What it means to become a designer at the edge of AI revolution

Hey there. Happy Saturday! Today, I want to share some thoughts on what it really means to be a designer at the edge of the AI revolution.

The Panic is Real.

Every designer I know is quietly asking the same question right now: will I still have a job? We watch AI generate ads in seconds, produce entire websites from a prompt, and mimic UI styles that took years to develop. The fear is understandable. But, I think it's pointing in the wrong direction.

The real question is: who will AI replace; and which ones will AI make ten times more powerful? These are not the same question. And the answer changes everything about how you should be preparing right now.

The people who are most afraid of AI and competing with AI are the ones most likely to be replaced by it. The ones who run toward it become something AI can never be, its most powerful user. To understand why, we need to stop looking forward for a moment. We need to look back.

Every revolution leaves something behind. And makes something priceless. When the printing press arrived, painters panicked. When photography was invented, illustrators panicked. When computers arrived, graphic designers who built posters by hand, cutting, pasting, composing physically, panicked. 

Every revolution in tools has been met with the same existential fear from the people whose craft it touched. But here's what actually happened each time: the revolution created mass adoption of the new, and simultaneously created a rising premium on the old.

Books didn't die when the Kindle arrived. They became objects. Things people chose deliberately, held with intention, collected and displayed. Hand-drawn illustrations didn't disappear when Photoshop arrived. It became luxury, something clients paid more for precisely because it was harder to produce. The newspaper didn't vanish when the internet arrived. In many cities it became an identity, a statement about how someone chooses to engage with the world.

The pattern is always the same. When something becomes infinitely reproducible, what was already rare becomes precious. When printing flooded the world with images, every visible brushstroke became proof that a human being was present: struggling, deciding, doubting. That evidence of a human hand didn't lose value. It became a value.

When everything becomes infinitely reproducible, what was already rare becomes precious.

AI is following the same pattern. Right now, we are in the adoption phase — the chaotic, exciting, slightly ugly phase where everyone is racing to use the new tool regardless of quality. Founders are cutting out designers entirely. Products are shipping that look AI-generated because they are. And nobody minds yet, because the speed feels miraculous.

But give it two years. When every startup looks identical, when every landing page shares the same AI-polished aesthetic, when every brand voice sounds like it was written by the same invisible hand — something will happen. People will start looking for the thing that feels different. Intentional. Human. And they will pay for it.

The question is whether designers will be ready when that moment arrives. Or whether they spent the transition hiding from the tool that could have made them extraordinary.

AI can design now, so what’s next?

Let me be honest about something most designers don't want to say out loud: AI can already do a lot of what designers do. Color theory, typographic hierarchy, layout rules, white space, visual balance — these are learnable principles. They are, in fact, the things we teach in the first year of design school. If it can be taught to a junior designer in a semester, it can be learned by a model trained on millions of examples.

Web design, in its most structural sense, is already largely replaceable. Combine readability with layout with typography with hierarchy and you get something AI handles with ease. The rules are knowable. The patterns are consistent. The output is predictable.

But there is a different kind of design work that AI cannot replicate — not because of any mystical human quality, but because of something more specific and more interesting.

AI cannot hold the whole picture. Not yet. Not in the way a designer who has worked across strategy, product, and brand simultaneously can feel when one decision has thrown an entire system out of balance. Founders who remove designers from their process discover this slowly and painfully. They start with good structure. Then developers respond to user feedback — one piece at a time, linearly, fixing what's broken without evaluating what the fix does to everything else. Six months later, the product is functional but incoherent. It looks like what it is: a series of isolated solutions that forgot they were part of a whole.

A developer sees a broken piece and fixes it. A designer walks in and sees what the product was trying to become and whether it's still on that path. That capacity — to walk into a messy product and feel the original intention underneath the accumulated decisions — is not a style skill. It is not aesthetic judgment. It is something closer to a narrative sense. Designers read products the way editors read manuscripts: not word by word, but as a living argument that is either holding together or quietly falling apart.

And there is something else. Something more fundamental to what designers actually are.

The Orbiting Designer: Doubt is the differentiator

Design, unlike most professional disciplines, has no finish line. This is not a weakness in the process. It is the process. A developer solves a problem and moves on. A strategist makes a decision and implements it. But a designer, a real one, never fully lands. They circle. They return to decisions already made and re-examine them. 

They sit in a meeting where everyone agrees on a solution and feel something is wrong before they can articulate why. They push back not because they have a better answer ready, but because their relationship with the product demands constant re-evaluation.

Most people call this indecision. I've come to believe it's the highest form of critical thinking available in creative work.

Because here's what that endless orbit actually produces: it produces products that remain coherent across time. It produces brands that feel intentional rather than accumulated. It produces the experience — rare and increasingly valuable — of encountering something that someone clearly cared about all the way through.

AI produces certainty. Clean outputs. Finished things. Design, at its core, is an act of sincere uncertainty. And sincerity is the one thing that cannot be generated.

When AI generates a design, it produces the most statistically likely version of good. It optimizes. It averages across everything it has learned. What it cannot do is doubt that output. It cannot feel the gap between what was made and what was meant. It cannot stand in a room of people who are satisfied and say: wait.

That capacity to wait, to hold quality to a standard that isn't finished yet, is what I now think of as the sincere hand. Not nostalgia for analog craft. Something more urgent: the only signal in a world of AI-generated output that proves a human being was genuinely present, accountable, and unsatisfied until it was right.

