Design in an AI world: the mockup was never the job

The anxious version of this conversation assumes the designer's job was making the artifact. It never was. Production got cheap and the actual job, deciding what should exist, got more valuable. Here is what that means in practice.

By Timothy Nice9 min readUpdated September 2026

What actually changed

One thing changed, and it changed everything downstream: the cost of producing a plausible version of almost anything fell to nearly zero.

Not the cost of a good version. The cost of a plausible one. A competent-looking landing page, a reasonable dashboard, a workable component library, a serviceable brand direction. All of it is now minutes of work. Everything people are calling a crisis is a consequence of that single change.

When plausible is free, plausible stops being worth anything. The value moves to the decisions that plausible cannot make for you.

What the job was always about

The designer's job was never the mockup. The mockup was the cheapest way to make a decision visible so other people could argue with it.

Strip away the artifact and the work that remains is: understanding a customer well enough to know what they will not tell you, choosing what to build and what to refuse, deciding what a thing means before deciding what it looks like, and holding a standard when everyone is tired and the deadline is Friday. None of that got easier this year. All of it got more visible, because the busywork that used to hide it is gone.

Which skills gain and lose value

SkillDirectionWhy
Taste and editingUp sharplyWhen you can generate twenty options, knowing which one is right is the whole job
Systems thinkingUpGeneration without rails produces six variants of everything; tokens and components are the rails
WritingUpYou direct models in words, and the product's voice is now most of its personality
Breadth across the stackUpOne person can now credibly carry strategy, design and build, and that person moves faster than a relay
Production speed in a toolDownPixel execution is no longer scarce; being fast in Figma is table stakes, not an edge
Deliverable craft for its own sakeDownBeautiful specs that document an unexamined decision are more expensive than ever
Process administrationDownHandoffs, tickets and rituals exist to coordinate specialists; fewer specialists, less coordination

How teams reorganize around it

The old shape was a relay: research to strategy to design to engineering to QA, with a document at each handoff. Each handoff existed because the next person could not do the previous person's work. That constraint is weakening, and the process built on it is the next thing to go.

  • Small groups of broad people beat large groups of narrow ones, because the coordination cost disappears.
  • The meaningful unit of work becomes a shipped outcome rather than a completed deliverable.
  • Review replaces approval. The question stops being is this done and becomes is this right.
  • Quality becomes everyone's job in practice, which means someone still has to own the standard or it evaporates.
  • Titles blur before org charts do, which is uncomfortable for about a year.

What founders should look for now

  • Ask what they removed from a project, not what they added. Editing is the scarce skill.
  • Ask to see something shipped and in use, not a case study with invented metrics.
  • Ask how they decided, not how they executed. Execution stories are now cheap to tell.
  • Prefer people who can carry a thing from decision to production over a perfect specialist you then have to staff around.
  • Be suspicious of portfolios where every project looks like every other portfolio. That is now a production signature, not a style.

What to do about it this quarter

  1. 01

    Use the tools seriously for a month

    Not a demo. Ship something real with them. Opinions formed without use are worth nothing right now.

  2. 02

    Practice deciding out loud

    For every project, write the three decisions that mattered and why. That document is your actual portfolio now.

  3. 03

    Widen by one adjacent skill

    Designers learn to ship. Engineers learn to judge visual quality. Founders learn to direct both.

  4. 04

    Raise your standard for the last ten percent

    Motion, empty states, copy, focus order. Everyone can now reach ninety percent, so the last ten is the whole difference.