Part IThe Story
The timestamp reads Sat, Dec 31, 2011 at 10:07 PM. It sits at the top of an online application to a program then still called 20 Under 20, and the applicant is a Brown University junior named Dylan Field. He had come across the form roughly half an hour before it closed. His parents, by his later account, assumed it was something of a joke, a teenage lark that would collapse under its own deadline. It did not collapse. It went in with minutes to spare. And the idea it proposed, the thing a nineteen-year-old offered Peter Thiel's foundation in exchange for $100,000 and the abandonment of his degree, had nothing to do with design.
It was about drones.
Field published the application in full years later, nearly unedited, "even in places that make me cringe," with his own commentary in italics. His stated reason was a theory of evidence. Founding stories, he wrote, "favor narrative simplicity over a deep dive into the messy, complex and often existential process of starting a company," and founders themselves are unreliable narrators: "As a species, entrepreneurs are optimists with very high pain tolerance. We forget the painful parts and skip over key details that no longer serve us." His conclusion was almost juridical. "The only 'True' way to recount a company's founding is through sharing artifacts from the past."
Fine. Here is the artifact. A kid at a laptop on New Year's Eve, pitching unmanned aircraft, about to drop out of the school he had called his "dream school," on the strength of one conviction he could articulate and a second he could not yet name. The first conviction was about a person: his teaching assistant, Evan Wallace, "the smartest, kindest, most humble person I had ever met." The second was about a gap, the one Field would describe on the Today show eight months later, between how much we can see and how little we can make. "I think everyone has this inner artist," he told the audience, "and we suppress that artist because our taste is so high but our skill level is a little lower. But I think the delta between that can be changed, fixed with technology."
Taste high, skill low. A delta that technology might close. It took four years to ship the first version of the product that tried, five to charge for it, thirteen to take it public, and roughly eleven more months for the market to decide the thing was worth about a seventh of what it had paid on the first day. Somewhere in that span the delta itself flipped. Machines learned to produce skill on demand. What they could not produce, Field came to argue, was the other half of the equation.
By the Numbers
Figma, from Fellowship to Float
20 of 500Thiel Fellows selected from student applicants in Field's 2012 class
4 yearsFrom founding in 2012 to Figma's first public release on September 27, 2016
$20BAdobe's abandoned offer for Figma, announced September 15, 2022
$1BReverse breakup fee Adobe paid when the deal died on December 18, 2023
$115.50Figma's closing share price on its first day of NYSE trading, July 31, 2025, against a $33 IPO price
$1.06BFigma revenue in 2025, alongside a net loss of $1.25 billion
$6.6B → $1.1BForbes estimate of Field's net worth, August 2025 versus June 2026
A Janitor, a Poet, and a Windows Commercial
Penngrove, California, is not much more than a crossroads in Sonoma County, the sort of place that people driving between Petaluma and Rohnert Park pass without registering. Field grew up there as an only child, the son of a respiratory therapist at Santa Rosa Memorial Hospital and a resource specialist teacher at Thomas Page Elementary School. His parents named him after Dylan Thomas. It is a detail that does a lot of quiet work: a boy christened for a Welsh poet who would spend his life building instruments for people who draw.
He was doing algebra at six. By middle school the curriculum had lost him so completely that, according to his father, speaking to the Press Democrat in June 2012, Dylan "mostly hung out with a janitor, who was kind of a math savant." There is a version of this anecdote that has already been polished into movie shorthand (the Good Will Hunting comparison writes itself, and at least one profile has written it), but the plainer reading is more interesting. The institution bored him. The interesting mind in the building was not on the faculty. He went and found it anyway.
The other half of the childhood is stranger and, in hindsight, more prophetic. Field was a child actor. He appeared in television spots for eToys.com and for Windows XP, and he did stunt work in a stage production of Peter Pan, where, the story goes, he once fell asleep on stage. He would later describe his childhood qualifications for commercial work with a product manager's cold eye: "My differentiators when I was five, five and a half I think, was that I could read and I could sit still and I was decently cute. And I hit puberty and those things were no longer differentiators. And then it was like, let's do some computer science." The theater stayed with him in another register. Years later, explaining why a design tool ought to feel like a room full of people rather than a file on a disk, he reached for improv: "The energy of play, of 'yes-and' in an improv sense — you really do want to enable that for people working in a collaborative space together."
High school was Technology High School, a STEM magnet on the campus of Sonoma State University. He built robots in FIRST Robotics and websites for friends. He interned at O'Reilly Media and, as a high schooler, turned up at Foo Camp in 2008, the invitation-only unconference where Tim O'Reilly's circle tends to gather. He worked with danah boyd, the researcher who had spent years studying how teenagers actually live online, and she wrote one of his letters of recommendation for college. He got interested in cryptography and gave a talk on hash functions, which had the unintended consequence, around 2009, of attracting emails from "these weird people talking about this bitcoin thing." He dismissed them as "crazy internet people with their internet money."
A pattern is already visible, though it would be a decade before anyone had a reason to look for it. The boy goes where the interesting people are, regardless of whether the institution has placed them there: the janitor, the researcher, the unconference. He asks. They tend to say yes.
— Dylan Field's father, to the Press Democrat, June 2012Pretty much everything we earned went to education.
That sentence was spoken in the weeks after Field had accepted $100,000 to leave Brown. It deserves to sit by itself for a moment. A respiratory therapist and an elementary school resource teacher had poured nearly everything into getting their only child to an Ivy League school, and the child had just been paid by a billionaire to walk away from it. "They totally did not want me to apply," Field told the Brown Daily Herald that September. He was careful, in 2012, to say he intended to return. Brown allowed leaves of up to five years. He wasn't feeling, he said, "like I was getting as much out of it as before."
He did eventually go back to school, in a manner of speaking. In February 2023, onstage at Stripe's South San Francisco office, the CEO of a company Adobe had tried to buy for $20 billion mentioned, almost in passing, that he was currently taking a math class.
The Bet Was on a Person
The money came from a man who had also been a lonely mathematical child.
