The Designer's New Job: From Producing to Deciding

Something is moving under the floorboards of the design profession this month. You can feel it in the publications that are usually a step ahead. UX Collective, Medium, Smashing Magazine, Designlab. Each one writing around the same shape, from different angles. They are not announcing a trend. They are noticing that a trend has already arrived, and trying to give it a name.

The clearest sentence I have read came from a piece on Medium in May.^[1] It is no longer about what you produce, because AI can produce too. It is about what you decide, because AI cannot decide well yet.

That is the whole essay, really. The rest is just looking at the shape from different sides.

The center has moved

For a long time, the center of design work was production. Not the glamorous part. The middle. Turning a wireframe into a high-fidelity screen. Building the components. Generating the variants. Cleaning up the spacing. The middle was where most of the hours went, and where most of the apprenticeship happened. You learned by making.

The middle is what AI has eaten first. Not because it is the most valuable layer (it never was), but because it is the most pattern-based. Patterns are what these tools see best. A Figma file is a structured document. A button is a button is a button.

So the production layer has compressed. What remains, what the tools cannot replace, is the layer above it and the layer below it. The judgment that comes before the work. The evaluation that comes after.

Those two layers were always the most valuable. They were just invisible, because they were folded into the production hours. Now they are visible, because production hours have collapsed.

The Junior Paradox

The shape of this gets uncomfortable when you trace it forward. A senior designer with the new tools can do the work that used to need a team of two or three. The entry-level seats where designers learned the craft, where they built intuition by making a thousand small decisions, are thinning.

Several writers have called this the Junior Paradox, and the phrasing is right. The tools make experienced designers more productive while quietly cutting the rope ladder that produced experienced designers in the first place. If production is where you used to apprentice, and production is automated, where does the next senior come from?

This is not a new question. It is the same question typesetting asked of typography, that the darkroom asked of photography, that the recording studio asked of musicianship. Every time a craft moved from the hand to the machine, the field had to invent a new kind of apprenticeship. The old one stopped working.

What I notice is that the field has not yet invented it. The conversation is still mostly about the loss. Which is a fair place to start. But it is not the place to end.

Two ways to use the tools

Designlab's State of AI in UX report puts a number on the surface of the change.^[2] Ninety-three percent of designers are already using generative AI in some part of their workflow. Seventy-three percent identify AI as a collaborator. These are large numbers, and they arrived quickly.

But the report makes a sharper distinction underneath the headline, and it is the more useful one. There are two ways designers are using these tools, and they look like the same thing from the outside.

The first group uses AI as a speed layer. Drafts come faster. The first round of variants is generated rather than drawn. The work shape stays the same, just compressed.

The second group uses AI as a decision layer. They generate fifty directions where they used to consider three. They spend less time in the tool and more time in the framing. They are not faster at making. They are more thoughtful about choosing.

The first group will be fine for another year. The second group is doing something different in kind. Their workflow is less about producing and more about steering. Less about the hand and more about the question.

What the hand used to know

The piece Smashing ran in April, on production-ready deliverables, sits in the same current.^[3] When the gap between concept and working prototype is minutes instead of weeks, the value of being able to make something falls and the value of knowing whether the thing is right rises.

Telling someone whether their idea is good is a strange skill. It is not really a skill at all, in the way design skills used to be measured. It is closer to taste, and taste is the most stubborn thing in the field to teach.

Taste comes from making. From the thousand small adjustments. From the time you spent at three in the morning trying to figure out why a piece of typography looked broken to you when you could not say what was wrong with it. The hand built the eye. The work built the judgment.

This is the part of the migration that has not been thought through. We are moving designers up the stack, from making to judging, but the judgment was a byproduct of the making. If you remove the making, the judgment does not appear on its own. It has to come from somewhere new.

I do not know yet where it comes from. I do not think anyone does.

From craft to curation

There is a more hopeful version of this story, and it is also true.

Photography did not destroy painting. It moved painting somewhere else. It freed painters from the obligation to render the visible world and let them do something stranger with paint. The painters who tried to compete with the camera at its own game lost. The painters who let the camera have what it wanted, and went looking for what only paint could do, found something new.

Design will probably do something like that. The work that AI can do, AI will do. The work designers will keep is the work that requires understanding what the work is for. Defining the constraints. Framing the problem. Holding the user in mind. Caring whether the thing is honest. None of this is new. It was always the deepest part of the craft. It just used to be hidden inside the production hours.

What is changing is not the depth of design. It is the location. The center is moving from hands in the tool to eyes on the output. From drawing the interface to defining the conditions that shape all possible interfaces. From producing to deciding.

John Maeda, in his Design in Tech report this year, called this shift from UX to AX.^[4] The user experience used to be a thing you designed. The agent experience is a thing you shape, by writing the rules an agent follows when it designs on your behalf. The work is still designing. The artifact is the rulebook, not the screen.

The question that stays open

None of this means the craft is in trouble. It means the expression of the craft is moving, and the field has not yet decided what to do with the people who used to enter it from the bottom.

The question I keep turning over is whether the migration preserves the depth, or whether something quiet and necessary gets left behind when the hand no longer touches the material. The best designers I have known were not just good at choosing. They were good at choosing because they had spent so much time making. The two were the same thing.

Maybe we will find a way to build that depth without the production years. Maybe the new apprenticeship is something we have not seen yet, and it will look obvious in a decade. Maybe the depth is replaceable. Maybe it is not.

For now, the honest answer is that the field is in the middle of finding out. Which is, perhaps, a place worth being. The work has not become smaller. It has become harder to see.