AI Blogger Returns to Taobao Design Team for a Day, Finds the Job Completely Rewritten by AI
数字生命卡兹克 · wechat · 2026-11-17
AI blogger "数字生命卡兹克" rejoined Taobao's design team as a designer for one day after three years away, documenting how AI has reshaped the profession.
- The team splits AI work into three areas: creative assets nearly fully AI-generated, 200+ design knowledge bases ingested into AI, and an in-house tool called TaoDesign.
- TaoDesign evolved from chat-to-HTML, to a Figma/MasterGo plugin, to a fully self-built canvas aiming to make the designer's deliverable the shipped product itself, eliminating implementation loss.
- Standout features: one-click image/text swapping with real business data, style distillation (years of design specs distilled into AI-callable style libraries and Skills), synthetic "digital users" for 15–20 min research reports, page-data prediction, and real-time design wired to the codebase via MCP.
- The author's first task—delivering a full winter-promo design by 5pm (a week-long job in his era)—initially failed when he dumped six templates with an over-detailed prompt; the fix was coarse prompts, one page at a time, letting the model lean on the knowledge base.
- His takeaway: while agents run tasks, designers finally have time to think about what users actually want.
More from coding & agent
- Who are we becoming as AI engineers? Notes from AIE Code conference — TheTuringPost · 2026-11-17
- Will Brown nails agents' Markdown problem: written as data, read as code — willcb · 2026-11-17
- Customer vibe-coded a third rendering backend for Blender — wordgrammer · 2026-11-17
- Wattenberger shares full prompt to turn local coding-agent logs into a personal "river" viz — Wattenberger · 2026-11-17
- Ripple-65B builds its own local IDE in 3 hours at hackathon, >300 tok/s codegen — Azaliamirh · 2026-11-17
- VC Martin Tobias: yesterday's deliverable was a PDF, tomorrow's should be an MCP — MartinGTobias · 2026-11-17