Peter Thiel was born in Frankfurt in 1967, moved to Cleveland as an infant, then to South Africa and South West Africa, and changed elementary schools seven times. In Swakopmund he attended a German-language school with uniforms and a teacher who struck students' hands with a ruler; he would later say the experience gave him a lifelong distaste for regimentation. Classmates remembered a smart, withdrawn boy whom others avoided because they knew he would not stay long. He became a chess Life Master, topped a California-wide math competition in middle school, co-founded PayPal, wrote the first outside check into Facebook, and in September 2010 walked onto the stage at TechCrunch Disrupt to announce he would pay young people to leave college.
The origin of the fellowship is less philosophical than its later defenses suggest. Thiel had been reading the script for The Social Network on a plane the day before Disrupt, disliked how it portrayed him, and in something close to a fit of pique arrived at the idea of an anti-Rhodes Scholarship. Michael Gibson, recruited to Thiel Capital only one day earlier to help Thiel teach a course at Stanford, got an urgent call to flesh it out. Gibson later recounted that the team was still working out details right up to the moment Thiel took the stage. A week after that, Danielle Strachman was hired to run it.
Larry Summers, a former president of Harvard, would call it "the single most misdirected bit of philanthropy in this decade." The early years were, by most accounts including Thiel's biographer Max Chafkin's, a mess: favoritism, unqualified staff, zero structure, fellows who felt like failures, and some who fell into depression and drug abuse. Thiel himself, a decade later, would offer the most precise verdict: "It has been an extremely successful and effective program. It's not scalable."
Into this improvised machine walked Dylan Field. The second class drew 500 student applicants; 40 became finalists and 20 were chosen. Field's parents, he said, warmed to the idea only as they met the other finalists during the selection weekend. He was named a fellow in May 2012 and immediately made the phone call that actually mattered. "After we got the fellowship, I called Evan and said, hey, you know, we've got actually funding now. Hundred thousand over two years. Do you want to go do this thing for real? And he's a very reserved guy, and he was quite excited about that."
Evan Wallace was a year ahead of Field at Brown, a graphics specialist who worked as a teaching assistant in the computer science department and who had built a WebGL demo impressive enough to land in Wired. He was one of a small number of people anywhere who understood what that technology, which let a web page reach a computer's graphics processor, might do to the category of software that until then had been chained to the desktop. He was quiet. He did not like hardware. He particularly did not want to drift into defense work, which the drone idea threatened to do. He completed his degree that spring and followed Field to California.
What Field was actually buying with Thiel's money was not an idea. He was candid about this from the beginning:
— Dylan Field, at Stripe, February 13, 2023I felt like it was super obvious that I could learn more from Evan than any other context. And even if the company was, like, a complete failure, I thought it would be successful for me because I would learn from Evan.
The Index Ventures partner Danny Rimer later reduced the same calculation to its investor's form. Field, he said, "believed in Evan, and he worried that if Evan got a job after graduation, they'd miss their chance." So the fellowship functioned less as seed capital than as a kind of hostage-taking in reverse, a way to keep the most talented person Field knew from vanishing into Google.
There is an irony in the lineage that Field himself seems to enjoy without ever quite stating. The fellowship's patron would become a political theologian of considerable notoriety, a funder of defense startups named after Tolkien artifacts, a man who wrote that he no longer believed freedom and democracy were compatible. Field took his money and his network ("the biggest thing, in addition to money the Thiel Fellowship offered, was community"), then went on to say things in public that Thiel's circle would not. On crypto, the most libertarian of technologies, he said in 2023: "I didn't expect to be talking about how we need regulation for crypto, especially given that crypto was such a libertarian idea. But I think that actually the lack of regulation right now is blocking crypto." On standardized testing, he said he "probably diverge[s] a bit from Silicon Valley consensus," and that he was "more aligned with some of the equity arguments that universities make there than I think a lot of folks are." The co-founder he dropped out for refused defense work.
He went on to mentor later fellows and to sit on selection for new classes. When asked what the program had actually given him, he did not talk about the check. He talked about a permission structure. "So many people just can't get in their heads that you have permission to go do whatever you want to do, and as soon as you got to have that unlock, it's so powerful."
The Idea Maze Has Dead Ends, and He Walked Into Most of Them
The early history of Figma reads like a list of things that were not Figma.
There were drones, which died almost immediately because Evan "was not that interested in doing hardware. He's like, hardware sucks, run debug cycles are really long." There was a WebGL tool that let you annotate a photograph and turn it into a navigable 3D scene, which was "pretty neat for 2012 in all in the browser," but whose only obvious application seemed to be "a better Craigslist housing section or something that just sounded not interesting." There was a face-swap app, delightful at parties and in investor meetings ("oh, here's your face, it swapped"), which Snapchat would popularize years later and which was, Field concluded, a one-off that didn't connect to anything else. The Brown Daily Herald in September 2012 described the company only as "a technology startup that will allow users to creatively express themselves online," whose first ideas involved 3D content generation and later photo editing and object segmentation. When TechCrunch reported the June 2013 seed round of $3.8 million, Field declined to be interviewed; the only public description of the business was a single line on the fellowship's website: "Dylan is building a browser based photo editing tool."
Then the meme generator. It is the best story in the canon, because it is the one that most resembles failure as it actually feels.
In the fall of 2012, Field talked Wallace into building a meme generator. Field's thesis was that "memes are going to be huge on the internet and meme generator websites are going to have a lot of traffic." He was, he later noted, not wrong. They built it in three days. A week in, the two of them confronted each other across the table. "I'm kind of going, and fuck, I dropped out for this. And Evan's going, I think I'm going to quit if we keep working on this."
The episode taught a rule Field would repeat to rooms full of aspiring founders for years: you know within days whether you want to keep working on something, and that knowledge is data. "Go for an idea that you really care about because even if it doesn't work, you'll still learn from it." Three to four years into a business you hate, he warned, "you're just going to burn out and you're going to quit. Or maybe you'll have like a zombie company."
The photo editor died for a different reason, and the reason is more instructive. After building much of it, the two of them did what they admit they should have done first: they looked at the curves. Megapixels per iPhone, rising and concave up. Digital cameras sold per year, falling and concave down. "The best camera is the one you have in your pocket," Field recalled, citing the old photographer's adage, and a browser-based photo editor was being built for a device people were abandoning. They scrapped it.
