直觉是假设发生器：Dylan Field 谈品味、简洁和尽快上线

The conversation was recorded live in front of an audience at Figma Config, and Field opens by naming the community as the highlight of the event. He is also candid that he would have improved his own demo. He says the conversation about AI-generated software that began at Config, covering craft, quality and design as a differentiator, is the one he hoped would start.

Field jokes that the Make Design feature gives you "the most obvious thing in the most obvious form possible," and he agrees with critics who say so. That framing sits inside a larger point he returns to throughout. In a world where AI creates more software, the need for unique, high-quality design only grows.

Asked about his line that design is art applied to problem solving, Field says it is not originally his and that other definitions exist, such as design as dialogue. He likes this one because design involves creativity and unique expression, but it must also match a real user need. Pure problem solving, he says, is utilitarian and lacks soul, and he finds the combination of the two beautiful.

Lenny cites colleagues who describe Field's product taste as a sixth sense and plays a clip of Rick Rubin explaining that his value comes from knowing what he likes and being decisive about it. Field answers with his own framework. Intuition generates hypotheses, which people then put forward, debate, and test against data until a working hypothesis remains.

To feed that process, Field says he reads everything about Figma he can find, from social media to support channels. He keeps asking questions to reach root problems, since people who say they need X often really want Y or Z. He also encourages his team to dig the same way.

On changing his mind, Field uses Figma's Pages feature as an example. He remains skeptical that it is the most elegant solution, but users and the team insisted they needed it. His general rule is that if a decision will not be fatal, he shares his skepticism and lets the person go ahead. Over time that builds trust.

What persuades him is something concrete that can be debated. He asks for examples and follow-up questions, and if nobody has answers yet, he will send the team off to find them. He admits this can frustrate colleagues who think the point is obvious. He sees the habit as a perfectionist streak that works as long as it does not bottleneck the organization.

Lenny raises Brian Chesky's remark from the previous year about getting rid of product management at Airbnb. Field tells a story about the aftermath: a product manager at Figma did a mock Chesky impression at a team dinner, just before Field walked in forty minutes late. He used the moment to tell the PM team he believed in them.

Field says the lines between product manager, designer and engineer are always blurry. PMs and designers need some technical grounding. Designers and engineers need a sense of business goals and user needs, and engineers and PMs need taste and care for craft. All of them should talk with users, and he suggests research could make it a four-legged stool.

Product people tend to fail, he says, when they treat the job too much like process. Good process helps, but a PM still has to talk to users, have a strategy and hold a point of view. They also need to bring the team along so everyone is happy at a milestone, because "if everyone's unhappy, you failed."

Asked whether he is bullish or bearish on product management, Field says everyone is currently learning to do a bit more of each other's jobs. He still sees immense value in product, design and engineering, and expects all three roles to persist. The best PMs, in his view, wrap strategy and point of view into a framework so everyone knows the destination and how to get there.

Lenny reports that colleagues say Field furrows his brow and insists there must be something simpler. Field says anyone who has built product knows how hard simplicity is, and that the more you add, the harder it is to stay coherent. He points to an essay his co-founder Evan introduced him to, which uses the term irreducible complexity: one plus one sometimes equals one and a half.

For Figma, the challenge is making the tool more powerful without making it more complex, which he calls extremely hard. If nobody watches, he says, a product becomes "a monstrosity really fast." He admits parts of Figma are too complex as a system, even though every local decision was reasonable, and says those parts need revisiting as a whole.

Asked whether simplicity is a top-down job, as HubSpot co-founder Dharmesh Shah described it, Field says everyone is responsible for it. He quotes a principle he attributes to others: "Keep the simple things simple. Make the complex things possible." It is unfortunately easy, he adds, to make simple things complex.

Lenny points out that Figma took three and a half years to launch and roughly five before its first paying customer. Field says that was way too long and tells people not to do it. He says stronger hiring earlier would have helped, and he credits Sho, who joined as an engineering director, with giving a presentation in his first week that catalyzed the team to ship.

His advice is to get product out as fast as possible, because the faster you ship the more feedback you get. FigJam and Figma Slides shipped quickly. Dev Mode took far longer, he says, because some directions did not work and the team had to really understand the developer user. It looks simple from outside, he notes, but took at least three times as long as FigJam.

On where to hold the bar, Field repeats a rule he learned from Evan: for a launch you get quality, features and deadline, and you choose two. Because software can be iterated on, you can ship a minimum feature set with a quality bar, which he calls a "minimally awesome product." Later iteration should improve quality as well as add features.

Field explains the story of how he found early users. Inspired by network-visualization work he had seen at LinkedIn using Gephi, he mapped the central designers in the Twitter graph around 2012 or 2013, then looked at their work. He contacted the ones whose work inspired him and offered to buy them coffee. Most said yes, and he used the meetings to learn and to show them Figma.

He describes a fanboy moment with designer Tim Van Damme, whose icons Field had used as test cases for the vector networks work. Van Damme is now on the team and doing the icons for the UI3 redesign. Lenny notes the story is usually told as a growth hack, but Field says it started as seeking feedback, and that this works especially well with designers because they give good feedback.

Field recounts a user research session with a designer that ran for hours and needed a bottle of wine because text editing in the product was so slow. He also retells the story, with some uncertainty about details, of turning the car around to help Krypton, the company later known as Coda, with an issue. Its founder, he says, did not realize at the time that he was Figma's first team customer.

Lenny notes that Field spots trends early, from WebGL to CryptoPunks, and asks what excites him now. Field names websim, a hallucinated internet where you type a URL or prompt and a model invents the site. Figma Ventures has invested in it. He sees it as a lean-forward entertainment tool and describes world building through an accumulating context window.

Lenny demos it live by typing a fake Gmail address for Field, and the invented inbox appears. Field also shares a colleague's story of a dream in which websim produced a marketing site for a "Frog Jam" version of FigJam, with lily-pad sticky notes. The point, he says, is how fun it is, even though generative UI is not what Figma is going for.

On scaling from intern to CEO of a company with over a thousand people, Field says the answer is all of the above, with a mindset of constantly adapting. Mentors can be the community, the people he hires, investors, coaches, or even the people he mentors. New founders and Figma interns have taught him things, so he tries to stay ready to absorb new information.

Reflecting on the community around Figma, he says serving people who shape the world's technology is a privilege and a responsibility he tries to treat as motivation rather than weight. He says that means continuing to simplify the product, champion design and quality, and advance the state of the craft. He adds that computing is still early and invites people to send him what they build.

In the lightning round Field says he is fascinated by how different each large language model is and what each does best. He teases that if you hack them the right way and get them in the right mood, they will do interesting things. He declines to elaborate beyond calling it his diplomatic answer.

His favorite piece of advice is that when people give you advice, they are really giving themselves advice in your shoes. He applies it to himself, noting that his own advice to Lenny would be advice to himself in Lenny's position. It is a useful reminder to weigh guidance against the giver's circumstances.

Lenny closes by asking about Field's stint as a child actor at five. Field jokes that his differentiators were that he could read, sit still and was decently cute, and that once puberty hit, computer science was the next step. A clip of an old eToys commercial plays, and Field quips that the ad made the company go bankrupt.