We're going to get to a world, we're already kind of there, where good enough is not enough. Good enough is going to be mediocre. And you're going to need to differentiate through design, through craft, through point of view, through brand, through storytelling and marketing. And I think the people that internalize that now, they're going to be winners.
That's my point of view is that this is what's going to matter at the top of the stack. And if you don't internalize it now, you got an issue.
Dylan, it's a pleasure to have you here. Thanks for doing this.
Jack, thank you. Okay, I want to start by teeing up a contrast between Figma in the early days, which was like a multi-year long build before you kind of got things going, and then the state of the world today where like AI startups are racing out of the gates and there's tons of competition and everything's frenetic.
13 years and it's a little different now, isn't it?
Yeah. So, you started in 2012.
Mhm. August 2012 was our official start. And then you got really kind of off to the races when like four or five years later.
Closed beta was launched December 2015. GA October 2016. Didn't start charging until summer 2017.
Same day as our CFO started. Fun fact.
CFO started the day you started charging.
Yeah. He was like a bizops guy then, but now a CFO.
Okay. So you had this 5-year period. And I guess when you look back on it, you could either feel like that was a little too long, we should have launched faster. But on the other hand, you built some hard stuff like a design product in the browser when people had never done it, and collaboration which was very difficult. So how do you make sense looking back now on that 5-year period?
Yeah, I think definitely there were ways we could have speedrun it. Hiring faster, noticing that we had product-market pull and that folks were literally begging us to go do things. When we got very long docs from people saying they were inspired by our user test, I probably should have known that people cared. But I was nervous. I took the roadmap feedback and thought it would take forever. In reality, I should have hired faster. We had the resources. I was just a little trepidatious. Beyond that, there was a lot to build. We had Evan, my co-founder, who loved cross-platform stuff and compilation. He made a way to do cross-platform targeting as a backup. At some point we ripped that out and moved faster. So there were things we could have done to speed it up, but not that much.
Yeah. I mean, there is this category of product like Airtables or email where you can't ship a quick v1. So maybe there's some element of that.
Yeah. And I think even after we launched in GA, we kept adding features and every time we did, retention would go up. There were folks in the beginning who liked the minimalism, but we added everything people needed for their workflows. You always have to find that balance between adding complexity and simplicity.
Did you have the sense in those early innings that anything you built worked? Were you having to make hard calls between A, B, and C, or was it just all improvements were good because of strong product pull?
Once we got out and had people using it, we separated the work streams into two: blockers and differentiators. Blockers were things literally blocking adoption. Differentiators were things like design systems and sharing components. We had to do both: evolve the state of the world and make it so more people could try the thing.
Yeah. When you think about that experience and then look at AI tools now, there's a bit going around about getting to a couple million ARR in no time. Founders feel self-conscious if they're not going from 1 to 10 million in a month. Investors perpetuate it. It's very different from how Figma got started. How do you make sense of seeing that around you?
Yeah. And again, don't recommend slow build either. But first of all, if we had today's tools like Figma Make or a prompt-to-app tool, we could build a lot faster. But my hot take is that there are a lot of really awesome companies right now that are not really AI companies. If you're looking at only AI stuff, you're missing a lot. For example, I invested in Ambrook, which helps farmers with taxes and compliance. Another company, Until Labs, is working on whole-body reversible cryogenics, but along the way they have a real problem: organs become available with a limited time window, and they can vitrify the organ and reworm it at destination, saving lives. Neither are AI companies, but they're great businesses. When you have strong trends like AI, there's a gold rush, but the people not working on that thing are missionaries at a higher rate. In the gold rush, there are a lot of missionaries but also a lot of people trying to go 0 to 10 next month.
Well, I think the other thing is that AI closes gaps and expands markets. So there's low-hanging fruit all over. If we're identifying companies growing fast, we're also identifying big markets people haven't tapped yet. That might be true for some companies, but some that go straight up go straight down. And then there will be amazing winners. But if the why now is AI, it's the same as everyone else's why now. If your strategy is a gold rush, you have to charge incredibly hard and be strategic. You have to know if you have that in you. Not everyone does. It's also a set of business decisions that look very different from the way you built Figma.
For sure. And it's also about whether it's long-term defensible. But if you get there first, you can go all sorts of interesting places. It's hard to puzzle through what will happen, leading to the climate where folks race towards opportunities and get backed at very large valuations before they even launch. That's not a judgment; it's a description.