What replaced it came from Field's other life. On his second stint at Flipboard, a six-month internship in Palo Alto as a technical product manager, taken by skipping the spring semester of his junior year, he had worked with the design team and grown frustrated with the tools. "During my Flipboard internship I was frustrated using Fireworks every day. Why couldn't design tools be more like Google Docs?" Around the time Figma raised its seed round, Adobe began sunsetting Fireworks. A product category had a hole in it shaped exactly like the thing two graphics-obsessed kids with a WebGL renderer could build.
And the market was about to expand under their feet. Field liked to cite IBM: in 2012, one designer for every 72 engineers; five years later, one for every eight. The cloud was cheapening development, so differentiation was migrating to design. "We didn't have the word product design," he said. "We didn't know why design was evolving and growing so much." They had stumbled into the right corner of the right decade.
Before committing, they called Aviary, a company that had already attempted something like it in the browser and pivoted away, and asked why. "They warned us, 'Don't do it. It's a terrible idea. Here are all the reasons why it's wrong.' We listened to them but thought we had a reason why it may be right."
The phrase he would later use for all of this was "why now," which he apologized for as startup cliché before using it anyway. His preferred flavor was technological. Drones and WebGL were the two things he saw changing in 2012. One required hardware and invited the Pentagon. The other required Evan.
Four Years Is a Long Time to Say 'Next Year'
At Index Ventures, it became a running joke.
Danny Rimer had first met Field in 2011, when Field was an eighteen-year-old intern at the Index portfolio company Flipboard, doing customer research. What struck Rimer was not the analysis but the energy: "He was genuinely curious, bubbling with ideas, and clearly lit up by the potential of building something people might love." A year later, LinkedIn's CEO Jeff Weiner mentioned to Rimer, "We have this incredible intern I'm trying to hold onto, but the kid wants to start his own company. It sounds interesting. You should talk to him." (Field's LinkedIn manager in the summer of 2010, Pete Skomoroch, would also write him checks in the seed and the Series A. The interns-into-investors pipeline was, in Field's case, almost the whole cap table.) Field and Wallace came to the Index office in San Francisco and pitched. They didn't yet know exactly what the product would be. Index wrote the check anyway, leading the 2013 seed.
"In the world of move-fast-break-things," Rimer recalled, "here were two folks who were saying, 'We're not going to have anything for two years, so we hope you're comfortable with that.'" It was more than two. Index partners would periodically ask Field when the product was launching. "Next year!" he'd answer. "We want to get it just right."
The man who nearly didn't fund them was John Lilly, the former CEO of Mozilla, by then at Greylock, with deep roots in design. Field approached him for the seed. Lilly's answer has become one of the more quoted sentences in Figma lore, and Field himself tells it as a wake-up call: "You know, I don't think you guys know what you're doing, but I'm very intrigued, so let's keep in touch." Lilly, unusually for a venture capitalist, meant the second half. Every few months they would get coffee. Lilly worked on the strategy with him, and it was through those conversations that Field abandoned his original plan, which was to build a design community first, a Dribbble-like place where designers could riff on each other's work, and figure out monetization later. "Talking with John got me to the point where I realized we need to start with a business tool. We'll build the community later." Eventually Lilly said that if Field ever thought about raising again, let him know. Field raised again a few weeks later, mostly, he admitted, "because I just wanted to work with him." Greylock led the $14 million Series A in December 2015. "At the time, it wasn't like there were metrics to look at," Field said. Lilly had conviction "in the space, a conviction in the attack, and he had conviction in me and Evan."
Stealth is lonely in particular ways. Recruiting was the worst of them. Field learned, sitting across from candidates, that the number of minutes before he got his laptop open and showed them the product mattered enormously to the outcome. He also had to recruit users before there was much to use. He wrote a script to scrape Twitter's follow graph, ran the result through Gephi, a network visualization tool he'd seen colleagues use at LinkedIn, and identified the most central designers. Then he looked at their work and wrote to the ones he admired. Most said yes to coffee, "especially if I was buying."
The growth-hack version of this story elides what it actually was, which was fandom. One of the designers was Tim Van Damme, whose icons Field had been using as test cases while building vector networks, Figma's novel approach to path editing. "I think the first time I met Tim was at Dropbox and think I had this total fanboy moment. I'm like, 'I've been tracing your icons.' He's like, 'Hi.'" Another, Payam, sat through a user research session so long and so painful, given how slowly text rendered in the early editor, that Field brought a bottle of wine to get them through. Payam then wrote the team a very long document listing everything he wanted.
The technical problem was not small. Running a professional vector editor at speed inside a browser in 2015 was, by most expert opinion, somewhere between ambitious and absurd. In an interview with VentureBeat that year, Field explained that the "viewer" was written in Skew, a new language at the time, while the rest of the application was in C++; WebGL, and later WebAssembly, carried the rendering. The free, invite-only preview opened on December 3, 2015. The public release came on September 27, 2016. Charging began in the summer of 2017, on the first day of the person who would become Figma's CFO.
Field has never romanticized the delay. "Way too long, don't do that," he said onstage at Config. On LinkedIn in 2025: "It took us 3 years to launch Figma (Advice: don't do that)." The confession recurs so often it has become a kind of signature. What he does not say, and what the record shows, is that very few things that later turned out to matter about Figma could have been built in a hurry.
The Intervention
Here is the part the narrative-simplicity version leaves out.
Sometime in the long pre-launch stretch, with roughly twelve people on the team, Field's father was dying of cancer. Field was twenty-two, twenty-three. He had never managed anyone; his only prior jobs were internships. "I was like, I think you're supposed to do one on ones. So I [did] that. Didn't really know what to talk about during them." He was perpetually optimistic about the ship date, which meant he was perpetually wrong about it, which meant that the people who had joined a stealth startup on faith were spending their twenties watching a deadline recede. He was, by his own account, "pushing super-hard on the team, but also giving them not a lot of empowerment."
People began leaving. Field later described facing a potential exodus. Then the senior members of the team took him aside, careful not to include the newer hires. As he told Sequoia's Kevin Lincoln in 2021, "It was like, 'You need to get some help.' Afterward I took a few days away from the office. It was just hard."
— Dylan Field, 2021I was just not a very good manager when I started Figma. I was an intern before that, so I had a lot to learn. I was always very optimistic; I thought that shipping was right around the corner, so I wasn't setting expectations correctly.
