I walked into a restaurant in Marylebone last month -- somewhere newish, opened in the last year, a friend had recommended it -- and within about four seconds I knew. Not that the food would be good. I hadn't seen the menu. Not that the service would be excellent. Nobody had spoken to me yet. I just knew. This place is right. The lighting. The materials. The sound. The proportions of the room. The way the chairs sat under the tables. The distance between tables. Something about the weight of the door handle when I pushed it open. All of it, in four seconds, collapsed into a single feeling: yes.

Try explaining that to someone.

I've spent fifteen years trying. Fifteen years in advertising, writing strategy briefs that attempt to translate gut feelings into slide decks. Fifteen years in rooms full of smart people, trying to articulate why something works. Why this campaign feels right. Why that brand identity misses. Why the first option is better than the second even though the second tests higher. The entire discipline of strategy is, at some level, the art of making the unconscious conscious. Taking the four-second feeling and turning it into a forty-slide presentation that a client can approve with confidence.

And I'm going to be honest: I've never fully cracked it.

The gap

There is a gap between experiencing taste and explaining taste that I don't think ever fully closes. You can narrow it. You can learn the vocabulary. You can study design principles and colour theory and typography and spatial design and all the formal frameworks that attempt to codify what "right" looks like. And that helps. Genuinely. Knowing why generous white space creates a feeling of luxury, or why a serif typeface signals tradition, or why warm lighting makes food look better -- that knowledge gives you language for things you previously could only point at.

But the gap doesn't close. Because taste, real taste, the kind that fires in four seconds before your conscious brain has even woken up, doesn't operate on principles. It operates on pattern recognition so deep and so fast that it bypasses articulation entirely. You've seen ten thousand restaurants. You've walked through a thousand doors. You've sat in countless rooms and noticed, without knowing you were noticing, which ones felt right and which ones felt wrong. And all of that accumulated experience gets compressed into an instant judgment that arrives as a feeling, not a sentence.

Taste arrives as a feeling, not a sentence. Four seconds. No words. Just: yes. Or: no. Explaining it takes fifteen years of vocabulary you might never fully find.

The frustration is real. I can walk into that restaurant and know. I can look at a website and know. I can see a brand identity and know. I can hear three seconds of a song and know. But "I just know" is not a thing you can put on a strategy slide. "It feels right" is not a rationale. "Trust me" is not a creative brief. And so you spend your career building bridges between the feeling and the explanation, and the bridges are useful, but they never quite reach the other side.

The brief-writing years

In my twenties and thirties, writing creative briefs at agencies across London, I got good at this translation exercise. Not perfect, but good. I could take a gut feeling about a brand -- this feels too corporate, this feels too try-hard, this feels like it's lost its way -- and break it down into components that other people could work with. The brand has lost its visual confidence. The tone of voice is trying to be younger than its audience. The colour palette suggests premium but the typography says high street. These are real observations, and they're useful, and they come from the same place as the four-second feeling. But they're after-the-fact rationalisations. The feeling came first. The explanation came later.

📋
The strategy director's confession: In fifteen years of writing creative briefs, the honest version of every brief I ever wrote was: "It should feel like THIS" followed by a gesture at something I couldn't quite describe. Everything else was scaffolding to get other people to the same feeling.

I used to think this was a limitation. That if I just learned more about design, about colour theory, about the formal principles of aesthetics, I'd eventually be able to fully articulate why something works. I'd close the gap. The feeling would become words and the words would be precise and everyone would understand.

I don't think that anymore. I think the gap is the point. The gap between feeling and explaining isn't a bug in how taste works. It's the feature. Taste is, by definition, the thing that operates faster than language. If you could fully articulate it, it wouldn't be taste. It would be a checklist. And checklists don't make restaurants feel right when you walk through the door.

Why AI can't get there

This is the thing that makes me optimistic about the future of taste in a world full of AI. And I say this as someone who builds with AI every single day. I've shipped fourteen products using Claude Code. I'm not a sceptic. I think AI is extraordinary. But it can't do this.

AI can analyse. It can tell you the hex codes in a colour palette. It can identify the typeface. It can measure the white space. It can compare a design to ten thousand other designs and tell you which ones it's most similar to. All of that is useful. None of it is taste.

AI can analyse every element of a design. It can't walk into a room and know in four seconds. That's the gap, and it's the most human gap there is.

Because taste isn't analysis. Taste is the thing that happens before analysis. It's the four-second judgment. The instant recognition. The feeling that this is right before you've identified a single specific reason why. It's the accumulated weight of every visual experience you've ever had, compressed into an instinct that fires faster than thought. AI can process information faster than a human. But it can't have experiences. It can't walk into a restaurant and feel the weight of the door handle and the warmth of the lighting and the proportion of the room and have all of that collapse into a wordless "yes." It can approximate that judgment through pattern matching, but approximation and feeling are different things.

The gap between taste and explanation -- the gap that frustrated me for fifteen years of brief-writing -- is actually the thing that keeps human judgment irreplaceable. Not because humans are better at analysis. We're not. But because taste isn't analysis. It's the thing that precedes analysis. It's the input, not the output. And you can't automate an input that lives in the body, in the accumulated experience of being a person who has walked through doors and sat in rooms and looked at things for forty years.

Learning to live in the gap

I've made peace with it. Not fully -- I still get frustrated when I can't explain why something works, especially when someone asks. But I've stopped seeing the inability to articulate taste as a failure of vocabulary. It's a feature of the thing itself. Taste is pre-verbal. It's supposed to be. The four seconds before the words arrive are where all the interesting stuff happens.

What I've started doing instead of trying to close the gap is learning to trust it. To use the four-second feeling as data rather than noise. When I'm building something and it doesn't feel right -- when the layout is technically correct but something is off -- I've stopped trying to identify the specific problem and started just responding to the feeling. Move this. Change that colour. Add space here. Remove that element. I can't always tell you why. But I can feel when it clicks.

That's what I mean when I talk about taste being the last unfair advantage. Not the ability to have opinions. Everyone has opinions. The ability to have fast, accurate, pre-verbal judgments about quality -- and the willingness to trust them even when you can't explain them. Even when the deck doesn't have a slide for "it just feels right." Even when the client wants a rationale and all you have is a feeling and twenty years of looking at things.

The taste you can't explain is still taste. It might be the only kind that matters.