Anthropic designer on using Claude to design Claude: wireframes first, borrow familiar interfaces
petergyang · x · 2026-11-16
Nate Parrott, the designer behind Claude Design at Anthropic, shared his process for building Claude features with Claude in a new interview:
- Have Claude ask the questions a good designer would: "What problem are we solving? What interfaces do people already know?"
- Start with wireframes so Claude doesn't get lost in fonts and colors; for an inbox feature, ask it to study existing inboxes and wireframe in the same structure.
- His key insight: "Claude has no idea what its design biases are" — to the model, it's always doing something for the first time, so it thinks every idea is original and needs human correction.
- Counterintuitive finding: with Claude Design they "could not ask too many questions" — users loved the friction, contradicting the traditional no-friction design dogma.
The quotes come from Peter Yang's podcast episode with Parrott.
Related event: Anthropic Designer Nate Parrott Breaks Down Claude Design Workflow(3 posts)→
More from coding & agent
- Dev open-sources local Apple Health MCP server that feeds workouts to AI agents and pushes plans back to Apple Watch — PrinzPwnage · 2026-11-16
- Coding agents crack a 1994 Jeep's ECU via a $3 Arduino and start controlling it — LeviTurk · 2026-11-16
- 'My favorite PRs': raylib author hails a change that adds features by deleting 100+ lines — yacineMTB · 2026-11-16
- Indie dev shows agent workflow turning before/after UI screenshots into beat-synced changelog videos — yihui_indie · 2026-11-16
- jscdec: Pure-JS V8 bytecode decompiler ships with Electron .asar/.exe unpacking — cyb3rops · 2026-11-16
- anti-slop repo adds 23 LLM-judged oxlint rules to block AI-written TypeScript slop — teropa · 2026-11-16