The designers who'll matter most aren't the ones with the best taste

For a long time, "good design" was mostly a visual claim. Clean layout, nice type, a coherent color system, maybe a delightful animation. That skill set still matters, but it's stopped being the thing that separates a good designer from a great one — mostly because AI tools can now produce competent visuals almost for free.

What can't be produced for free is judgment about a genuinely messy problem. And that's quietly become the actual job.

Polish was never the hard part

Most design problems that show up in a brief look simple on the surface. Make sign-up easier. Make the dashboard clearer. Make people trust the product more. The polish — the part that's easy to screenshot for a portfolio — was rarely where the real difficulty lived.

The real difficulty is underneath: sign-up is hard because three teams disagree about what data is actually required. The dashboard is unclear because the business itself hasn't decided which one metric matters most this quarter. Trust is low because of a support failure eight months ago that no interface can fully undo. These aren't visual problems. They're systems problems wearing a visual costume.

Designers who only ever practiced the visual layer are the ones most exposed right now, because that layer is exactly what's gotten cheap and fast. Designers who practiced diagnosing the system underneath it are becoming more valuable, because that skill was never about pixels in the first place.

What "solving complex problems" looks like in practice

It rarely looks dramatic. It looks like:

• Noticing that a feature request is actually two conflicting requests wearing one name, and asking which one the team actually meant

• Realizing the real constraint isn't the UI, it's that two backend systems can't agree on a single source of truth, and no amount of clever layout fixes that

• Sitting with ambiguity long enough to find the actual tradeoff, instead of jumping to the first clean-looking solution

• Being willing to say "the interface isn't the problem here" even when everyone was hoping the interface would be the fix

None of this shows up nicely in a portfolio shot. It shows up in how a person talks about their process, and in whether the thing they shipped actually held up once it met real users and real constraints.

Why this matters more, not less, going forward

As production gets faster and cheaper, the scarce resource stops being "someone who can make it look good" and becomes "someone who can figure out what it should even be." That's a harder thing to fake, a harder thing to automate, and a much harder thing to hire for — which also makes it the thing worth actually getting good at.

Taste still matters. It always will. But it's no longer the whole job. The whole job is becoming: can you sit inside a messy, contradictory problem long enough to find the shape of it, before you ever open a design tool at all.