Part 1 -- The Designer Is Dead.

In 2001 I bought a copy of Sagmeister's Made You Look. I had no formal design education, no built visual language, maybe a year of actual work behind me. The book hit like a revelation. I thought I'd found the shortcut, that cleverness was the thing, that if the idea was smart enough, the rest would follow, and I had plenty of clever ideas.

So I pivoted hard into clever. I equated the idea with the work. I left execution entirely out of the equation.

Looking back, I cringe. What I produced looked like it understood something it didn't. Smart on the surface, empty underneath. Not actually clever , because cleverness without execution isn't cleverness. It's a pose.

What I'd missed was everything underneath that book. The obsessive craft. Specific solutions to specific problems built over years. The judgment that made the cleverness land. I'd seen the output and extracted the wrong lesson.

There has always been a path. You started at the bottom of a studio doing the work nobody else wanted: production files, comp variations, resize requests, the fifth version of a banner that would run once and be forgotten. It was unglamorous. It was repetitive. And it was, for the designers who came out the other side of it, the actual education.

Not because the work was meaningful in itself. Because doing it built something that couldn’t be taught directly: an eye. A sense of what holds and what doesn’t. A feel for hierarchy, weight, proportion, the moment something tips from clear into cluttered. You learned by doing it wrong a hundred times and feeling the wrongness — in a rejected comp, a misread layout, a headline that almost worked but didn’t. That feedback, accumulated over years, is what eventually becomes taste.

That path is being removed. Not all at once, and not without warning — but systematically, and probably permanently. The entry-level production work that served as the training ground is being automated. Not because studios want to remove opportunity, but because the tools now make it irrational not to.

The problem isn’t that AI does the work faster.

The problem is that the work it’s replacing was never really about the output. It was about the person doing it.

The shortcut that skips the point

When a junior designer uses AI to generate options, iterate layouts, and produce polished-looking work before they’ve developed the judgment to evaluate it, something specific goes wrong. It’s not that the output is bad — often it’s competent. It’s that the designer never develops the ability to know why it’s good, or isn’t, or is almost right but slightly off in a way that matters.

The eye develops through struggle with specific problems. When the struggle is removed, so is the development. You can produce work that looks senior without becoming senior. And that gap — between the surface of the work and the judgment behind it — is invisible in a portfolio and catastrophic in practice.

There is a second problem, quieter and more corrosive. AI models are trained on aggregated output — on the accumulated visual and verbal production of the culture at large. They are, by design, oriented toward the center. Toward what is already known to work, already recognized as good, already legible to the widest possible audience. That’s their function and their limitation.

A designer who forms their eye through AI-assisted production is, in a real sense, being trained by an averaging machine. Their instincts are calibrated against a statistical mean. They learn to produce work that is correct, that reads as competent, that no one will object to — and that carries none of the particularity, the lived specificity, the vernacular of a genuine creative point of view.

Vernacular isn’t style. It’s the residue of a specific person having worked through specific problems in a specific time and place. It can’t be prompted. It can only be accumulated.

The designers who will matter in five years are not the ones who learned to use the tools most fluently. They are the ones who used the tools while also developing, through some other means, a perspective that is genuinely their own. The question is what that other means looks like — because the traditional path is no longer reliable.

I see my 25-year-old experiments posing across portfolios and brands now. The cleverness is there. The execution is not. The shortcut found a faster vehicle, but it arrives at the same place I did in 2001 — work that looks like it understands something it doesn't.