CHAPTERS
- 0:00 – 0:39
Welcome + who Dylan Field is and what Figma set out to build
Sarah and Elad introduce Dylan Field, founder/CEO of Figma, and frame the conversation around life after the Adobe deal, collaboration products like FigJam, developer workflows via Dev Mode, and Figma’s early AI direction.
- •Dylan’s role and Figma’s position as a dominant collaborative design tool
- •Core theme: a “single multiplayer canvas” for building digital products
- •Preview of topics: Adobe acquisition fallout, FigJam, Dev Mode, AI products
- 0:39 – 1:55
Figma’s original vision: closing the gap between imagination and reality
Dylan reflects on the earliest Figma pitch and how the company refined a broad “North Star” into something concrete: making design accessible while still powerful. He notes how today’s AI moment makes the original ambition feel newly achievable.
- •Initial mission: eliminate the gap between imagination and reality
- •Reframed North Star: “make design accessible to all”
- •Balancing power with learnability and simplicity
- •Design expanding beyond “pixels and software”
- •AI as a force bringing the two visions back together
- 1:55 – 4:21
Why the Adobe acquisition made sense—and why it ultimately didn’t happen
Sarah asks about the canceled Adobe acquisition and how Figma navigated the emotional and strategic uncertainty. Dylan explains the rationale for joining Adobe to accelerate the vision, and how regulatory resistance forced the company to walk away after a long process.
- •Deal rationale: accelerate Figma’s roadmap and learn from Adobe’s team
- •Regulatory scrutiny as the blocking factor
- •16-month period of uncertainty and eventual decision to call it
- •Internal communication to keep the team aligned as the path narrowed
- •Relief of starting the year with clarity (and independence)
- 4:21 – 7:16
Post-deal clarity: sprinting through the roadmap and building platform foundations
Elad notes Dylan’s renewed energy; Dylan describes how the team kept building through the acquisition limbo. He outlines Figma’s platform pillars and why the last 16–17 months were about laying the right foundations for what comes next.
- •Regulatory process was draining, but execution didn’t stop
- •Three pillars: FigJam (ideation), Figma Design, Dev Mode (design-to-code)
- •Major platform components (e.g., Dev Mode) weren’t even announced at signing
- •Focus on foundations to enable future expansion
- •Future direction: AI across idea→design→code and better developer support
- 7:16 – 9:35
Why FigJam exists: remote collaboration, whiteboarding, and ‘fun’ as a feature
Dylan explains that FigJam came from observing existing behavior inside Figma and then accelerating it—especially during the pandemic shift to remote/hybrid work. He argues that “fun” isn’t superficial; it helps create psychological safety and increases participation in ideation.
- •Product expansion driven by real user behavior already happening in Figma
- •Pandemic as catalyst: teams needed shared spaces to collaborate
- •FigJam shipped quickly (beta in ~6–8 months)
- •Differentiator: “fun” supports safety, play, and participation
- •Tactical examples: stickers, emoji reactions, cursor interactions, cursor chat
- 9:35 – 10:49
FigJam evolving beyond brainstorming: meetings, decisions, and AI setup assistance
Sarah shares how FigJam helped with onboarding and team cohesion; Dylan observes how customers use FigJam as a structured meeting/decision tool, not just a brainstorm board. He flags this as a key place where AI can help—by generating or setting up meeting artifacts faster.
- •FigJam used for visual thinking even by non-designers
- •Onboarding and belonging benefits for new teammates
- •FigJam as a “better way to run a meeting,” not only ideation
- •Boards can function like visual docs/knowledge bases
- •AI strategy: speed up setup and structure for outcomes
- 10:49 – 12:49
Dev Mode: making Figma truly useful for developers (and fixing developer pain)
Dylan details Dev Mode as a developer-focused view of designs to reduce confusion, mistakes, and friction in implementation. He explains the impetus—developers were a large user segment with lower satisfaction—and the research, acquisition, and iteration that led to today’s product.
- •Developers are ~1/3 of weekly engaged Figma users
- •Problem: insufficient value for developers and lower NPS
- •Common pains: navigation difficulty, building the wrong thing, missing context
- •Background work: research, multiple pivots, acquisition of Vizzlie
- •Dev Mode features: notations/properties, change tracking, workflow context
- 12:49 – 13:16
Extensibility and workflow integration: plugins, Jira, and design-system mapping
The discussion expands from Dev Mode’s core to the broader ecosystem: integrating the tools developers already use and keeping design systems consistent from design to code. Dylan hints at ongoing evolution beyond the current Dev Mode starting point.