What he did next is the most consequential personnel decision in Figma's history and among the least glamorous. He hired a manager. Sho Kuwamoto, a veteran builder of creative software who had spent years at Adobe, the very company Figma was designed to displace, joined as director of engineering. In his first week, Field recalled, Sho gave a presentation: "Here's what we got to do, here's the gap. Everyone agrees on it. Let's go." The phrase Field used for Sho's arrival was "very catalytic," and the reasoning was about learning rather than delegation. "It meant that I could go learn from him, and I was still learning from him. Today." Sho later became Figma's VP of product. Elad Gil, onstage at Stripe, joked that Bill Gates used to churn through COOs every eighteen to thirty-six months, learning from each in turn. "In this case," Field said, "Sho is still here like seven years later."
Many of the people in that room eventually left. Field described the episode as "a bit of a cultural reset moment, not just for me, but for the greater team as we went up to launch." And he drew from it the one piece of advice he chooses, whenever he is given a final word in front of founders: "Don't feel shame over them. Talk to other people about them, and you'll be surprised at the amount of help they can give you."
The arc tends to get written as a redemption story, a founder who became the leader his company needed. That is roughly true. But the specific mechanism matters. The young man who as a bored middle-schooler went looking for the most interesting mind in the building, and who dropped out of college to learn from his TA, solved his management crisis the same way: he found someone who knew more than he did and put himself in the student's chair. "One thing that's kind of cool," he said, "is you can recruit your own mentors to the company."
Multiplayer, Which Nobody Asked For
When the Figma team interviewed designers about real-time collaboration, the kind where several people edit the same canvas simultaneously and see each other's cursors move, roughly one or two out of hundreds said they wanted it. "Most designers, we asked, hey, do you want to do this activity with the browser in real time? And they're like, no, I have no interest in that at all."
They built it anyway. It took about nine months.
The conviction came from three places. Field had watched design work at Flipboard, where files were the unit of collaboration and therefore the unit of pain: designers lost hours to version control and still couldn't work together. He and Wallace had both spent time "playing around with Google Wave, which was a thing at some point for a brief moment," Google's ambitious, failed experiment in live collaborative documents. And the alternative, once they had a browser-based editor, was simply intolerable. "If it's last edit wins, and we're both in the tab at the same time, and you make an edit, and my thing reloads, and I'm making edit your thing reloads, it just sucks."
"When you kind of have those insights, even if no one's asking for them," Field said, "you have to prioritize them sometimes."
The design community did not initially greet Figma with gratitude. There was, in the period after launch, open hostility: a browser tool for serious design seemed to many working designers a contradiction, and an online editor that strangers could peer into seemed to some a violation. Field later called the posture he adopted toward the "haters" a kind of patient engagement; Sequoia titled its 2021 profile "How Figma's Dylan Field Won Over the Haters." The mechanism was listening, conspicuously and constantly. Field reads, he has said, everything people post about Figma that he can find, across social media and support channels, and drops much of it into an internal Slack channel. During the pandemic, according to Forbes, he spent much of his time reading support tickets, answering users on Twitter, and visiting customers in person in Ukraine and Nigeria. In 2018 he called Rimer asking for recommendations in Lagos, where he was going for a Figma meetup. Rimer had never been. "I want to go wherever the Figma community is," Field told him, repeatedly. More than half of Figma's revenue eventually came from outside the U.S.
The first team to use Figma together was at Krypton, the company that became Coda, led by Shishir Mehrotra. There is a story about Field driving home from installing the software there, learning that something had broken, and turning the car around to fix it himself. (He has insisted, onstage, that he certainly was not reading email while driving.) Shishir, who became a mentor to Field, did not know until years later that his team had been the first. "He goes, 'Wait a second, I am?'"
The most telling admission from this era concerns product-market fit, and specifically Field's failure to notice it. A potential customer once wrote the team a twelve-page document listing everything they wanted built, at a time when "the editor was so slow that if you typed text, it took forever." In retrospect, Field says, that was the moment. He didn't believe it until Microsoft told the company it needed to start charging, because Microsoft wanted to spread Figma internally and could not do so under the free terms. "In reality we'd had it for like a year and I wish that I had recognized that sooner." The lesson he extracted is that "it's really easy to just deny that you have product market fit," and that product-market pull, the feeling of customers "trying to get the product out of us," is the signal worth trusting.
There is one more artifact from these years worth mentioning, because it shows the tool working in a way no deck could have predicted. When Notion's founders, Ivan Zhao and Simon Last, laid off their two employees and decamped to Kyoto to rebuild their product from scratch, Zhao spent eighteen-hour days designing the new Notion in Figma. The tool for people whose taste exceeded their skill had become the workbench of people with plenty of both.
By 2020, Figma had $75 million in revenue. Forbes reported that Joe Biden's presidential campaign managed its visual assets in Figma, and that "when toilet paper ran out across the U.S. in 2020, Kimberly-Clark drafted reorder forms using Figma's tools." Valuations ran from $115 million in 2018 to $2 billion in April 2020 to $10 billion in June 2021. That year, Evan Wallace stepped back from the company to work on other projects. Forbes called both founders "near-billionaires."
Twenty Billion Dollars and a Breakup Fee
On September 15, 2022, Adobe announced an agreement to acquire Figma for about $20 billion in cash and stock, the largest acquisition in Adobe's history. Field would stay on as CEO. Adobe shares fell 17 percent. Pete Skomoroch, Field's old LinkedIn manager, tweeted his congratulations: "Dylan was my intern in 2010, we kept in touch as friends, and years later I was able to invest in Figma's seed and Series A rounds."
Designers were not pleased. Many feared that the tool they'd adopted precisely because it wasn't Adobe would be folded into Creative Cloud, repriced, and smothered. The Guardian's John Naughton compared the deal to Facebook's acquisition of WhatsApp, a purchase meant to neutralize a rival before it became a threat; Adobe XD had been losing share to Figma since at least 2021. On November 2, 2022, the U.S. Department of Justice opened an investigation. In February 2023 the European Commission began its own review. The U.K.'s Competition and Markets Authority dug in.
