"I don't know when it happened, but my job changed shape."
That's an anonymous creative writing to Creative Boom, describing work that no longer arrives as a blank page. It shows up half-made—generated, roughed out, "nearly there"—and the job is to fix it. Tidy the type. Sort the spacing. Make it look like a human cared.
The money's fine, mostly. But there's an ache underneath, and the designer names it with unusual precision: "I used to feel like a maker. Now, I feel like a finisher. A correction service."
Then the line that should stop the whole industry cold: "I worry I'm getting slower at the part that mattered most—the thinking, the having of ideas—because I rarely start anything from scratch now."
I worry I'm getting slower at the part that mattered most—the thinking, the having of ideas—because I rarely start anything from scratch now.
Sit with that. The ache isn't fear of replacement. It's the quieter, worse suspicion that finishing other people's half-made things is actively making you worse at the part of the work that made you a designer.
This week the story gets darker, because the people leading the transition are now openly admitting they don't know where it goes, and the "maker-to-mender" framing turns out to be diagnosing the wrong disease entirely.
Here is the consensus, stated cleanly by someone who should know. AWS Vice President of Design for Applied AI Solutions Hector Ouilhet: "We spent our careers taking an input, executing a task, and returning the output. But the thing that got us into this craft isn't that. It's something higher-order: taste, discernment, judgment. That's the work only we can do."
Several leaders at Figma's Config Leadership Collective reportedly said a version of this. Expertise isn't dying; it's migrating. Away from execution, toward judgment. The pixels commoditize; the point of view does not.
I believe this is true. I also believe it's internally contradictory in a way nobody's saying out loud.
Because judgment is not a skill you can be promoted into. It is not a credential that activates once execution becomes beneath you. It is the residue of having executed—badly, repeatedly, and long enough to learn what "better" actually feels like.
Michael Polanyi called this tacit knowledge: "we know more than we can tell." The professor writing in UX Collective puts the operational version bluntly—recognizing quality "still requires the calibration that only comes from having struggled to produce it yourself, made the wrong call, and learned what better actually looks like."
Read that again against Ouilhet's line. He's describing taste as the thing that got us into the craft. But that phrasing hides the mechanism. Taste didn't get anyone into anything. It came out—slowly—of years of making things that weren't good yet.
So here's the contradiction. The industry is telling designers that judgment is the new value at the exact moment it's dismantling the process that manufactures judgment. AI is absorbing the execution—the failed drafts, the wrong calls, the calibration—that was never the product. It was the training.
The maker-to-mender panic, then, is a category error. Designers think they're mourning the loss of execution. They're actually mourning the loss of the loop that turned execution into discernment.
Eilidh McDonald names the specific casualty: "if we're all AI creative directors now, we'll miss out on those wonderful serendipitous moments when you discover something new, in the process of trying to make something else."
That's not nostalgia. That's a description of where judgment is grown. You don't discover the better idea by evaluating options—you find it by being elbow-deep in the wrong one. Mending machine output doesn't generate those moments. Correction is a closed loop; it can only ever return the output toward what already exists.
Now run it forward one generation and it gets structurally worse.
The professor offers the detail that should keep design leaders awake, reportedly: "A designer with twenty years of experience may have only two or three years of experience working in an AI-first environment." Portions of expertise now have a half-life. The profession is evolving faster than careers can accumulate.
Which means the mentor is no longer ahead of the mentee in the way mentorship has always assumed. Airbnb Vice President of Design Teo Connor reportedly says the honest thing: "Everyone's a beginner, and everyone should be experimenting." OpenAI's Ian Silber: "In just a few months our entire industry has shifted."
That's leadership admitting, on record, that they're figuring it out in real time. Fine. Beginner's mind is a healthy posture. But stack it against the pipeline problem and the recursion turns vicious.
Trace it. Juniors now generate interface concepts, flows, and personas in seconds—and never spend the years failing at execution that used to build calibration. Their mentors have only two or three years of AI-first judgment themselves, and are being asked to teach a version of the craft they're still learning. Nobody in the chain is completing the loop that produces taste.
The mender role, in other words, requires exactly the calibrated eye that the mender role no longer produces. It consumes judgment. It doesn't manufacture it.
A profession that runs this way eats its own supply of expertise. This generation of menders can correct machine output because they built their eye the old way—on blank pages and bad drafts. The next generation won't have that. They'll be menders who never made, evaluating work against a standard they were never forced to internalize.
You might hope systems thinking is the escape hatch—that even if surface craft commoditizes, platform judgment survives. But the systems-thinking argument doesn't dodge this; it inherits it. Experience architecture is itself a tacit skill, reportedly "accrued through real experience over time" (though in the source that phrase describes taste). The apprenticeship problem doesn't disappear when the work moves up the stack. It generalizes.
So what actually rebuilds the loop?
The Creative Boom community, without using this language, is describing an engineered answer. Claire McDivitt: "Make sure there is always something in your life where you're still responsible for the first mark." A personal project, a sketchbook, a sculpture, an idea that exists purely because you wanted to explore it.
We usually file these under coping strategies—self-care for the burned-out creative. That undersells them badly.
They are, right now, the last deliberate mechanism for re-manufacturing taste. The first mark is where calibration lives. If your paid work never lets you make a wrong call from nothing, the personal project becomes the only place you still can.
And there's a companion skill that has to be taught rather than assumed. From UX Collective: "Learning to say 'good enough' is a design skill, not a compromise." You cannot judge good enough without a felt sense of the gap between "fine" and "exactly right"—which is, again, tacit, and again, earned only by having closed that gap yourself.
Here's the shift the profession has to make. Judgment cannot be treated as an inheritance that automatically transfers once execution leaves the room. It's a byproduct of struggle, and if the struggle is gone, the byproduct doesn't appear on its own.
Which means judgment-building now has to be engineered—deliberately, structurally, on purpose. Protected time on the blank page. Mentorship that admits it's teaching a moving target and builds failure back in rather than optimizing it away. Personal work reframed not as therapy but as the training reps that paid work no longer provides.
The alternative is the mender role as a dead end: a profession that declared judgment its new value while quietly severing the only process that ever produced it.
The anonymous designer already senses this. "I'm getting slower at the part that mattered most." That's not fatigue. That's a working eye reporting that it isn't being trained anymore.
Taste was never a moat you inherit. It's a residue you earn. And right now we're being handed the moat while the well that fills it runs dry.