On exploring the idea maze, vibe coding, and preserving craft even when the models are “cooking.”
With AI tools and models coming out faster than anyone can keep track of, designers and product builders are left wondering: Which technologies will stick? And how will they reshape the way we design and build software? Against this backdrop, Figma Co-founder and CEO Dylan Field and Y Combinator President and CEO Garry Tan sat down for a wide-ranging conversation about the staying—and growing—power of design, and the difference between making something work and making it work well. They also touch on the spirit of play and experimentation that has propelled Figma’s growth, what Dylan would tell founders starting companies today, and the early pivots that kept Figma from a potentially short-lived existence as a meme generator.
Read their conversation—edited and condensed for clarity—below.
On AI and design
How do you view AI changing the face of design right now?
Design has an opportunity to differentiate software in this age of AI. We’ll see whole new workflows emerge where designers, engineers, and product will work together in different ways than they’re doing today, or at least different cadences.
Current AI tools like ChatGPT or Midjourney enhance specific tasks and augment human capabilities in narrow domains, whereas many believe that Artificial General Intelligence would possess human-like cognitive abilities across various fields, potentially reasoning and learning like a human.
So far, AI is very much in the tool category. People are using AI to just do more and explore more. We talk a lot about lowering the floor, raising the ceiling, making it so more people can participate in the design process, but it’s also about raising the ceiling in what you can do.
I always had this image in my head of the idea maze, where you go down all these different branching paths as you ideate through a product journey. You get more breadth now in terms of the things that you can try, but it still requires depth to fully explore.
We’re sort of in this uncanny valley where prompt apps don’t give you well-designed apps yet. How do you see this playing out?
It’s interesting to look at the language people use. They talk about “getting locked in” or “I’m cooking,” “vibe coding,” et cetera. Maybe people are using this language because it’s the feeling you get when you’re developing something really fast.
There’s always this call and response with tools as you’re trying to get thoughts out of your head onto a screen. The faster you make that feedback loop, the more you can get into that flow state. Also, the more fun you can have. One of our values as a company is play. We’re always trying to make it so that whether it’s as a culture or as a product, Figma is fun to use and is a tool that just lets you creatively express. The opportunity is: How do you create that feeling—even if you’re not an expert—and have that rapid ideation loop going at all times? But you also want to give people a way to not just get started and prototype rapidly but also get to the finish line. That’s where the disconnect is—and not just for design, but also for code.
The faster you make that feedback loop, the more you can get into that flow state.
Are founders focused enough on design?
People seem to, year over year, care more about design than ever. People also understand the value of design. The question is not just, can you make it work? It’s how it works that actually matters.
The question is not just, can you make it work? It’s how it works that actually matters.
How are you thinking about the evolution of AI models in relation to design, especially when considering the two regimes of image diffusion and code generation and the potential of multimodal models?
It almost goes back to the question, "What is design?" Maybe these models are not great at design yet because on the art side, you’ve got diffusion, on the problem-solving side, you’ve got LLMs. It’s not clear that people have figured out how to marry techniques together yet. There are so many definitions of design, but one I like is [that design is] art as it applies to problem solving. A designer’s role is to bring in all this context—like the current cultural moment, or the brand experience, how it all ties together—so much more than just, “here’s this two-liner problem I want to solve.”
As more time has gone on, I’ve become more confident in the designer’s role. I believe it’s going to be one of the critical roles in building software in the future. I want to be careful with the word “designer,” too, because I think a lot of people will probably call themselves that in the future. People who call themselves developers today might call themselves designers in the near future—not because they’ve stopped writing code, but because they’re thinking more holistically about the product experience.
Do you think that people are going to do more chat-based interfaces, or is it more visual design? Or is it even going to be terminal and prompting?
Telnet (short for “telecommunications network”) is a client/server protocol that enables bidirectional communication with remote systems over local networks or the internet, allowing access to virtual terminals.
I’m seeing pretty much everything right now, which is exciting. It feels like we’re kind of exploding out the possibility space of how you actually can work. Prompting feels like the telnet days of AI. There’s so much more ability to work with tools through different interfaces. There’s a lot more to try, like different design patterns and things like AR, VR, XR. We’re going to see entirely different dynamic interfaces start to pop up. I just love how humans interact with machines. We don’t have BCIs yet, but one day that’ll be the final frontier.