Do you ever think about building a company in the opposite ethos of the way you built Figma? Does that sound fun or tiring to you?
The hard-charging part is what we are doing at Figma. The team is working incredibly hard right now, myself included. We see massive opportunity. Figma Make is one example. We have a differentiated approach where you can go from Figma Design, pull context into Figma Make, create a better designed application, and then go back to Figma Design to tweak. Eventually it'll be a round trip. We've talked about doing more workflow assistant type things in Figma Design where you can prompt there too. I'm so excited about making it so more people can create designs consistent with their brand language and design system. If you scribble something on a napkin, it won't be treated seriously as if it looks like your usual UI. Doesn't mean you don't need a designer. We're going to get to a world where good enough is not enough. Good enough is going to be mediocre. You're going to need to differentiate through design, craft, point of view, brand, storytelling, and marketing. The people that internalize that now will be winners. That's my point of view.
That's my point of view is that this is what's going to matter at the top of the stack. And if you don't internalize it now, you got an issue.
This was one of the topics I wanted to ask you about. As far as you can see, do you think the role of the human designer will flip from maker to editor, or become only focused on the most creative flourishes? Where does this go?
I think first of all, I've been in this place where maybe the roles are all merging together. I was saying that before AI was the topic, even 10 years ago. Now I feel like everyone is thinking PM, engineer, designer, researcher. The general pull of AI makes everyone feel they should be more generalist. But looking at how things are playing out, it's less roles merging and more responsibilities getting murky. Everyone has their specialization but also increased ability to have impact elsewhere. So a designer might commit code, a product manager might make a prototype instead of a PRD.
Yeah. I mean, you've always had full-stack product builders at startups, but often they work with a product manager and designer. Now we're in a place where you should be hiring a designer right away. Design is everything going forward. It's the top of the value stack. You're thinking about business logic, user problems, the system, brand, culture, affordances, style, structure. All these decisions are at the root of how you win or lose as a business.
Yeah. And there's good AI tooling around a lot of that. Getting user feedback used to require scheduling 30-minute calls; now you can do it other ways. A designer can get feedback scalably, make good enough versions to start playing with, and get something into production from a cockpit. I'm not saying in the next year or two a designer will scale the company as one person. Engineers are more needed than ever to architect systems properly. If you let agents run right now, you'll have a mess with security vulnerabilities and hacks.
Do you think this is a moment-in-time issue or a perpetual issue?
It's for sure the issue right now and will be for a while. The developer will still be in that outer loop even as models improve, as will researchers. Math is probably underappreciated as one of the most deterministic areas where RL should shine. Even if we have an ASI mathematician, mathematicians won't be out of a job. They'll be prompting and reasoning, like fishermen fishing for theorems.
Do you think we'll have even more designers and engineers, but much more productive?
Much more productive, but let's be clear about productivity. In engineering, it's building something well. In design, it might be further exploration of the option space. Right now you're constrained by timeline; you can only explore so much. If you can explore further and figure out how options work in the decision tree, you can go deeper and apply more craft. Design is inherently nondeterministic, unlike math, so AI's role in design will be very different.
Do you buy the argument that we'll have lots of small companies with a billion in revenue, or will competition rise and companies stay the same size but software gets way better?
There are already companies smaller than you'd expect with large revenue runs. The question is how much they scale before they break. I talk to founders who are proud of how much they've accomplished with few people, but then they're desperate to hire because they have problems at scale. I don't know if team sizes will decrease a lot. There are a lot of problems to deal with. AI makes some roles more efficient, not all. As AI improves, we get more efficiencies, but there's also more work to be done.
Totally. If engineers get more productive because of codegen, you'll start building a lot more product and testing it, so you'll probably need more designers. Competition is important. If you have more software, company A does the same with less, but company B does more with the same and wins.
I see it as for Figma, we just went through headcount planning. It didn't cross my mind to reduce headcount. We have so much we want to do. Let's increase headcount and hire great people to do amazing stuff for our users. That's the mindset more companies will find: there's more possibility, I can do more, I need more people. Some companies will try to get more efficient, but that might be a recipe for long-term ineffectiveness compared to companies growing with the technology.
On this competition point, I always read Figma as having somewhat low competition relative to being such a big market. You had Adobe and Sketch, but the market didn't catch on as fast as I thought. Is that right?
