A collection of thoughts on design, systems, and the role of AI in shaping how products are built.
Design tools have long acted as passive canvases, leaving designers responsible for enforcing structure, consistency, and intent. AI-led tools change that dynamic by actively interpreting rules, constraints, and relationships — and why human judgment must remain the source of intent, not an afterthought.
Continue reading →When AI started executing my directions well, I concluded it had gotten better at design. I was wrong about what that meant. The effort it takes to direct AI precisely enough to get useful output — that effort is the design. AI executes. It doesn't decide.
Continue reading →AI can generate interfaces, but it cannot design them. The difference lies in relational meaning — the invisible logic that connects every decision in a product. Why prompting isn't precision, why remixing isn't reasoning, and what only a designer can provide.
Continue reading →AI improvises. Design can't afford improvisation. How diffusion models and LLMs each break down differently in product work, why inconsistency compounds as a product scales, and how to flip the relationship so the designer leads and AI assists.
Continue reading →Three practical ways to use AI in the early stages of product design work — learning new frameworks faster, generating ideas by analyzing what's already working, and synthesizing user research data without losing the signal.
Continue reading →Three more ways AI earns its place in a design workflow — refining design system tokens and type scales, organizing assets for developer handoff, and making sense of the raw output from product strategy workshops.
Continue reading →