On December 18, 2023, the two companies announced they were mutually abandoning the deal; Adobe cited "no clear path to receive necessary regulatory approvals from the European Commission and the UK Competition and Markets Authority." Adobe paid Figma the $1 billion reverse breakup fee. CB Insights estimated that, on earnings metrics alone, Figma was now worth somewhere between $8.3 and $9 billion.
What a founder does on the day a twenty-billion-dollar outcome evaporates is one of those things that rarely gets recorded. What Field did shortly afterward did get recorded, and it is an unusual piece of organizational design. Figma offered employees a program it called Detach: anyone who wanted to leave could do so with three months' pay, no questions asked. The logic was that some people had stayed because they expected to become Adobe employees, and that a company moving forward should be composed of people who chose to. Roughly 4 percent took it.
It is worth noticing how closely this rhymes with the earlier intervention. In both cases a gap had opened between what the team had signed up for and what the team was going to get. In the first case, the team forced the reckoning on Field. In the second, Field forced it on the team, gently, with a check.
The Weather App
The age of generative AI arrived at Figma the way it arrived everywhere: suddenly, and with a strong sense that whatever you were already doing was now either obsolete or about to become enormously valuable, and you could not quite tell which.
Field's public answers were consistently on the side of design. In 2023, asked by a16z's David George whether AI would take designers' jobs, he walked through the history of platform shifts: "You have the printing press and then you have to figure out what you put on a page." Mobile, which ought to have meant fewer pixels and fewer designers, produced "the biggest explosion of designers." He noted, with a certain dryness, that OpenAI was at that moment "on a hiring acquisition spree trying to get product people and designers." At Stripe that February, he'd predicted that by 2030 or 2033 an AI would co-author a paper in a pure mathematics journal, and that all software would eventually be "browser based and collaborative."
Then, at Config in June 2024, Figma shipped an AI feature called Make Designs. On July 1, the designer Andy Allen posted two screenshots side by side. On the left, a weather app Figma's new feature had generated. On the right, Apple's iOS Weather app. They were nearly identical. The post accumulated roughly 1.5 million views.
Field responded within a day, on July 2, in a thread that did not do what most corporate statements do:
— Dylan Field, July 2, 2024Ultimately it is my fault for not insisting on a better QA process for this work and pushing our team hard to hit a deadline for Config.
The feature was pulled. When AI generation came back, it came back with different framing and eventually a different name, and with a self-deprecation that seemed calibrated to defuse exactly the fear the weather app had triggered. Onstage at Config, Field had already warned that the generator was "going to give you the most obvious thing in the most obvious form possible." People online, he noted with some amusement, then complained that it gave them "some obvious thing." "I agree," he said.
The weather-app episode clarified Field's thesis rather than undermining it. If a model, asked for an app, will reproduce the most average app in its training data (one of the most-copied pieces of interface design in the world, in this case), then the model is extraordinarily good at precisely the part of design that was never the point. "Prompt to app doesn't give you an app that is well designed yet," he said. AI, he began to argue, is useful as a cliché detector: it tells you the obvious answer so you can go somewhere else.
The argument came to its sharpest form in 2025. "If you really believe that development gets easier and it's more simple to create software, it's faster to create software," he said at Y Combinator, "then like what is your differentiator? It's design, it's craft, it's attention to detail, it's point of view." To Jack Altman, he said it more bluntly: "We're getting into a world where we're kind of there, where good enough is not enough. Good enough is gonna be mediocre." To Latent Space: "Brand, it's point of view, it's taste, it's craft, it's design."
Notice the inversion. In 2012, on the Today show, the problem was that our taste was high and our skill was low, and technology would fix the skill. In 2025, technology has more or less fixed the skill, or at least the first draft of it, and the scarce thing has become taste. The delta did close. It closed from the side Field originally aimed at. And the residue it left, the thing machines could not supply, was the thing Field had assumed everyone already had.
He defined taste carefully, as more than preference: a conscious ability to judge quality and explain why. He defined intuition as "a hypothesis generator," whose outputs must then be debated and tested. And he defined design, most often, as "art applied to problem solving," crediting the line to someone else. "If you lose the art and you're just solving the problem, it's totally utilitarian and it lacks soul."
At Config 2025, on May 7, Figma shipped Figma Sites, Figma Make (built on Anthropic's Claude 3.7), Figma Buzz, and Figma Draw. "This year, we shipped more than we have in the previous 10 years combined," Field wrote. Not everyone welcomed it. Among the two hundred comments under his announcement were designers asking him to stop changing the interface ("We have memories of how to use it") and one who accused the company of "pivoting hard toward becoming the Canva of design tools." The community that had once feared Adobe would smother Figma now feared Figma becoming something other than what they'd loved. The same week, the founder of 37signals, Jason Fried, wrote: "Really impressive work. Beautifully presented as well. No afterthoughts here. Bravo."
In October 2025 Figma bought Weavy, a one-year-old Israeli AI startup, for more than $200 million and opened an R&D center in Tel Aviv. In February 2026 it announced integrations with Anthropic's Claude Code and OpenAI's Codex. And in April 2026, Anthropic's chief product officer, Mike Krieger, co-founder of Instagram, resigned from Figma's board after reports that Anthropic's next model would include design tools that competed directly with Figma. Figma's stock rose about 5 percent on the news, as though the market preferred its rivals off the board.
Number Goes Up, Number Goes Down
Before Figma there was a JPEG of a pixelated face.
In January 2018, Field bought his first CryptoPunk, one of ten thousand algorithmically generated 24-by-24-pixel portraits on the Ethereum blockchain. His thought at the time, he later said: "this is probably the stupidest thing I've ever done." His wife, Elena Nadolinski, had become deeply interested in Ethereum and would found Iron Fish, a privacy-focused crypto project. Field's own theory of the market came from Neopets, the virtual-pets website of his childhood, which "had an entire virtual economy." When Ethereum collectibles took off, "it literally felt the exact same."
In February 2021 he sold one Punk for $1.5 million. Weeks later he sold another for $7.5 million, at the time the highest price ever paid for a CryptoPunk, and went on the Clubhouse show The Good Time Show to describe it as "a digital Mona Lisa." As of March 2021 he held eleven more, plus Autoglyphs and Beeple works; Forbes put his collecting profits at $9.5 million. By 2023 his tone had cooled. "You'd be cautious of any market where as money comes into the market, prices go up, and as it goes out, prices go down." His advice: "Buy things that you're going to want to keep forever." He had, he said, sold only two in his life.