- •Dev Mode as a container for dev workflow context
- •Plugin integrations (e.g., Jira) and custom extensions
- •Bridging design systems between design and code
- •Maintaining consistency and tracking changes over time
- •Roadmap emphasis: Dev Mode is “just the start”
- 13:16 – 15:05
Figma’s approach to generative AI: hybrids, usefulness bar, and shipping carefully
Elad asks how Figma is incorporating generative AI; Dylan describes early signals (scaling laws, diffusion models) and a conceptual model: design as “art applied to problem solving,” suggesting hybrid AI approaches. He emphasizes taking time to ship something truly useful.
- •Early AI inflection points: scaling laws + diffusion improvements
- •Concept: design blends art (diffusion) and problem-solving (LLMs)
- •Hybrid model approaches may be important for design tooling
- •Product stance: ship when it’s at least clearly useful
- •Figma plans to share more as AI capabilities solidify
- 15:05 – 17:14
How AI changes design work: lowering the floor, boosting efficiency, crossing boundaries
Dylan lays out two major impacts: making design accessible to more people (“lowering the floor”) and removing repetitive work for professional designers. He also points to AI’s potential to connect idea→design→code and to enable agents that cross tool/data boundaries.
- •AI can broaden who can participate in design
- •AI can remove repetitive tasks and increase designer throughput
- •Interfaces matter: better AI interaction inside the tool
- •Opportunity across transitions: idea→design and design→code
- •Agents crossing boundaries between data/tools are underexplored
- 17:14 – 21:14
The iterative human–AI loop: direct manipulation, faster feedback, and ‘latent space’ navigation
Dylan highlights the importance of tight iteration loops: the more immediate the feedback between user and model, the more powerful the creative workflow becomes. He discusses direct manipulation interfaces and speculates on richer ways to navigate latent spaces beyond “magic prompt phrases.”
- •Direct manipulation + real-time updates create a visceral, productive loop
- •Parametric/slider-based control in 3D as an example of the future
- •Core requirement: faster human↔model iteration cycles
- •Prompting as a primitive interface; better controls will emerge
- •Idea: define controllable semantic vectors (e.g., optimistic↔pessimistic)
- 21:14 – 25:31
Will AI replace designers? More software, higher abstraction, and role evolution
Sarah asks about job displacement fears; Dylan argues that near-term reality is augmentation, not replacement, because design depends on deep context and human understanding. They predict more software will be created, with engineers shifting toward higher-level intent and design-oriented work as code abstraction rises.
- •Why replacement is hard: emotion, context, cultural signals, affordances
- •Near-term trajectory: augmentation, access, and efficiency gains
- •Engineering abstraction will rise; fewer humans writing every line of code
- •More (and better-designed) software likely gets produced overall
- •Some “no UI” futures exist, but demand for interfaces persists
- 25:31 – 29:25
The future of AI UI and multimodality: chat, voice, and where new interfaces fit
Elad probes how AI changes UI—chat-centric interfaces, agentic worlds, and multimodality like voice. Dylan argues new media rarely replaces old media entirely; instead, each interaction pattern finds its niche, and exploration (even hype cycles) is healthy for HCI progress.
- •Skepticism that everything collapses into chat as the universal UI
- •New interaction modes tend to add options rather than remove existing UIs
- •Voice is valuable in certain contexts but not for all-day information work
- •Mixed feelings about gesture-heavy ‘Minority Report’ interaction
- •Pro-experimentation stance: exploration helps discover what works
- 29:25 – 31:27
Investing and ecosystem curiosity: why Dylan spends energy outside Figma
Sarah asks about Dylan’s breadth of interests and investing; Dylan says he genuinely enjoys helping founders and learning from building across the ecosystem. He frames it as energizing and mutually beneficial, even if the motivation is partly personal curiosity.
- •Investing as a way to help people bring ideas to life
- •Learning across tech trends and startups feeds back into leadership
- •Personal motivation: fun, curiosity, and service to others
- •Shared context: Sarah/Elad as investors too
- •Acknowledgment that the rationale isn’t purely strategic—it’s energizing
- 31:27 – 36:30
Leadership and culture at scale: evolving as a CEO while keeping Figma’s ‘feel’
Sarah asks how Dylan’s leadership changed from year 1 to now; Dylan describes learning management through each growth stage and balancing empowerment with efficiency. He notes the culture’s continuity—creative, maker-oriented, humble—while acknowledging that different values (like play vs. execution) get emphasized at different times.
- •Leadership evolution: continuous learning across company stages
- •Tension to manage: empowerment vs. efficiency/streamlining
- •Cultural constants: creativity, making, thoughtfulness, humility
- •Values shift by season: currently “sprinting hard,” emphasizing “run with it”
- •Scaling risk: ownership ambiguity; encourage initiative when something needs doing
