AI is Making Designers ‘Coders’ and Stealing Our Best Debates

When I started my career, having an engineering background as a designer made me an anomaly. I could hand over working prototypes, drop code snippets, or dive into a new framework to see what was possible.

The real value of that coding experience—outdated and limited as it was—wasn’t just the technical artifacts I produced. It was the trust it built with engineers.

Rather than tossing designs over a wall, I was invited directly into the development loop. In turn, I brought engineers into the design process. When engineers understand the core problem and build genuine empathy for the user, they aren’t just building to spec; they’re invested in the solution.

My favorite engineering partners were the ones who pushed back. They questioned my design decisions not just on feasibility, but on utility. Sure, handoff moments could be tense, especially with tight deadlines or tight budgets. But that tension was healthy. It forced me to rethink my work, often leading to a simpler, far better design. And because I understood the basics, I could challenge them back—whether that meant digging through Stack Overflow or spinning up a local environment to prove a concept.

Beyond collaboration, coding things myself forced me to slow down and really wrestle with the constraints.

Today, AI tools are making "designers who code" the norm. Designers can now generate browser-ready prototypes with Claude or push code directly to a GitHub repo without knowing fundamental programming principles.

Excellence without understanding misses the point.

Is an AI generated prototype really building the same bridge between disciplines? I don’t think so. When we shortcut the learning process, we risk skipping the exact moments where healthy tension happens. I worry that without that shared struggle, we lose the deep trust and collaborative friction that actually lead to great products.