On July 31, 2025, Figma listed on the New York Stock Exchange under the ticker FIG. The IPO priced at $33. Shares opened at $85 and closed at $115.50, a jump of roughly 250 percent, the largest first-day pop for a billion-dollar tech IPO in years, and Fortune put the fully diluted value near $68 billion, more than triple what Adobe had offered. Demand was such that many investors got only a handful of shares; trading was briefly halted. Field signed the guestbook on the exchange floor. His 9 percent stake made him, by Forbes's count, worth about $6.6 billion. "Dylan is, by far, the most humble billionaire I've ever met," said Joshua Browder, a fellow Thiel Fellow. Michael Gibson, who had helped run the program, offered a more pointed comparison: "He's kind of like the anti-Steve Jobs. When it comes to Jobs' legend as this hard-charging a--hole, Dylan is the opposite."
Then the number went down. After Figma's first earnings report in September 2025, the stock fell more than 50 percent from its peak. By June 2026 it was below $20, the company's market value under $10 billion, roughly where CB Insights had pegged it on the day the Adobe deal died. Forbes's estimate of Field's net worth fell to $1.1 billion. Figma's 2025 revenue was $1.06 billion; its fourth quarter alone brought in $304 million, up 40 percent year over year. Its net loss for the year was $1.25 billion. It had 13 million users and 1,886 employees. The fundamentals and the price had come apart, as they do in any market where money coming in makes prices go up.
Field's statement, which he says he gave the team before the listing, on the day, and after, was the one he had arrived at with the Punks:
— Dylan Field, Access podcast, September 2025Number goes up, number goes down. And we don't control that number, we control the inputs. And we have to educate the market.
Biographies have started to appear. One on Amazon, Who is Dylan Field?: His Journey from Math Whiz to Figma, compresses the arc into the shape its subtitle promises, the prodigy who became the founder. It is the narratively simple version, the one Field himself distrusts. His own preferred version is messier: a kid who could sit still, an intern who couldn't run a one-on-one, a CEO who blamed himself publicly for a weather app, a billionaire who lost five-sixths of it on paper in under a year and told his team it didn't matter.
Where Great Ideas Come to You
At Config in June 2024, inside a set built to resemble a podcast studio on the stage of San Francisco's Moscone Center, Lenny Rachitsky asked his last question. He had found something on YouTube. The lights dimmed. On the big screen, a five-year-old boy, small and solemn, in a world of primary-colored boxes, sat entirely still. A voiceover asked where you would find a world of ideas for your child. The boy, who could read and could sit still, had been cast for those two qualities.
Only at eToys, the announcer said. Where great ideas come to you.
A thousand designers laughed. Somewhere in the hall sat Field's mother, the resource teacher from Thomas Page Elementary, who had helped pour nearly everything the family earned into his education, and who had not wanted him to apply. "That's my mom," Field said, pointing. "My mom's in the audience and she says yes."
He asked for one more comment before they cut the mics.
"That commercial made that company go bankrupt."
Part IIThe Playbook
What follows is not a list of virtues. It is an attempt to extract the operating logic behind a specific career: one that began with a deadline met by thirty minutes and a pitch about drones, ran through four years of stealth, a management intervention, a regulatory collapse, an AI embarrassment, and a spectacular IPO followed by a spectacular decline. Each principle is grounded in the record above. Several of them run against the grain of standard founder advice, and some run against Field's own later advice, which is part of the point.
The principles
- 1.Bet on a person before you bet on an idea.
- 2.Let boredom tell you where not to go.
- 3.Read the curves after you've built, if not before.
- 4.Treat a rejection as the first meeting of a relationship.
- 5.Recruit your mentors onto the payroll.
- 6.Build the feature nobody asked for, if the alternative is unbearable.
- 7.Learn to recognize pull, because you will deny it.
- 8.Go to the community; don't wait for it to come to you.
- 9.Pay people to leave when the deal changes.
- 10.Own the failure in the first twenty-four hours.
- 11.When skill becomes abundant, sell taste.
- 12.Control the inputs; ignore the number.
Principle 1
Bet on a person before you bet on an idea
Field's Thiel Fellowship application proposed drones. What he actually wanted was to keep Evan Wallace from taking a job. He has been explicit that the bet was on learning from Wallace, so explicit that he framed even total failure as a win: "even if the company was, like, a complete failure, I thought it would be successful for me because I would learn from Evan." Danny Rimer at Index made the same calculation from the other side of the table, writing a check to two founders who "still weren't exactly sure what the product would be."
The practical implication is that the idea is a variable and the team is a constant. Field and Wallace burned through drones, 3D scenes, face swap, photo editing, and a meme generator. The combination that survived all of those deaths was the pair of them plus WebGL. Ideas can be iterated. Wallace was not going to be available again once he graduated.
This principle has a cost Field seldom dwells on: the company's center of gravity was a co-founder who eventually left, in 2021. The bet paid out across nine years. It was never meant to pay forever.
Tactic: Before committing to a venture, write down the one person you would most regret not working with in five years, and design the opportunity around getting them into the room.
Principle 2
Let boredom tell you where not to go
Field's middle school bored him, so he found the janitor. Brown stopped giving him "as much out of it as before," so he left. The meme generator, which was a perfectly reasonable business thesis, bored both founders inside a week: "fuck, I dropped out for this," and "I think I'm going to quit if we keep working on this." They killed it.
The standard founder advice is to follow passion. Field's version is more operational and more useful: treat your own boredom as a fast, cheap signal that arrives within days. "I think we knew like days into working on any project, like whether we wanted to work on it or not." A company takes years, he points out; three or four years into "an idea that you hate, you're just going to burn out and you're going to quit." The underrated corollary is that boredom also identifies where interesting people are not. He kept leaving rooms that had gone dull and finding the room where the most curious person was.
Tactic: For any new project, schedule a candid check-in with your co-founder at day three and day ten, and ask only one question: do we want to keep working on this?