On Figma’s origins and first users
How did you come up with the idea for Figma?
It was a process. Going all the way back, we started talking about what would become Figma in late 2011. I was at Brown at the time, Evan was my TA. We were asking the question of "why now"? We started with technologies that were shifting and would create all sorts of opportunities in late 2011. For us, that was WebGL, which Evan had a lot of expertise in. We explored a lot of different stuff. We did meme generation at one point, which was probably the low point of Figma. We built a really cool photo editor, too, but it was just not something that made sense because we were doing it in the browser.
Adobe Fireworks (formerly Macromedia Fireworks) was a bitmap and vector graphics editor designed for web designers to create website prototypes and application interfaces.
Then we saw [Adobe] Fireworks get killed and we thought, “Wait, maybe there’s an opportunity here.” We formed the thesis that as software gets easier to build, design becomes more important. At the time, it felt like this tiny market, but in retrospect, it was a tiny market that was rapidly growing.
A lot of VCs might have looked at how many designers there are in the country and then multiplied that out by whatever they thought they could spend, and then come to the conclusion that the number was too small. Whereas you made a bet. Was it hard to convince people that this was a good market to enter?
If you looked at the seed pitch for Figma in June 2013, it was all over the place. A lot of our framing up to investors was, “Oh yes, [design] is a small thing right now. We think it’ll grow.” Eventually, our users pulled us into the process we’re in now: How do you go from idea or brainstorming in FigJam to slides where you align, to going to design where you’re trying things in prototyping, to getting development, shipping to production?
Read Shishir Mehrotra’s 10 rules for leading great team meetings, where he explains why the best product teams often put as much care into how the work gets done as what product gets shipped.
Our early adopters were minimalists who liked how few features we had, and companies like Notion and Coda (called Krypton at the time) who resonated with cloud-first performance in the browser because they wanted that for their teams, too. I remember going to Shishir [Mehrotra]'s office, then Coda’s CEO (now CEO of Grammarly), and after months of trying to get a team to adopt Figma, getting them onboarded and driving away only to get an email that the fonts broke. We turned the car around immediately and fixed it—a great example of going the extra mile for your first customers.
What features really wowed users early on?
When we launched our closed beta in December 2015, there was no multiplayer. It was just kind of like in the browser, a new design tool, high quality, but minimal feature set. In the months following, Evan started working on our real-time collaboration multiplayer engine.
When we finally launched the GA in October 2016, the comments we got were like, “if this is the future of design, I’m changing careers,” or like “a camel is a horse designed by committee.” It was very much this reaction from the design community of, "I’m the designer that’s going to be in the corner. I’m going to do this amazing work, and I’ll do the grand reveal," which is kind of the agency mindset. Versus the product team approach of we’re all in together, and we’re going to work through this and figure it out.
One of our early users tweeted, “we’re gonna do a design party,” and invited everybody to a Figma file. At the time, there were not many limits on how many people could be in the file, so for the 48 hours after launch, the team was just firefighting, trying to make sure this one file, the design party, didn't crash. It started as people making fun of Figma and turned out to be a great advertisement. People started to jam in Figma and make stuff together. It showed what we were trying to do, even though people were maybe not buying into the message at first.
On building for the long term
How have you developed the design culture inside Figma?
We’ve been fortunate to have incredible designers at Figma. We are always trying to balance approachability with power, making sure that Figma is easy to use for both new and experienced users. This leads to the question of how to make the simple thing simple, but keep the complex thing possible. These challenges are best solved by working with the best people, and the best people often come up with the best solutions. Our internal culture at Figma encourages everyone in the company to help make Figma better, with initiatives like Maker Week. Many of our best features have come from these efforts.
Any advice for founders who’ve made it from zero to one but are struggling with scaling?
Be mindful of the loop you’re in and be self-aware. What are you doing the most of right now? Then go replace yourself in that task. If you’ve got the resources, if you don’t have the resources, figure out how to get them either by being profitable, by raising funding, or by being really clever. If you can keep doing that loop, that’ll lead to team building or lead to figuring out the way to delegate work properly. The danger is becoming so reactive that you never self-improve the organization.
Given how AI is accelerating development, what would you tell founders starting today?
Move as fast as you can. A lot of people come to me and say, “I think I should do this two-year build.” As someone who did one, I say, don’t. Almost no one has to do that at this point. If there’s a way to move quicker, you should do it.
