Ex-Apple design lead: Most people only tap 1% of AI's creative potential
HamelHusain · x · 2026-09-02
Anshu Chimala, who led software engineering and design teams at Apple for 12 years, published a piece on Lenny's Newsletter on why most people get generic "slop" from LLMs when designing—and how to unlock the model's hidden creativity.
Core thesis: LLMs are next-token predictors, and post-training pushes them toward outputs that average everyone's preferences, while good design is the opposite—risky, bold, opinionated. You must push the model out of the average and toward the fringes of the bell curve where unseen ideas live.
Demos include a conversational calorie tracker built in 3 prompts with Claude Fable 5, a space exploration game in 2 prompts with Claude Opus 5, and a dynamic landing page with Claude Opus 5 + GPT-5.6 Sol. The full end-to-end process is paywalled.
Related event: Ex-Apple design lead shares 8 tips to unlock AI's creativity(3 posts)→
More from coding & agent
- Claude Fable 5.1: Costs Drop 45%, Agent Benchmarks Double — kimmonismus · 2026-09-02
- Claude Code 2.1.257 Released: Defaults to Fable 5.1 — ClaudeCodeLog · 2026-09-02
- Claude Code 2.1.257: Defaults to Fable 5.1, Hardens Auto Mode Security — ClaudeCodeLog · 2026-09-02
- Mental Health FAQ: RAG Chatbot for Mental Health Inquiries — madeofAjala · 2026-09-02
- Bezalel unifies memory, email, and MCP tools for AI agents — Rasmic · 2026-09-02
- Anthropic Is Back? Fable 5.1 Hands-On: Major Gains in Coding and Writing — every · 2026-09-02