Principle 3
Read the curves after you've built, if not before
The photo editor died because of two lines on a chart: megapixels per phone rising and concave up; standalone cameras sold falling and concave down. Field admits they "probably should have started there." But the sequencing is instructive. They built first, then looked at trend data, then scrapped the product, then went to talk to designers and found the Fireworks-shaped hole and the IBM ratio moving from 1:72 to 1:8.
Field's preferred lens is the "why now," and his preferred version is technological. The trick, in his telling, is that technology shifts generate options, and you then need market curves to tell you which option points downhill. Building gave them conviction about what WebGL could do. Curves told them where to point it.
Tactic: Pick two quantitative trends that your product depends on, chart them over five years, and kill the project if either one is bending the wrong way.
Principle 4
Treat a rejection as the first meeting of a relationship
John Lilly's no became the most famous line in Figma's history: "I don't think you guys know what you're doing, but I'm very intrigued." Field took the second clause literally. He kept going back for coffee every few months, let Lilly reshape the strategy from "community first" to "business tool first," and raised the Series A essentially so that he could work with Lilly. Aviary's "Don't do it. It's a terrible idea" got the same treatment; Field listened to every reason and kept the one that let them disagree.
The interns-turned-investors pattern belongs here too. Pete Skomoroch, his 2010 manager at LinkedIn, invested in the seed and the A. Rimer met him as an eighteen-year-old Flipboard intern. Jeff Weiner, LinkedIn's CEO, made the introduction. Field's cap table was, largely, a record of people who once evaluated him in a lesser role.
F
From Fellowship to Float
Each round arrived through a relationship, not a cold process.
2010
LinkedIn internship under Pete Skomoroch, later a seed and Series A investor.
2011
Meets Danny Rimer of Index as an 18-year-old Flipboard intern.
2012
Thiel Fellowship awarded in May; Field leaves Brown; Figma founded with Evan Wallace.
2013
$3.8M seed led by Index Ventures. John Lilly passes, but keeps in touch.
2015
Lilly and Greylock lead a $14M Series A in December; invite-only preview opens December 3.
2016
Public release on September 27.
2017
Figma begins charging, prompted in part by Microsoft.
2020
$50M Series D led by Andreessen Horowitz at a $2B valuation.
2021
$200M Series E at $10B; Evan Wallace steps back.
2022
Adobe agrees to buy Figma for about $20B on September 15.
2023
Deal abandoned on December 18; Adobe pays a $1B breakup fee.
2025
IPO on July 31; closes first day at $115.50.
2026
Shares below $20 by June; market value under $10B.
Tactic: After every investor or partner rejection, ask what they would need to see to change their mind, then send them that evidence within ninety days.
Principle 5
Recruit your mentors onto the payroll
The intervention ("You need to get some help") could have ended Field's tenure. Instead it produced the hire of Sho Kuwamoto, an Adobe veteran, whose first-week presentation finally gave the team a shared gap to close. Field's framing is the unusual part: he did not hire Sho to take things off his plate; he hired Sho so he could learn. "I'm still learning from him today." And: "One thing that's kind of cool is you can recruit your own mentors to the company."
Most first-time CEOs experience senior hires as threats to their authority or as rescue. Field experienced the hire as tuition. The same instinct that sent a bored middle-schooler to the janitor and a college junior to his TA turned an organizational crisis into a curriculum. He extends the idea in every direction: mentors, he said at Config, can come from the community, from investors, from coaches, and from interns. "Interns at Figma have been mentors to me, in many ways."
Tactic: When hiring your next senior leader, list three things you personally want to learn from them, and set a standing monthly session where they teach you.
Principle 6
Build the feature nobody asked for, if the alternative is unbearable
One or two designers out of hundreds wanted real-time multiplayer editing. Figma spent nine months on it. The justification was not market research; it was the felt experience of the alternative. In a browser, without multiplayer, "you make an edit, and my thing reloads, and I'm making edit your thing reloads, it just sucks."
Field's general framework for feature requests is ordinary (plot effort against impact, do the cheap high-impact things, keep a long-term roadmap, manage tech debt). The exception he carves out is what made Figma. "When you kind of have those insights, even if no one's asking for them, you have to prioritize them sometimes." He later formalized it: separate the work into blockers, which prevent adoption, and differentiators, which move the industry. Run both.
◆
Blockers vs. Differentiators
How Figma sorted early product work, per Field's later accounts.
| Item | Source of demand | Category |
|---|---|---|
| Fast text rendering | Early testers (one needed a bottle of wine) | Blocker |
| Pages | Widespread user demand; Field remains skeptical | Blocker |
| Real-time multiplayer | One or two of hundreds asked | Differentiator |
| Vector networks | Founder conviction | Differentiator |
| Charging for teams | Microsoft insisted | Pull signal |
The asymmetry with Pages is worth noting. Field shipped a feature he still doubts ("I'm not sure we'd come to the same implementation of Pages that we are at today") because "the world told me" it mattered. Conviction overrides demand for differentiators. Demand overrides conviction for blockers.
Tactic: Label every roadmap item as a blocker or a differentiator, and protect at least one differentiator that no customer has requested.
Principle 7
Learn to recognize pull, because you will deny it
Figma had product-market fit for about a year before Field admitted it. The twelve-page wish list from a prospective customer, arriving while text typed at a crawl, was the signal; Microsoft demanding to pay was merely the one he could no longer ignore. "It's really easy to just deny that you have product market fit." He would have hired faster, he says, and spent capital sooner.
The denial has an understandable source. Founders who survive a long stealth period are trained to discount enthusiasm, and Field's optimism had already burned his team with missed ship dates. But the asymmetry is costly: underreacting to pull leaves growth on the table at the moment it is cheapest to capture. "As soon as you have that notion or some sense there's product market pull, you kind of do have to hit the gas too, which is scary."
Tactic: Define, in writing and in advance, the specific customer behaviors (unsolicited feature documents, demands to pay, internal spreading) that will trigger you to accelerate hiring.
Principle 8
Go to the community; don't wait for it to come to you
Field mapped Twitter's design graph with Gephi and then bought coffee for the people at its center. He brought wine to user tests. He reads, by his own account, everything about Figma he can find. He flew to Lagos for a meetup in 2018 and spent the pandemic in support tickets and in customers' offices in Ukraine and Nigeria. "I want to go wherever the Figma community is."
