The New Full-Stack Designer
Marvin Schwaibold, design leader at Shopify, discusses how AI tools like Codex and Claude are transforming design workflows from canvas-based design to rapid prototyping, emphasizing that designers now function as editors and curators rather than craftspeople. The episode includes detailed design reviews of five websites, highlighting principles like clear hierarchy, progressive disclosure, and prioritizing user interactions over visual complexity.
Summary
Marvin Schwaibold, who leads the product design studio at Shopify and previously ran Molly design studio in Brooklyn, shares his evolving design process in the age of AI. Rather than replacing designers, AI tools like Codex and Claude Code have shifted the designer's role from creating detailed Figma mockups to directing AI through prompts and breakdowns of complex tasks, then curating and quality-controlling the output. Schwaibold emphasizes that designers remain in the driver's seat, responsible for understanding what looks good and steering AI in the right direction. At Shopify, his team works on major initiatives including redesigning the admin experience and building Sidekick, an AI infrastructure layer that lets merchants manage their stores through conversation rather than interface manipulation.
Schwaibold's previous studio, Molly, was known for meticulous motion design, typography, and what he calls the 'primal system'—a structured approach to content that allows users to explore progressively rather than being overwhelmed. Projects with Collins and other high-end clients showcased how small, dedicated teams could create exceptional work through focus and craftsmanship. The transition to Shopify allowed his studio to scale impact to millions of users while maintaining the same level of care.
A significant portion of the episode involves Schwaibold demonstrating his weekend side project: creating a grass shader from real video footage. This example illustrates his iterative AI workflow—initial failures, breaking down the problem into components (terrain mapping, grass blade variations, lighting), using Figma as a visual scratch pad to communicate intent to AI, and leveraging tools like DialKit to expose variables for fine-tuning. He created 91 prompts to get the animation feeling right, demonstrating that AI's infinite patience enables exploration that would be impractical with human collaborators.
The design review section analyzes five websites: Nstack (too many competing calls-to-action), Skill Sync (dense features hidden behind tabs, small typography), Instaplay (unclear primary goal, forcing sign-up too early), Mintlify (excellent example of clean, professional hierarchy), and Nullify (scroll-jacking, poor carousel interaction patterns). Common feedback emphasizes choosing one primary action per page, progressive disclosure, large tap targets, native scroll behavior, and using space to separate concerns. The review of Instinct's space technology site praises its visual storytelling while noting serif font misuse and non-swipeable carousels.
Schwaibold advocates for designers becoming 'full-stack'—understanding code, building prototypes, testing on real devices—while also specializing in areas like motion and information design. He argues against over-relying on AI-generated animations and illustrations without intentional art direction, citing examples like Gemini's helix animation where specialized teams spent significant time perfecting a single interaction.
Key Insights
- Designers are now more editors than ever before—AI can execute complexity and generate outputs at scale, but designers must decide what looks good and steer AI in the right direction, as AI itself lacks aesthetic judgment.
- The process of designing hasn't changed fundamentally; it's still extremely iterative and sequenced, but tooling has shifted from Figma canvas work into AI-driven workflows where designers quickly move from sketches to real, testable demos.
- At Shopify, the design studio builds all prototypes and demos directly into the product interface using a dev UI panel, allowing multiple designers to branch off with their own versions and merge them back together in the same language.
- Breaking down complex problems into discrete components—like explaining to AI where light is coming from in a grass shader, creating topographical maps, and identifying grass blade variations—is essential for getting AI to produce desired outputs.
- Most websites fail because they try to show too many important things at once, forcing users to be 'dictated to' rather than allowing exploration; successful design means choosing one primary thing to rally around and building narrative progressively.
Topics
Transcript
[0:00] Today, I'm joined by Marvin Schwaibold, who leads the product design studio at Shopify. He'll give us a behind-the-scenes look at his process, >> [music] >> where we'll see how he's using the latest AI tools to bring his designs to life, then review some viewer-submitted websites. Welcome to another episode of Design Review. Marvin, thanks for joining us. >> Thank you for having me. >> Before we get into your work at Shopify, um I'd love to hear more about your [0:30] process and the tools that you're using today to design with AI. >> I feel like the process itself hasn't really changed. It's still extremely iterative uh and and sequenced out. There's a lot of There's…
Full transcript available for MurmurCast members
Sign Up to AccessMore from Y Combinator
We’ll Be Inferior to Our Grandchildren—and That’s Great
A discussion about scaling societies and the relative ratio of good versus bad behavior, with one participant expressing concern that future generations will be intellectually superior to current humans, similar to how humans are vastly superior to ants.
YC Debates: Should We Slow Down AI?
Jacob Coxin argues governments should substantially intervene to slow frontier AI development, citing rapid capability growth (superhuman AI researchers by 2027), alignment risks, and misaligned corporate incentives. Robin Hanson counters that typical technological adoption patterns suggest slower real-world impact, that regulation is dangerous without concrete problems, and that liability-based approaches are preferable to centralized global regulation.
How To Build The Future: Eric Glyman, Co-CEO of Ramp
Eric Glyman, co-CEO of Ramp, discusses how the company evolved from Parabus to build a $60 billion financial operations platform, and how Ramp is leveraging AI agents and internal tools like Inspect, Research, and Glass to transform organizational productivity and reshape the future of work.
Grok Bot: What Changes When You Can Build AI Teammates
Lauren Tan and Roman Ugarte discuss Grokbot, a platform for creating AI teammate agents with their own identities and tool access that can work autonomously. The conversation covers how AI agents are transforming software engineering and non-technical work, with Lauren showcasing her 4,000+ PRs/month workflow using specialized bots in Slack and describing the shift from execution-constrained to execution-abundant organizations.
Grok Bot: What Changes When You Can Build AI Teammates
Lauren Tan and Roman Ugarte discuss Grockbot, a product enabling users to create AI teammates with their own identities and capabilities. They share how AI agents are transforming software engineering and business operations, demonstrating workflows where engineers ship thousands of PRs monthly by leveraging specialized bots integrated into team communication tools.