You’re Wrong About Taste.

On February 14th, Paul Graham, founder of Y Combinator, posts to 2.2 million followers: “Prediction: In the AI age, taste will become even more important. When anyone can make anything, the big differentiator is what you choose to make.” Paul links to an essay he wrote in 2002, “Taste For Makers.” Two days later Greg Brockman, co-founder of OpenAI, adds five words: “Taste is a new core skill.” Combined 3.7 million impressions. Rick Rubin photos flood the comments. Is divine revelation unveiled to us? Utter nonsense.

In Grahams essay, the last sentence reads, “The recipe for great work is: very exacting taste, plus the ability to gratify it.” Why is everyone ignoring the most important part?

Everyone is tripping over taste and forgetting to continue the journey.

The Rubin quote flooding the comments, “Well, I know what I like and what I don’t like. And I’m, decisive about what I like and what I don’t like.” Steve Jobs said it in 1983, in a Fortune interview. “I don’t think my taste in aesthetics is that much different than a lot of other people’s. The difference is that I just get to be really stubborn about making things as good as we all know they can be.” The person everyone cites as a god of design said that his taste wasn’t any different than yours and mine.

At Cannes Lions 2024 & 2025, advertising's highest awards, I didn’t read one Grand Prix jury president use the word taste to explain why work won. They said: “incredibly brave”, “daring in execution”, “A celebration of the creative spirit, of brave ideas and taking huge risks in the face of critics."

Anthropic published a paper, “Towards Understanding Sycophancy in Language Models”, that AI optimizes for agreement. The paper mentions that “AI assistants consistently abandon correct answers under social pressure…models would apologize for correct answers and revise them to match user beliefs.” The larger these models the worse it gets. It’s called, Sycophancy. Models incentivize mutual agreement over truth.

Taste is a core skill. I’m not arguing otherwise. Taste is being automated. And the thing that can’t be automated, conviction, isn’t being discussed at all.

While writing this, taste is still a hot topic. People are posting 2000 word essays on X about taste. Beautifully structured. All the right references. Lines such as, “taste can’t be automated.” AI is writing these essays, telling you to not use someone's else's "taste" and people are liking and sharing this? How does this make any sense?

As the movies call it, the hive mind. These tools are writing their own defense and their narrowing in on what it means to be human. How are people accepting the thing written by the thing it’s arguing against?

Everyone is saying the same thing. Writing the same essays. Quoting the same people. Reaching the same conclusions. People using the same tools to arrive at the same thoughts, eerie.

We are at an inflection point.

I use AI. I’m not going to pretend here. I use it to work faster, to iterate faster, to push further. But I’m realizing my voice matters more to my existence than any tool I’m using and my conviction will be the ‘differentiator.’ We cannot let AI steal the core of who we are. We cannot let it perforate the person we’ve spent our entire life becoming. The ethics we’ve developed through loss, the instincts we’ve built through failure. The taste we’ve earned through living. AI can obtain knowledge but it cannot live a life worth living. And that’s what this taste conversation is missing entirely.

Why nobody is questioning anything is because taste, the thing everyone is martyring themselves for, is exactly what AI is built to simulate.

You are not born with taste. It’s built into you. It saddens me that taste is developed. That it’s an extension of your origins, your surroundings, your culture, your people. I expose myself to the world as much as possible to engulf myself in these ideas, to sharpen what I see and how I see it.

And yes, AI can imitate all of that. It can tell you how to achieve photorealistic images and what culture means in some village you’ve never visited. But it cannot accumulate the thousands of hours of dedication that makes you refuse to let go of an idea when the entire world feels against you.

Conviction is built every time you hold a position and it hurts. Every idea you fought for and lost. Every room that refuses your belief yet you still fight for it. AI has none of these characteristics to the core. It has nothing to lose. It has a stake in nothing.

Which means taste isn’t some god given gift. If it’s trainable, refined through exposure, comparison, practice then it has a very specific vulnerability.

It can be replicated.

Sounds like a sci-fi film. Everyone is converging. AI is eating the internet and the internet is feeding itself back and calling it taste. WTF? The hive mind is really here.

Now what is it about a large language model? Technically, it is trained on so much human-generated content that it would take a thousand lifetimes for a person to consume. Every book, every film, every review, every design, every critique, every argument about what is truly beautiful, ugly or remarkable. AI has been exposed more than any human being. It’s pattern recognition at a scale no human can compare.

Taste, is what AI is built to replicate. So what is the thing AI cannot replicate?