The last gate — And why designers must own it

Here is how I have restructured my own relationship with AI, and why I think it points toward the right model for designers who want to be relevant in five years.

I use AI extensively. For research, for pattern recognition across industries, for brainstorming at the beginning of a problem when I need to expand the possibility space fast. I use it to generate initial directions I might not have reached on my own, and to produce output that becomes raw material for the real work.

But nothing leaves my hands without passing through my judgment. Not because I distrust AI's output, sometimes it's remarkable. But because the moment something reaches a founder, a team, a user, it carries authorship. It makes a claim. And that claim has to be mine.

This is not a quality control step. It is the definition of what a designer is in the age of AI: the last gate. The person who is accountable for the result, not just the process.

Most designers are so afraid of AI replacing their execution that they haven't noticed the more important shift happening: the role of the designer is moving from maker to decision-maker. From the person who produces the work to the person responsible for whether the work is right.

Founders who cut designers out control the output. Designers who own the last gate control the outcome — forever, even after they've left the room.

This is actually a promotion. If designers can claim it.

The founders who are running entirely AI-generated processes right now are, in many cases, producing faster and cheaper. But they are also producing something that looks like everyone else. They are the last gate, but they don't know how to judge what's passing through it. They know it when something is broken. They cannot feel when something is merely adequate.

That gap, between broken and merely adequate, between adequate and genuinely right, is where the designer lives. And in a world where AI eliminates the broken, the distance between adequate and right becomes the entire competitive landscape.

Taste Infrastructure can become a thing now

For most of design history, taste has been considered personal. Uncopyable. Something you either had or you didn't, and even if you had it, you couldn't hand it to anyone else. The best you could do was produce work that demonstrated it, and hope clients recognized what they were getting.

AI changes this. And I think it changes it in a way that is genuinely new in the history of creative work. If you can explain why something works — not just feel that it does — then your taste can become a system. It can be encoded into rules, hierarchies, guardrails, component decisions, and voice principles that guide output even when you are not in the room. Your judgment can be made legible. And anything that can be made legible can, eventually, be made scalable.

This is what I mean by Taste Infrastructure. Not a style guide. Not a brand manual. Something more dynamic: a designed system that produces outputs consistent with a specific aesthetic philosophy, regardless of who or what is generating them.

Think about what this means practically. A company running entirely on AI-generated content still needs their output to feel like them, not like everyone else using the same models. The designer who can build the system that makes that possible is not selling hours of execution. They are selling something far more valuable: the ability for a company to be self-sufficient without losing coherence.

Taste was always considered personal, uncopyable. Now it can become a product — something built, sold, and scaled. The designer who can manufacture taste wins everything.

Every company that wants to differentiate in an AI-saturated market will eventually face the same problem: their tools are the same as everyone else's tools. The output will converge toward sameness. The only way out is to encode something distinctive upstream — before generation, not after.

That upstream encoding is a design problem. And it requires exactly the skills that survived every previous technological revolution: the ability to perceive a whole, hold it in mind, evaluate it against an intention, and articulate why it works or doesn't.

My perspective in the picture

I spent six years studying art, taste and visual craft. Then I spent five years teaching design to students who had never done it before. After that, I spent 9 years practicing product design in Silicon Valley. 

When you teach design, when you sit across from someone who has made something and you have to explain not just that it isn't working but why, and how, and what specifically to change,  you are forced to develop something most practicing designers never build: a verbal language for aesthetic judgment.

Most designers with great taste are mute about it. They feel it. They produce from it. But they cannot reliably explain it to someone who doesn't already share it. This is why designers have historically struggled to lead organizations — not because their judgment is poor, but because their judgment is illegible to people who think in language, logic, and strategy.

Teaching forced me to make my judgment legible. To find the words for why this composition creates tension and that one creates resolution. Why this hierarchy guides the eye and that one abandons it. Why this color palette signals trust and that one signals noise. Year after year, student after student, I built a vocabulary for something that most designers leave unspoken.

You spent years teaching others to see. That is exactly what qualifies you to teach machines to see.

That vocabulary is, I now realize, the foundation of everything that matters in the AI transition. Because AI doesn't respond to feelings. It responds to articulation. The designer who can tell AI exactly why something isn't right yet, and keep pushing until it is, will produce work that no one else can produce with the same tools.

And the designer who can encode that articulation into a system, who can build the infrastructure through which a company's entire creative output flows, is not competing with AI at all. They are becoming the intelligence that AI learns from.

What Comes Next

The designers who will matter most in five years are not the ones who learn the most AI tools. They are the ones who use AI to think bigger while using themselves to think better. Who treat AI as a way to expand their inputs infinitely while taking complete ownership of every output. Who understands that the value they carry is not in their hands,  it's in their judgment, their orbit, their sincere uncertainty, their ability to feel the whole.

The world right now is in its adoption phase. Everything is fast and cheap and slightly unfinished. Founders are cutting designers out and feeling smart about it. In two years, maybe three, the market will have its answer about what was lost.

I don't want to wait for that moment. I want to be already building the systems, the infrastructure, the encoded taste, so that when companies realize they need a designer who can do more than execute, there is something real to offer. Not just presence in the room. Not just a seat at the table. An architecture of judgment that scales.

That is what I am thinking now. And I think it is the only answer worth building toward. The Taste Architect doesn't design things. They design the conditions under which good design is always possible.

Share my newsletter

Loving my content so far? I’d appreciate if you can share my newsletter to a friend 🙂