In the winter of 1770, an inventor named Wolfgang von Kempelen rolled a wooden cabinet into the court of Empress Maria Theresa. On top of the cabinet sat a life sized wooden figure dressed in traditional Ottoman robes, complete with a turban and a long pipe. This was The Mechanical Turk. It was marketed as the world’s first chess playing automaton. It defeated Benjamin Franklin. It defeated Napoleon Bonaparte. For nearly eighty years, people across Europe and America marveled at its incredible logic.
There was only one problem. The Turk was a lie.
Inside the cabinet, hidden behind a series of deceptive gears and sliding panels, sat a very small, very talented human chess master. The machine wasn't thinking. It was merely a vessel for human intelligence wrapped in the illusion of clockwork.
As we navigate 2026, the parallels are impossible to ignore. We are surrounded by digital cabinets. We call them Large Language Models. We call them Generative AI. We treat them as if they are thinking, feeling entities. But when you peel back the layers of hype and stock pumping, you realize we are often just looking at a much faster, much more expensive version of Kempelen’s hoax. We have created a machine that can calculate, but we have yet to build one that can actually feel.
The Great Divide: Math vs. Magic
To understand why AI is currently hitting a wall, we have to distinguish between two very different types of intelligence.
The first is Objective Intelligence. This is the world of math, logic, and verifiable facts. Machines are brilliant at this. If you need to calculate the trajectory of a rocket or search through a billion lines of legal code to find a specific clause, the machine wins every time. It is a calculator on steroids. There is no debate about whether two plus two equals four. There is a right answer and a wrong answer.
Then there is Subjective Intelligence. This is the world of taste, charisma, and lived experience. It is the "je ne sais quoi" that makes a brand iconic or a piece of music move you to tears. This is the world where there is no "right" answer, only a "felt" one.
The machine can mimic the patterns of taste. It can look at a million Pinterest boards and tell you that "mid century modern" is popular. But it doesn't know why it’s popular. It doesn't understand the nostalgia for a specific era of optimism. It doesn't understand the warmth of walnut wood against a cold floor. It just knows that the word "walnut" often appears next to the word "aesthetic."

The Strawberry Test and the Illusion of Logic
For a long time, the tech giants told us that these models were learning to reason. Then came the Strawberry Test.
If you ask some of the most advanced models in the world how many "r"s are in the word "Strawberry," they will frequently look you in the digital eye and tell you there are two. This isn't a glitch. It is a fundamental revelation of how the technology works.
AI does not "see" the word strawberry. It doesn't understand the concept of a red fruit or the letters that form its name. It processes "tokens." It is essentially a giant game of "Predict the Next Word." It predicts that "two" is a statistically likely answer because of how it was trained or how its tokenization system breaks down the word.
This is the ultimate proof that the Turk has no brain. It is predicting, not reasoning. If a machine cannot count the letters in a common fruit, why are we trusting it to handle the nuanced "taste" required to build a brand or lead a creative project? The "r" in strawberry is a tiny crack in the glass that shows the entire structure is built on probability, not understanding.
The Rise of AI Washing
In 2026, we are seeing a massive wave of what experts call AI washing. This is the corporate practice of taking a standard automated process and slapping an "AI Powered" sticker on it to pump stock prices or justify mass layoffs.
We saw the precursor to this with Amazon’s "Just Walk Out" technology. It was marketed as a miracle of computer vision and AI sensors that allowed shoppers to grab items and leave without a cashier. The reality? It was reportedly powered by over one thousand workers in India who were manually watching camera feeds to see what people put in their carts.
The Mechanical Turk lives on.
Today, companies like Anthropic and others are pushing the "Superintelligence" narrative. They want you to believe that the machine is just one more update away from becoming a god. Why? Because that narrative justifies the five trillion dollar chip factories and the massive energy consumption. It distracts from the fact that in real world economic tasks, AI still fails about 96 percent of the time.
When a task requires a tiny bit of human judgment, the machine stumbles. When a task requires an understanding of social context, the machine produces noise.

Why AI Produces Noise, Not Art
The biggest problem with relying on AI for "taste" is that the models are trained on the average. They are trained on the entire internet. And if you have spent any time on the internet lately, you know that the "average" is pretty mediocre.
When you ask an AI to design a logo or write a blog post, it looks at everything that has already been done and gives you the middle point. It gives you the most probable result. This is why so much AI generated content feels like "filler." It is technically correct but emotionally vacant.
Taste is about the "grain." It is about the specific, the weird, and the unconventional. True taste often goes against the grain of what is popular. It is the ability to say, "I know everyone is doing X, but Y feels more authentic right now."
A data center cannot replicate a childhood memory of a summer afternoon. it cannot replicate the feeling of a first heartbreak or the specific charisma of a leader who can rally a room with a look. Because the AI has no lived experience, its version of taste is just a mask. It is a photocopy of a photocopy. It is high volume noise that lacks the soul of human judgment.
The Economic Failure of the 96 Percent
We are currently in a strange economic moment. Corporations are firing thousands of creative professionals and replacing them with prompt engineers. They are betting that the "good enough" output of the machine will be enough to sustain their growth.
But the data is starting to show a different story. In tasks that require high level creative problem solving or complex project management, AI success rates are abysmal. It can write a basic email, but it can't navigate the political nuances of a merger. It can generate a generic image, but it can't create a visual language that defines a generation.
Leaders who are navigating this growth period need to be careful. If you outsource your taste to a machine, you are effectively deciding to be average. You are choosing to join the sea of noise. In a world where everyone has access to the same "average" generator, the only thing that will actually hold value is the human "je ne sais quoi."

The Human Edge in 2026
The real winners of this era won't be the people who use AI to replace their thinking. They will be the people who use AI as a tool for the "Objective" and keep the "Subjective" strictly human.
Use the machine to crunch the numbers. Use it to organize your calendar. Use it to find that one specific line of code that is causing a bug. But do not let it tell you what is beautiful. Do not let it tell you what is funny. And definitely do not let it tell you what your brand should stand for.
Taste is a muscle. It is developed through years of looking at art, traveling to new places, talking to people who disagree with you, and paying attention to the world around you. It is a biological process that requires a heart and a nervous system.
The Turk in the machine is still just a man in a box. The "AI" we see today is a reflection of us, but it is a flattened, smoothed out reflection. It has no edges. And in the world of creation, the edges are where the magic happens.
We don't need more noise. We don't need more filler. We need leaders who aren't afraid to trust their gut over a probability score. We need creators who realize that their lived experience is the most valuable asset they own.
The machine can play chess. It can even win. But it will never understand the thrill of the game. It will never understand why we play in the first place. That part is ours. And as long as we hold onto that, the Turk will always just be a wooden figure in a turban, waiting for a human to tell it what to do next.


