The senior designer test, in the age of AI

Every designer can now produce a polished mockup in twenty minutes. That used to take a week. The fact that the floor moved up doesn't mean the ceiling did — and the gap between the two is where senior designers now live.

Here's a test I've been using on myself. It works for hiring too.

I call it the second draft test.

The first draft is the easy part

Give a designer — any designer, any seniority — a brief and a Figma file and Claude open in the next tab. In an hour they can produce a credible first draft. Clean layout. Reasonable hierarchy. The right component patterns. AI made this true. It's no longer a useful test.

The first draft is the easy part. The first draft is the floor.

The second draft is everything

Now show that same designer the work and ask: what's wrong with it?

This is where the ceiling shows up. A junior designer looks at their first draft and sees a finished thing. A senior designer looks at the same draft and sees a thousand small wounds — the hierarchy is slightly off, the empty state was generated and not designed, the copy is competent but generic, the loading behaviour wasn't considered, the mobile breakpoint has a touch target three pixels too small, the colour token used here doesn't exist in the system, the user it was designed for is "everyone" which means "no one."

Then they fix it. Not by re-prompting. By looking, and knowing, and caring. By having an opinion about what good means.

"AI gave everyone a competent first draft. Seniority is now defined by what you do in the second draft — which is exactly where it should have been all along."

Why I'm optimistic about this

The cynical read is: AI is going to flatten the design industry, juniors won't learn the fundamentals, everything will be slop. I'm not buying it.

What I think is happening is: the visible work — the deliverable, the screenshot, the artifact — has become commoditized. The invisible work — the taste, the rationale, the second draft — has become more valuable, not less. The market is going to figure that out within two years, and the designers who treated their first draft as the work are going to find themselves explaining what they actually contribute.

The designers who treated the first draft as a starting line are going to be fine. Better than fine, because their starting line just moved up.

What to do about it

Three things, in order.

Use AI heavily. Not optionally. The designer who refuses on principle is the calligrapher who refused to use a printing press. Move on.

Train your eye relentlessly. Study real interfaces. Critique your own work in writing. Read the books that are still in print fifty years later — Tidwell, Norman, Tufte. The fundamentals are what tell you what's wrong with the AI's draft. Without them you're just rearranging output.

Develop a stance. Have opinions about what good design is. Defend them. Be willing to be wrong about them in public. AI cannot have a stance — it can imitate one, badly. A stance is the most senior thing you can bring to a room.

That's the test. That's also the career.

— Victor