The growth-hack reading misses the posture. Field's outreach was fandom ("I've been tracing your icons"), and designers, he notes, "are really good at giving feedback." The early users became, in his words, "friends, Figma customers, and evangelists." The community he once wanted to build first, Dribbble-style, came later and organically, because he had spent years being a member before he was a vendor. His advice to anyone without geographic access: "Don't ignore the power of a cold email," especially one with a clear question.
Tactic: Identify the twenty most respected practitioners in your users' field, study their work closely, and ask each for thirty minutes of feedback with a specific question you genuinely want answered.
Principle 9
Pay people to leave when the deal changes
When the Adobe acquisition died, Figma offered Detach: three months' pay to anyone who wanted out, no questions asked. About 4 percent took it. The program did two things at once. It released the people who had mentally become Adobe employees, and it converted everyone who stayed from passive participants into people who had actively chosen the next chapter.
The precedent was Field's own intervention years earlier, where an unaddressed gap between expectation and reality had nearly broken the company. Detach is that lesson institutionalized: when the implicit contract changes, make the new contract explicit and make exit cheap. Founders often fear that offering an exit will cause a stampede. Field's data point suggests the opposite.
Tactic: After any event that materially changes what employees signed up for, offer a time-limited, generous, no-questions exit and announce the take-up rate.
Principle 10
Own the failure in the first twenty-four hours
The weather-app screenshots posted on July 1, 2024. Field's thread posted on July 2. It named a cause (inadequate QA), assigned responsibility to himself ("it is my fault"), and identified the pressure that produced the mistake: "pushing our team hard to hit a deadline for Config." The feature came down. When AI generation returned, the framing had been recalibrated toward modesty.
Speed mattered, but specificity mattered more. Generic apologies protect the company; Field's apology protected the team by locating the error in his own deadline pressure. It echoes the 2021 confession about being "not a very good manager" and the repeated "don't do that" about the four-year launch. Field has made public self-criticism a recurring instrument, and it has bought him credibility with a community that is primed to distrust tools that threaten its craft.
!
Three Public Admissions
Field's recurring pattern of naming his own mistakes.
| Admission | Year | What it protected |
|---|---|---|
| "I was just not a very good manager when I started Figma." | 2021 | Team trust |
| "It is my fault for not insisting on a better QA process." | 2024 | Community trust |
| "It took us 3 years to launch Figma (Advice: don't do that)." | 2025 | Founder credibility |
Tactic: When something you shipped fails publicly, respond within a day with the specific root cause, your personal share of responsibility, and the concrete change you're making.
Principle 11
When skill becomes abundant, sell taste
In 2012 Field diagnosed the problem as taste outrunning skill. By 2025, AI was supplying skill in bulk and the scarce resource had flipped. "If you really believe that development gets easier... then like what is your differentiator? It's design, it's craft, it's attention to detail, it's point of view." Good enough "is gonna be mediocre."
This is both an analysis and a strategic bet. Figma's AI products are positioned as first drafts and cliché detectors, not replacements: they show you the obvious answer so a human can choose a non-obvious one. Field's definition of intuition as "a hypothesis generator" fits the same frame: the machine generates, the human selects and defends. Even his view of AI-driven productivity points this way: when engineers ship more, companies need more designers, not fewer, to decide what the extra output should be. The risk, which the April 2026 Krieger resignation made visible, is that the model makers decide taste is also something they can sell.
Tactic: Audit your product for the three decisions an AI would make most predictably, and make each of them deliberately differently, with a written reason.
Principle 12
Control the inputs; ignore the number
Field told his team on IPO day, before it, and after: "number goes up, number goes down. And we don't control that number, we control the inputs." Within a year the stock had fallen from $115.50 to below $20 while revenue grew 40 percent year over year in the fourth quarter. His personal paper fortune dropped from $6.6 billion to $1.1 billion.
The principle predates the IPO. His CryptoPunk education taught him to distrust markets where "as money comes into the market, prices go up, and as it goes out, prices go down," and to buy only "things that you're going to want to keep forever." Applied to a public company, the logic is the same: build something you'd want to keep forever, explain it patiently ("the market doesn't come out of the gate understanding Figma"), and measure yourself by shipped products, customers, and revenue. That is easy to say at the peak. The test is saying it at the bottom, which is where Field now is.
Tactic: Pick three operating metrics you fully control, report them internally every week alongside the share price, and never discuss the price without them.
Free playbook
Get The Business Model Playbook
58 business models, one visual page each: how the money flows, the metrics that matter, and who runs it. Free when you join the Faster Than Normal email.
Free. No spam. Unsubscribe anytime.
Part IIIQuotes and Maxims
In their words
— Dylan Field, the Today show, August 2012I think everyone has this inner artist, and we suppress that artist because our taste is so high but our skill level is a little lower. But I think the delta between that can be changed, fixed with technology.
— Dylan Field, on the meme generator, February 2023I'm kind of going, and fuck, I dropped out for this. And Evan's going, I think I'm going to quit if we keep working on this.
— Dylan Field, Figma's Story, Part 1As a species, entrepreneurs are optimists with very high pain tolerance. We forget the painful parts and skip over key details that no longer serve us.
— Dylan Field, Y Combinator, 2025If you really believe that development gets easier and it's more simple to create software, it's faster to create software. Then like what is your differentiator? It's design, it's craft, it's attention to detail, it's point of view.
— Dylan Field, Config 2024Keep the simple things simple. Make the complex things possible.
Maxims
- Bet on the person. The idea will change five times; the co-founder you dropped out for should not.
- Boredom is data. If you know within days that you hate the work, believe yourself.
- Read the curves. A beautiful product aimed at a declining trend line is still a dead product.
- A no is an opening. The investor who says you don't know what you're doing may be the one who teaches you.
- Hire your teachers. The best senior hire is someone you intend to learn from, not someone you intend to hide behind.
- Build what's unbearable to live without. Customers rarely ask for the differentiator; they ask for the blocker.
- Pull is easy to deny. Decide in advance which customer behaviors will force you to accelerate.
- Make the exit generous. When the deal changes, pay people to leave and keep the ones who choose to stay.
- Apologize specifically. Name the cause, name your share, name the fix, within a day.
- Control the inputs. Number goes up, number goes down; taste, craft, and shipped work are what you own.
