WhoBuilt

Exploring the idea maze, vibe coding, and the power of “locking in”

Mar 13, 2025· 38 min· 66K views·Watch on YouTube

Speaker

Dylan Field, co-founder and CEO of Figma, talks with Y Combinator host Gary about AI and design, Figma’s early product choices, and lessons from growing the company.

AI expands the design process

Field sees AI mainly as a tool that helps people explore more ideas and lets more people contribute to design. He describes product development as an idea maze: AI can widen the paths people try, but exploring them deeply still takes work. He expects designers to remain critical because good design draws on user research, context, empathy, and judgment.

Fast feedback makes building more engaging

Field connects phrases like “vibe coding” and “locking in” to the rapid call and response between a person and their tools. Faster feedback can help people reach a flow state and make creating more enjoyable. He also points out that a quick prototype may still need substantial work to become a polished design or maintainable codebase.

Figma emerged through exploration

Field and co-founder Evan Wallace began exploring startup ideas in 2011, initially focusing on WebGL after setting aside drones. They tried projects including meme generation and a browser photo editor before settling on design software, based on the idea that design would matter more as software became easier to build. Figma’s early pitch was unfocused, and it took another year and a half to two years before people outside the company used the product.

Users and collaboration shaped the product

Early Figma users valued its minimal feature set and browser-based workflow, though the product still had broken features and no multiplayer when its closed beta began in 2015. After the 2016 general launch, a public design party brought users together in one file and strained the servers, demonstrating collaborative use even as some designers resisted the approach. Field says users also pulled the team toward products such as FigJam and Slides, while internal testing and customer feedback helped guide improvements.

Grow by making room to learn

Field says the founders spent several months pivoting before committing to Figma, then needed patience while building a usable product. In hindsight, he wishes he had hired faster when a user sent a detailed, 12-page list of requested features. His broader advice is to notice what occupies the most time, replace yourself in that work when resources allow, and keep looking for ways to ship sooner.

Takeaways

  • Use AI to explore more options, then invest time in understanding and refining the strongest ones.
  • Shorten the feedback loop between an idea and something users can try.
  • Treat repeated user requests and detailed feedback as possible signs of product pull.
  • Dogfood the product and fix basic workflow problems before expecting new use cases to emerge.
  • As a founder, regularly identify the work consuming your time and delegate it when you have the resources.

Mentioned in this talk

Search inside this talk

What was claimed 14

Related talks