October 19 2026
(from Gemini)
- Co-founder & CEO of Figma
- Design is the New Code & Product Differentiation: Launched in August 2012, Figma has evolved from a core software interface design tool into a visual platform spanning whiteboarding, slide decks, marketing, and diagramming. Defining design as "creativity meets problem-solving," Field argues that as AI coding models mature, code is no longer a primary bottleneck or competitive moat. Although app store launches are growing exponentially due to AI-assisted development, overall user app consumption remains flat. Consequently, value and differentiation are shifting up the stack to design, brand, marketing, a bold non-consensus point of view, and the ability to elicit a genuine emotional reaction from users.
- The $20B Adobe Deal & Antitrust Myopia: After building Figma for over a decade, Field agreed to a $20 billion acquisition by Adobe, feeling Figma had finally reached sufficient scale (1,100–1,200 employees) to join the graphic design pioneer without being absorbed and disappearing. However, a grueling 16-month regulatory review across the US, EU, and UK (CMA) ultimately blocked the merger. Field criticizes regulators for defining markets by looking backward at historical boundaries rather than forward at how AI is expanding the entire software landscape. He points to the blocked Amazon–iRobot (Roomba) acquisition as a prime example of idealistic antitrust policy backfiring: blocking the deal forced Roomba into bankruptcy and effectively handed the robotic vacuum market to competitors in China.
- Post-Merger Crisis Leadership & The Severance Playbook: Throughout the 16-month regulatory limbo, Field kept Figma's "foot on the gas" so the business would remain strong regardless of the outcome. When the deal officially collapsed on the first day of winter break in late 2023, Field called a Monday noon all-hands. The prevailing mood among employees wasn't devastation, but relief that the exhausting 16-month ambiguity was finally over. To ensure 100% alignment for the independent startup grind ahead, Field sent a company-wide message offering a no-shame voluntary severance package: anyone who felt burnt out or had joined expecting a corporate job at Adobe could take a few months of pay and leave on good terms (with the sole condition that they wait six months before reapplying). Only a small percentage took the buyout, leaving a focused, re-energized team that later took Figma public.
- Surviving the "SaaS Apocalypse" & Competing with Frontier Labs: Despite partnering with Anthropic via MCP connectors, Figma faced a surprise competitive narrative when Anthropic launched "Claude Design," prompting online commentators to declare "Figma is dead." Field attributes the clash to internal communication mix-ups inside hyper-scaling AI labs, noting that Figma's business actually accelerated to 48% year-over-year revenue growth following the announcement. He emphasizes that the market often mistakes raw model progress for product progress. Building professional software requires crafting an instrument-grade feel and mastering complex user workflows over many years. He points to Cursor as proof that app-layer SaaS is far from dead: even before training its own models and while competing directly against frontier labs, Cursor won by obsessing over its target users and seeing around corners.
- Lowering the Floor vs. Raising the Ceiling of Creativity: Retracting an earlier prediction he made at Figma's Config conference, Field notes that AI has not raised the creative ceiling—it has only lowered the floor. While lowering the floor democratizes digital production, prompting an AI model without advanced scaffolding yields results anchored to the statistical mean (average) of its training distribution. Only humans raise the ceiling by investing care, taking bold risks outside the Overton window, and creating things never seen before. Because workplaces are now flooded with low-effort AI slide decks and memos, Figma instituted an internal cultural rule: any internal writing generated with AI must be labeled "Written with AI" at the top, preserving organizational trust and signaling when a human has genuinely invested care into a document.
- Code as Canvas Material, Figma Weave & The "AI Psychosis" Loop: While AI is excellent for overcoming the "blank canvas problem" by generating an initial draft to react against, relying solely on prompts filters a creator's vision through the model's biases. Figma's product strategy treats code as a material directly on the canvas—allowing users to alternate between "building to think" with AI and molding the details by hand. Through generative shaders and a node-based product called Weave, designers can pipe AI model outputs into other inputs to sculpt video and photo work like clay. Field also observes a recurring "AI psychosis" loop: whenever a new model drops, algorithmic feeds surface cherry-picked results that make the model look omnipotent for a few days, until users quickly spot its repetitive tropes and must work to push past them.
- Verifiable vs. Non-Verifiable Domains & Human Agency: While AI models are rapidly stacking IQ points in formal, verifiable domains like mathematics and cybersecurity—highlighted by OpenAI releasing 700+ papers on famous math problems overnight—progress in non-verifiable domains like design and human taste remains incremental. Unlike current reinforcement learning systems, human brains possess an immediate loss function, an innate drive for novelty, and the ability to hold user needs, brand culture, and business constraints in their heads simultaneously. Ultimately, Field frames the AI era as a personal choice between consuming and creating: people can either passively let AI think for them and join the hive mind, or use AI as leverage to learn faster, climb steeper mountains, and become bold makers.
Comments