LukeW: design systems get agents 90% there, visual feedback loops close the gap
LukeW · x · 2026-11-05
Luke Wroblewski's new article 'Design Feedback for AI Agents' addresses how to close the gap when coding agents apply design systems to production code.
Key points:
- Design systems can't (and shouldn't) cover every layout/functionality combination — they must stay lean to avoid spiraling in team size and complexity
- Their two-stage process: first pass driven by design system + agent instructions gets 90% there (earlier models only managed 50-60%)
- The remaining gap is closed with a rapid visual feedback loop: show agents where things don't line up and have them fix each issue until all variants and composite layout problems are resolved
- Feedback methods tried include Figma comments and annotated screenshots
A directly referenceable practice for teams running coding-agent workflows with design specs.
More from coding & agent
- Code mode isn't about MCP: it's about whether tool results go to context or program variables — ritakozlov · 2026-11-05
- Agent Arena: frontier models show 55-minute gap on long-horizon agent tasks — arena · 2026-11-05
- Agent-based kernel optimization faces a fidelity-throughput tradeoff: random data misleads, cold starts cost 12.7 minutes — simran_s_arora · 2026-11-05
- Solo dev builds window-free OS where a local Qwen3.5 9B generates every screen — leonardo_c · 2026-11-05
- Apple team ships Siri Evaluations: don't skip evals when wiring App Intents into Siri AI — rudrank · 2026-11-05
- RECAP: RL recipe for context compaction in long-horizon agents, +4.3pp on SWE-bench — Michael_D_Moor · 2026-11-05