It was early March, I was sitting in my third coffee chat of the week with a designer five years into the field, and she said the word again. Taste. I’d heard it in pitch meetings, in performance reviews, in podcasts about AI and creative work, in Lenny’s newsletter, in VC blog posts about what makes a good founder. Everyone was suddenly talking about taste as if it were the most important skill in the field.

Honestly, they’re not wrong.

The skill that compounded hardest in the first year of the AI-design transition has a name now, and the name is taste. Editorial judgment about which of N generations to ship, which direction to push when five are plausible, which of a dozen microcopy variants captures the thing you actually meant. That work is where the value sits now.

But “taste is what matters” is a glib thing to say, because taste is expensive to develop, uneven across the population, and built through an apprenticeship path that’s closing fast. The profession’s gradient is about to get steeper. This essay is about what that means, and what we owe each other because of it.

What is taste, concretely?

I want to avoid the mystical version of this conversation. Taste is not a vibe. It’s a specific, describable skill with components that can be named.

Taste is the ability to quickly distinguish “this is the one” from “this is almost the one,” at the level of:

Composition. Where things sit on the page, how they balance, where the eye goes first, whether the negative space is working or fighting.

Typography. Whether the type pairs are doing what they’re supposed to, whether the scale is right for the context, whether the line height is comfortable, whether the mix of weights has rhythm.

Information density. How much is on the screen, how much should be, what can be cut without loss, what needs to be added to avoid feeling thin.

Emotion. Whether the piece feels the way it should feel. Trustworthy for a financial product. Warm for a healthcare product. Sharp for a developer tool. Whether it could be mistaken for belonging to a different category.

Brand fit. Whether it looks like it belongs to this specific company, this specific product, this specific moment, or whether it could be anyone’s work.

Taste is the ability to make those judgments fast, in sequence, across many pieces of work, without losing consistency. The person with strong taste looks at ten generated comps, points at one, and is right about 70% of the time in a way that their team can feel even if they can’t articulate it. The person with weak taste picks the one that looks most finished and is right about 40% of the time.

That gap, 70 versus 40, used to be a curiosity. In a process where the production was slow and deliberate, a mediocre picker could still ship decent work because they had time to course-correct. In a process where production is fast and cheap, the mediocre picker ships wrong-but-polished work at high volume, and nobody on the team has the time to fix it.

Why taste used to be undervalued

For most of my career, taste was a step in the design process that was small and fast, surrounded by steps that were big and slow. A designer would spend two days producing comps, ten minutes picking which one to push forward, another day polishing, another day soliciting feedback, another day iterating. Taste was the ten-minute step. Nobody really optimized for it because the optimization wouldn’t buy you much. Even a 50% improvement in taste efficiency saves five minutes in a cycle that takes a week.

So the profession optimized for the slow steps. Figma proficiency. Production speed. Documentation hygiene. Handoff discipline. Auto-layout mastery. The things that took the ten-hour tasks and made them nine-hour tasks, because those savings compounded.

Taste was talked about in blog posts and at conferences, but it wasn’t really a hiring criterion, except implicitly through the portfolio review. And the portfolio review was noisy because a junior designer with a good mentor could ship work that looked like it came from a senior’s taste even when the junior was mostly executing. The evidence of taste was hard to extract from the evidence of execution.

Why taste is now overvalued

So what changed? The slow steps got fast.

Production is close to free for a lot of design surfaces. Documentation mostly writes itself when the system is model-readable, which was last issue’s argument. Handoff has been collapsing into generation. Figma hygiene matters less when Figma is a polish surface rather than a production surface.

What’s left is the part that didn’t scale. You can generate ten comps in forty seconds, but picking the right one still takes a human with good judgment, and that judgment is where the whole cycle lives or dies now. A team with strong taste ships better work at the new speed. A team with weak taste ships faster bad work, which is worse than the slow bad work was, because there’s more of it.

So taste is suddenly the bottleneck. It’s the only part of the process that didn’t get automated. It’s also the most expensive part to develop, because it’s not a technical skill. You don’t learn taste from a bootcamp or a certification. You learn it from looking at thousands of pieces of great and near-great work over a long period of time, with someone more experienced than you pointing out what’s working and what isn’t, and slowly internalizing the patterns.

That’s a problem. Taste is now the scarcest skill in design, and the path for developing it was built assuming it was a slow, background process that happened in parallel with years of execution work.

