Once upon a time, designers donned many hats.
You were the person who could sketch a logo in the morning, build the brochure by afternoon, and code the landing page by evening. The T-shaped designer hadn't been invented yet — you were just a designer, and design meant doing all of it.
Then came specialization. UX research became a discipline. Visual design separated from interaction design. Frontend engineering split from product design. Job descriptions multiplied and narrowed.
And it made sense, for a while.
The Case for Specialization
Large organizations need specialists. When you're building a product used by millions of people, you need someone who thinks about nothing but navigation patterns, and someone else who thinks about nothing but data visualization. The problems are genuinely hard enough to warrant full-time attention.
The specialist era gave us careers. It gave us WCAG, design systems, and an entire vocabulary for talking about craft.
But it also gave us something less useful: a generation of designers who'd never shipped code, sitting across from a generation of engineers who'd never talked to a user, and both of them wondering why the handoff kept failing.
What's Shifting Now
The tools changed.
Figma put design and engineering closer together than they'd been in years. Framer made it possible to design in the browser without a computer science degree. AI is compressing the distance between an idea and its implementation even further.
The thing that used to require a team of three can now be done by one person with the right range of skills. And suddenly, the generalist is interesting again.
The New Generalist
The designer-developer isn't new. But the context around them has changed.
The old generalist was a generalist by necessity — there was no one else to do the work. The new generalist is a generalist by choice — they've decided that the constraint of having to own the full stack of execution produces better work.
I count myself in this group. My best work has come from moments when I was close enough to the implementation to understand what was possible, and close enough to the user to understand what mattered.
The pendulum hasn't stopped. But it's clearly swinging.
