AI didn't lower the craft bar. It exposed who never had it.

Show two designers the same AI-generated screen and you'll get two different reads. One sees a finished artifact. The other sees four decisions nobody made. That gap has always existed. AI just made it visible faster.

4 July 2026 · 4 min read

Run this anywhere: show two designers the same AI-generated screen. Same prompt, same output, twenty seconds each to respond.

The first said it was ready to ship. The second pointed at the empty state, asked what happens when the list is long, noticed the primary action was competing with the nav, and flagged that the error copy promised something the backend couldn't do. Same screen. One of them saw a finished artifact. The other saw four decisions nobody had made yet.

That gap has always existed. AI didn't create it. AI just made it visible in twenty seconds instead of six weeks.

Mediocrity was always here. It was just slow.

The panic goes like this: AI floods the industry with plausible-looking mediocrity, the bar drops, craft dies. I've been designing for seventeen years and I think this has it exactly backwards.

Mediocre work was always being produced at scale. It was hidden inside process. Six weeks of discovery, a workshop, three rounds of stakeholder review, a handover deck. By the time the artifact arrived, it carried so much ceremony that questioning it felt rude. The process was the alibi. Weak judgment could hide inside it for an entire career, because nobody could separate the quality of the thinking from the length of the journey.

Then the journey collapsed. When a screen takes minutes instead of weeks, all the ceremony evaporates and you're left staring at the only variable that remains: was this a good decision or not? Speed didn't lower the bar. Speed removed the scaffolding that was holding people above it.

Process was where weak judgment hid

I build AI tooling inside a design org as part of my day job. Prototyping tools, heuristic review workflows, the unglamorous plumbing. The single clearest thing I've learned: AI amplifies whatever judgment you bring to it, in both directions.

Give it to someone with taste and they move frighteningly fast, because they can evaluate output as quickly as it's generated. Accept, reject, redirect. Their bottleneck was never production, so removing production cost them nothing and gained them everything.

Give it to someone without taste and they also move fast. That's the problem. The output looks finished. It's fluent, aligned to a grid, reasonably worded. Every historical signal we used to detect unfinished thinking (rough edges, lorem ipsum, misaligned boxes) has been polished away by default. The work looks done, and whether it is done now depends entirely on the person looking at it.

We spent two decades building processes to compensate for the fact that we couldn't measure judgment directly. Now the artifact arrives instantly, naked, with no process wrapped around it. Judgment is measurable again. Some people are discovering theirs was never tested.

Craft is not years served

Here's where I need to be careful, because there's a lazy version of this argument and I don't want to make it.

The lazy version says juniors are cooked and seniors are safe. That's wrong. Second-year designers will dismantle an AI-generated flow because they can articulate exactly why the hierarchy failed. Fifteen-year veterans approve rubbish because their seniority was built on managing process, not on seeing.

Craft is not tenure. Craft is whether you can say why. Why this spacing, why this word, why this state matters, why this pattern will break at scale. If you can say why, AI is the best thing that ever happened to your career, whatever year of it you're in. If you can't, no amount of experience will hide that anymore, and honestly, it shouldn't have hidden it in the first place.

If you're early in your career and this frightens you, it should do the opposite. The path to craft used to be gated behind years of production grind. That gate is gone. You can now generate a hundred variants and force yourself to articulate why one works and ninety-nine don't. That's a taste gym no previous generation had access to. Use it.

Taste is the bottleneck again

For most of my career the scarce resource was production capacity. Enough hands to make the things. Entire org structures, entire careers, were built around rationing that scarcity.

Production is no longer scarce. Judgment is. The ability to look at fluent, plausible, instantly-generated work and know whether it's right: that's the whole job now. The rest was always scaffolding.

So no, AI didn't lower the bar. It walked into the building, turned all the lights on, and the bar is exactly where it always was. Some people are standing above it. The rest are learning that the process was never going to hold them up forever.

The best time in a decade to have taste is right now. The worst time in a decade to fake it is also right now.