring in their expertise as well. It’s like getting inspiration from a model, just like they get inspiration from looking at other designer’s work on Dribbble, Behance, or in Figma’s community. A generative model is your operating lead in the space. It’s kind of the same thing in some ways. People may not like me saying that, but I think inspiration can be drawn from lots of places. How do you work with an AI agent versus an AI agent replacing a designer? I think across all creative work, that is going to be a really important thing to figure out.
Let’s just take the bear case, which is that the tech we have today is the tech we get. There’s no evolution. Never mind that every day on Twitter I’m seeing two new papers that are mind-blowing. Let’s just say that stops tomorrow. So the bear case is what we have now in AI. If we can take that and just apply it to new areas of the economy, that is already going to be very disruptive. We would continue to see the base case for some time, then it asymptotes. At some point, we don’t see much more around AI for a while. The capabilities kind of max out across a few domains. Again, it’s even more disruptive. In a bull case where there’s AGI [artificial general intelligence], I don’t even know how to think about that. I think that the world could get really weird, really fast.
Then we’re all out of work. That’s the most deflationary.
No, work might just change in a very fundamental way.
If there’s AGI? Yeah, we’ll all be working for the robot.
I made a tweet recently. I’ll read it to you.
Oh my God.
I did a poll. I said, “AGI has arrived. Everyone is saying it’s way smarter than humans and well-known experts claim it’s ‘human aligned.’ Recently political leaders asked the AGI to propose an ‘optimal system of government,’ (and resulting constitutional amendments). Would you consider the proposal?” The options were yes, yes if still democracy, no, and see results; 21.5 percent said see results. The remainder: 35.3 percent said yes, 25.5 percent said yes if still democracy, and 17.7 percent said no.
I thought that was really interesting. First of all, I was surprised that so many people were open to nondemocratic systems of government, which is not something I would have expected going into that question. Also, I thought the no answers would be higher, because if you think about it for a second, if AGI existed, it would be able to basically market to you whatever idea it had in a way that would presumably sway your opinion. The fact that less than 20% of people said no, it’s like, okay, people are open to the idea of an AGI-proposed utopia.
Now how do you know if it’s actually human aligned? There are a lot of alignment issues that come up with that too, that I think are just not even on people’s radar yet. Again, AGI is too complex. I don’t know. It breaks my brain a little bit.
Yeah. I think you and I are going to be lucky enough to not have to experience it. I worry about our kids. All right, we have to get out of here. I can’t believe we got all the way to AGI in this interview. Dylan, you’ve given us so much time. I really appreciate it. I think I know what’s next for Figma, but what should people be on the lookout for?
So much. There’s going to be a ton that we do around Figma Design over the next year, and I’m excited for people to see that. We’re just at the start for design systems, and FigJam is in collaboration with that. Like I said, we are launching a bunch around collaboration today. There will be new areas where you’re able to add nouns to the FigJam canvas, like tables, music, voting integrations, mobile comments, and notifications. There’s so much more coming for FigJam. We’re trying to make it so that you’re able to have A-sync/B-sync, and make it so that your means are able to run better, whether you’re in the file at the same time or not. There will be lots more over the next year about that.
Going back to the start of the conversation, how do you navigate that process and product development, from idea to design and production, and make it a really great process? To do that handoff, whether it’s a design system or your design file, with engineering more effectively and make it a great experience for developers in Figma, that is another area that we really care about.
All right. This was great, man. Thank you for getting into it. I loved it. You have to come back soon. We’ve never had anybody in the middle of an acquisition, so now we have to follow up after and see how it went. So come back soon, Dylan. It was great talking to you.
I look forward to it. It was great to see you, and thanks for having me.
Decoder with Nilay Patel /
A podcast from The Verge about big ideas and other problems.