The uncomfortable distribution

Taste is unevenly distributed in the profession. Senior designers with fifteen or twenty years of experience who’ve spent that time immersed in the field, looking at work, reading It’s Nice That and Brand New and Sidebar, collecting references, going to shows, studying the history of typography and composition, have very strong taste. They developed it by osmosis during the years they were doing production work. The production work paid for the time it took for the taste to develop.

Junior designers don’t have strong taste yet. That’s not a knock. You can’t skip the thousand hours of exposure. The junior designers I’ve watched stay on a good taste trajectory are the ones actively studying the field: subscribing to newsletters, building personal reference libraries, analyzing other people’s work, asking why a thing works instead of just admiring it. The junior designers who are not on a good trajectory are the ones treating design as a skill you learn once (how to use the tools) rather than a skill you build over decades (how to see).

The uncomfortable part is that the apprenticeship path that built taste for previous generations was: do execution work for years while absorbing taste by osmosis from the seniors around you. You’d spend Monday through Thursday making comps, and on Friday your senior would review them, and over the course of eighteen months your sense of what works would calibrate against theirs. The osmosis was free because the execution work you were doing was paid for.

That path is closing. Execution work is cheap now. Teams that used to hire three juniors to do production under one senior are hiring one mid-level designer and giving them Claude. The juniors don’t get hired in the first place, so they don’t get the eighteen months of calibration, so their taste doesn’t develop, so they don’t become the seniors with strong taste ten years from now.

This is a known problem in a lot of professions that went through similar transitions. The junior associate at a law firm who used to do document review is now an AI prompt operator, and the partners are starting to worry about what happens to the pipeline in ten years when nobody has done the document review that used to build the pattern recognition. The junior developer who used to write boilerplate is now prompting Copilot, and the seniors are wondering whether they’re training the next cohort or just saving money this quarter.

Design has the same problem. The profession’s gradient is about to get steeper. The designers at the top will have even more pricing power than they do now, because their taste is rare and the other skills that used to mitigate the rarity are gone. The designers in the middle will be fine if their taste is strong, and in trouble if it isn’t. The designers at the bottom will struggle to get the experience that would let them climb.

The slow steps got fast. Taste stayed slow. The thing that was a ten-minute step became the entire game.

What seniors owe juniors now

I’ve been thinking about this a lot, because the traditional answer (give them execution work and let the taste develop by osmosis) doesn’t work anymore. The execution work isn’t there to give.

I’m not a manager and I don’t run a design org, so take this as an outsider’s read. The replacement isn’t obvious, but I think it has to involve deliberate taste training. Not project assignments. Not production tasks. Actual structured practice in looking at work, evaluating it, and defending the evaluation.

The shape of this, on my team at least, has been this. Take thirty minutes a week and spend it together looking at other people’s design work. Could be a recent Klim Type Foundry specimen. Could be a redesign a competitor just shipped. Could be a Pentagram case study. Could be the new Linear changelog page. Look at it, say what’s working and what isn’t, disagree out loud, defend the take. Junior designers develop taste by watching seniors argue about design in front of them. That argument used to happen implicitly during execution reviews. It has to be made explicit now, because the execution reviews don’t happen as often.

The other piece is this. Give juniors generated work to critique, specifically. Not their own work. Not work anyone cares about shipping. Generated options, produced on demand, for the purpose of the exercise. Five versions of a landing page. Three versions of a dashboard. Ten options for an illustration style. Pick the best one. Defend it. Then pick the worst and say why. Rapid, repeated, with feedback.

That’s more explicit teaching than design has historically done. It might feel weird. It’s also genuinely the only way I can see to build the skill the profession is about to ride on, for a cohort that doesn’t get the on-the-job exposure that built it for us.

Where this lands

I don’t have a neat ending for this one.

The taste tax is real. It’s going to reshape the profession’s gradient. Senior designers with strong taste are about to have a great decade. The apprenticeship path for the next generation is broken, and it’s going to take deliberate work to rebuild.

The rebuilding is everyone’s job, but it’s especially the job of the seniors. If you’re in that bucket, you got your taste from somewhere. Somebody pointed things out to you. Somebody argued with you about what was working. You owe that to the cohort behind you, and you have to find new ways to give it because the old ways don’t really scale anymore.

If you’ve found a deliberate-practice setup that actually moves the needle on junior taste, please tell me about it. I’d love to write that up.

Jameson