AI can produce screens. It can't decide which problem is worth solving.
The nature of design work has changed more in the past two years than in the decade before it. Not because the market demanded new skills, but because a large share of the daily work no longer requires a person.
AI can now design a screen, write the interface copy, propose a full user flow, build components, even synthesise research findings. These aren't future promises - they're part of how I already work.
But there's one thing it still can't do: know which problem needed solving in the first place.
I used to think depth was the risky bet
Early on, I believed safety was in breadth - knowing a little of everything so I'd never be cornered. I now think the opposite.
Once a machine can produce "a little of everything" in seconds, shallow breadth stops being an advantage. Depth becomes the differentiator.
Depth is no longer a design skill
Depth today means understanding the whole picture before opening any tool. It means answering three questions:
Why does this problem exist at all?
Who actually benefits from solving it?
AI made execution cheap. What's scarce now is knowing which problem is worth solving - and owning that decision.
Which constraints are real - and which did we impose on ourselves?
The third question is the most neglected and usually the most expensive. Many decisions that look like external constraints are internal habits nobody has audited.
Breadth isn't a tool count
Breadth is language, not tooling. Talking feasibility with engineers, business impact with product, risk and trade-offs with decision-makers. The value isn't knowing every discipline — it's moving a decision between teams without it losing meaning.
Acceleration doesn't transfer accountability
AI accelerates your execution. It does not absorb the consequences of a wrong call. When a product fails, nobody interrogates the tool.
That's why the most valuable designers today aren't the ones shipping the most screens. They're the ones thinking most clearly before they design. They use AI to shorten the path — not to outsource the thinking.
The takeaway
AI will raise the minimum quality of design for everyone. That's a new floor, not a ceiling.
The ceiling sits somewhere else: in the depth of your judgment, the clarity of your thinking, and your willingness to own the decision.
I shared this idea there too - join the comments and tell me your experience.