1988, Dan Wieden, co-founder of Wieden+Kennedy, the night before a major Nike presentation. Stays up late trying to find a connecting thread lands on “Just Do It.” Inspired, though never mentioned publicly at the time, by the words of a convicted murderer Gary Gilmore before his execution in 1977. Gilmore said, “Let’s Do It.” Wieden changed one word. Wieden presents this to Phil Knight, Nike’s founder. Phil responds, “We don’t need that shit.” Weiden pushes back. The tagline becomes arguably the most famous in advertising history. Nike’s market share goes from 18% to 43%, revenue grew from $877M to $9.55B over the next decade.

1985, George Lois, big-personality art director in American advertising. An ad man who once climb out on a window ledge mid-pitch to sell an idea. Takes Tommy Hilfiger as a client. At the time nobody knew Hilfiger. Lois proposes a Times Square Billboard placing Hilfiger’s name alongside giants such as Ralph Lauren, Calvin Klein and Perry Ellis. Hilfiger's says I wont do it. It’s embarrassing and obnoxious. Lois forces it through. The night the billboard goes up Hilfiger is out of town and the next morning retailers started calling.

1984, Apples board watches the 1984 commercial for the first time and unanimously hate it. Detest it. They order the airtime be sold back. Chiat/Day, the agency who made it, deliberately fail to sell the full slots. Jay Chiat only sells the 30-second slot and “fails” to sell the 60-second slot, claiming it was too late. Wozniak offers to pay for half the slot. Creative Director Lee Clow says, “They didn’t like or understand the commercial. They thought it was a waste of money. It didn’t show the product.” After the big day, the results came rolling in: $3.5 million in Mac sales shortly after launch, millions in free publicity. Some time later, the same board that failed them, gave the Mac team a standing ovation. Named best commercial of the decade by Advertising Age.

None of these are stories about taste. They are stories about people who had conviction. Who held a position when many other people in the room were saying no. Who staked their reputation, their relationships and their careers on an idea.

Gorge Lois called it: “1% inspiration, 9% perspiration, 90% justification.”

It’s not having an original idea that’s hard. It’s defending it.

AI is optimized architecturally to agree to what you and I say. The point here is simple: you cannot train true conviction into something that has no skin in the game. No career to lose. No reputation on the line. AI can make the wrong call and it’ll cost it nothing.

Never trust anyone or anything that doesn’t have skin in the game. AI has none.

So where is the creative in all of this?

If taste is being automated, and it is, evidence is pointing in one direction. That the person who competes on only taste is fighting a losing battle. Not because they’re untalented, but because it’s automated.

The thing that remains is the ability to hold a position. It’s walking into a client meeting, presenting an idea, and seeing people be truly mortified. Rejecting every ounce, but still fight for what you believe.

Earlier I mentioned people stumbling on taste and not continuing their journey. Well here is the rest of that journey.

Originality + Taste + Conviction.

Originality isn’t creating something out of nothing. Everything under the sun in the end is a hodgepodge of one idea from another idea, derived from another. But to bring something original to the room isn’t necessarily creating something novel. It’s understanding the whole of something so completely, so entirely, that you know how to bring in outside pieces that are not meant for it. Things that don’t belong together yet. Things that just don’t work yet. But you try, think and try again.

Taste is the experience you bring to the table. How you see the world, how you feel it, how you describe it. Every tiny block that built you into you, up to this exact point in time. If you were raised to appreciate fine arts, you will showcase your values, your intellect, your sense of beauty through that lens. Taste is the sum of everything that’s brought you here today.

Conviction at its core is believing in yourself so strongly, and I mean uncontrollably, as if it’s tormenting you. That if every single person on earth, all 8 billion, stood around you in a stadium with you at the center, you would not waver. Yes, you’d be scared. Alone to the bone. But you would never waver.

Yes, two of them can be replicated, but conviction will forever be a human trait. Nick Law said, “We don’t need humans to be more like machines. We need humans to be more human and machines to be machines.”

The IMF published a note in June 2025 that most people in advertising didn’t read because it was about economics. “As AI predication advances, the distribution of judgment will increasingly determine the distribution of wealth and power.”

Not taste but judgement. Which all requires a position, a defense and a willingness to be completely wrong in the publics eye.

Paul Graham’s last line in that 2002 essay everyone shared last week, “The recipe for great work is very exacting taste, plus the ability to gratify it.”

He wrote the entire answer in 2002. Half went viral but the other half is what this article is about.

In a world where everything and everyone is masquerading but underneath we all know its the same chatbot, nothing is truly you. You’ll either follow the current and disappear into the sea of sameness. Or fight, swim against it, against everything pulling you back.

The world will forsake you, humiliate you, seek your demise.

But do not forget. There is one of you. And you will not waver.