AI makes ten. A person picks one. Making gets cheap; choosing gets rare. The eye that tells good from great — a curator's sense — is the real skill of the AI era.
When you work with AI, you've probably lived this moment. You ask for one image, and ten of them show up. One has beautiful light; one has good composition; one has uncanny color. Which do you pick? Most people freeze right here. This is the moment you realize choosing is harder than making. Why? The answer is in a surprisingly old principle: when making gets cheap, choosing gets rare.
This essay explains that principle from start to finish. If you've never done creative or curatorial work, you can still follow along — we'll go slowly. Today's example is ChatGPT and image generators, but the principle applies to any tool, in any era. Whatever AI comes next, the spine of this essay still holds.
Start with the principle. Every line of human work has a scarcity at its center. When something becomes easy, its value drops — and something else becomes valuable instead. Scarcity never disappears. It just moves.
Think back. There was a time when "a person who could take photos" was respected. Then smartphones arrived and everyone could shoot. From that point on, who was truly valued? The person who could pick one great photo out of thousands. The real skill of the Instagram era wasn't shooting. It was editing. In other words — choosing.
AI is heading down the same road. Algorithms can already do about 80% of your work — writing, images, code, design, translation. Making has become cheap and fast. So what's the remaining 20%? Choosing. And everything rides on that 20%.
The common sense blurs only in front of AI. Many people get excited — "I can draw now," "I can write now." That's only half-true. The real question is "can you see which drawing is good?", "can you tell which sentence is alive?" Making and choosing are different skills.
Why does the sense blur? Because everything AI produces looks plausible. All ten images are pretty. All ten sentences make sense. So you feel like "it doesn't matter which one I pick." It does. Of those ten, nine are ordinary and one survives time. The eye that finds that one — that's the real skill of this era.
Let me give you a concrete scene. Imagine someone who loves music. In the past, they had to sing, play instruments, and mix. Not anymore. There's an AI that makes music in their style — just one prompt line.
Love drawing? There's an AI that draws for you. Love architecture? There's an AI that designs buildings for you. Making is all AI now. The key is this: you become the center. Your taste, your style, your eye becomes the filter. AI pours out raw material, and you pick "this is mine."
ChatGPT crossed 100 million monthly users in two months — the fastest service to reach that mark in history. The number says one thing: the tools of making are now in everyone's hands. So the next question follows naturally. In a world where everyone can make, who survives? The one who chooses well.
The easiest analogy is a traditional market. Picture a stall with 100 fish. A sharp-eyed homemaker walks up, and within 3 minutes picks the one fish that's perfect for tonight's grill. From the outside it looks like "they're just picking randomly" — but in those 3 minutes, the eyes are reading dozens of signals: the color of the gills, the clarity of the eye, the firmness of the body, the edge of the smell.
The market pours out raw material. The homemaker is the chooser. The eye that tells good fish from great fish — that's the skill. The AI era has the exact same structure. AI is the market; you're the homemaker.
Let me show the numbers. One run of an image AI gives you 4 pictures. Five runs is 20. A full day's work easily stacks to 100. Of those 100, how many end up in your portfolio? Usually 1-2. One out of a hundred. In ratio — 1%.
Here's the first aha moment.
Making got 100× faster, but choosing got 100× harder.
When output multiplies by 100, the burden of choosing multiplies by 100. If you can't choose, you sit on 100 images and cry. If you can, you throw away 99 and complete a piece with 1. AI won't replace you. You'll be replaced by people who use AI. And people who use AI — in the end — are people who choose well.
So how do you train the curator's eye? It's simple. Every time AI hands you a result, ask yourself one question.
"Which of these looks most like the me I know?"
It isn't asking "good or bad." It's asking "mine or not mine." Taste is trainable. Look at 100 outputs a day for a month, and in a month you'll pick a better 1 out of the same 100.
| The old skill | The AI-era skill |
|---|---|
| Knowing a lot | Having a clear taste |
| Memorizing fast | Throwing away fast |
| Arguing logically | Choosing creatively |
All three rows share the same core. The past rewarded people who accumulate. Now it rewards people who filter.
Here's the training routine. Don't memorize — just keep three things.
1. Throw away 10 a day — of every AI result, delete 9
2. Write one reason — beside the one you kept, one line on "why I chose this"
3. Re-read your portfolio weekly — chosen things stack into a taste
On day one, your reason ends at "it's just pretty." After a week it becomes "the light comes from the left," "the subject's gaze points off-frame." After a month, those reasons fuse into your own style sentence. That's how you become a curator.
Let's wrap up.
AI made making cheap. Writing, drawing, coding, composing — all of it happens in seconds. So where did the scarcity go? It moved to choosing. In an era when AI pours out 100 images, the eye that picks 1 of them is your real skill.
We all have to become curators. This isn't only an artist's story. Writers, journalists, programmers, art directors, designers, architects, illustrators — every creative field is shifting into the same shape. And underneath it, there has to be a dream. You need to know what you want to make. Only then can you find yours among 100.
Three years from now, the AI's name will be different. The principle in this essay still works. When making gets easier, choosing gets harder. And people who choose well have always been rare. Tech changes. Principles don't.
Throw away well. Choose well. Stack well.