The Creativity Paradox of AI
I see AI illiteracy everywhere.
I see people convinced AI will replace everyone. I see people getting scared. I see people that are way too enthusiastic about AI as well as people refusing to touch it. Some people ask loads of questions, and some ask none.
There is one thing that we all have in common - nobody knows what the future looks like.
Not the CEOs. Not the AI companies. Not the governments.
The only certainty is that change is coming quickly.
This change is going to be faster and more disruptive than anything humanity has experienced so far. OK, maybe not more disruptive than discovering fire. But who knows how many people got burnt before they learned how to use it?
This gave me an idea to create a little passion project fully built with Claude Code - a webapp that targets the most common questions and fears about AI and leaves you with something tangible to work with instead of endlessly debating abstract futures.
Check out the “Still Yours” app here:
Please share it with anyone who can benefit from knowing a little bit more about AI, or just to spark a conversation. Show it to your friends, parents, grandparents, partners, kids.
I am not a software engineer. But as a PM I have a solid understanding of the tech behind the scenes.
Many people have this fantasy:
- AI writes code.
- Therefore anyone can build software.
- Therefore building software is now easy.
And yes, shipping something now is accessible to anyone - you, me, tech people, non-tech people, artsy people, old, young, you name it. Basically anyone who has enough patience. Technically, this is true. Practically, there is a catch.
I discovered that AI can build incredibly fast. But was speed the real bottleneck before? For years we have treated execution as the scarce resource.
Need a website? Find a developer.
Need illustrations? Find a designer.
Need copy? Find a writer.
Need research? Spend days collecting information.
AI dramatically reduces the cost of execution.
What it does not reduce is the cost of deciding what should exist in the first place.
When almost anything can be built, choosing what is worth building becomes the harder problem.
Claude Code, Lovable, Gemini, ChatGPT - they can all ship. They can all generate solutions. But they cannot read your mind.
And that is where I ran into an unexpected problem: knowing exactly what I wanted did not make AI easier to use. It made it harder.
Yes, if you are asking questions, and doing research - knowing more means giving better questions, digging into more detail, and capturing any gaps in AI’s thinking.
However, creation from scratch is another story.
I started wondering whether the real competitive advantage in the AI era is no longer technical skill, but taste.
Not taste in the artistic sense.
Taste as the ability to recognize the difference between “good enough” and “this is exactly what I meant.”
AI can generate a thousand versions.
But it cannot tell you which one aligns with your intention.
That responsibility still belongs to the human behind the keyboard.
People imagine experts benefit less from AI because they can already do the work. My experience suggests the exact opposite. I am a perfectionist. I am critical. And I was getting very frustrated by not being able to fulfil my vision at a faster pace.
As a non-programmer, I often felt that if I wanted the product to match my standard exactly, I might as well learn to code it myself. The number of times I had to point Claude Code to the correct piece of code while debugging was surprising.
At one point I caught myself doing something I did not expect.
I was no longer asking AI to create.
I was reviewing.
Reviewing designs.
Reviewing text.
Reviewing logic.
Reviewing assumptions.
The role slowly shifted from maker to editor.
And I am not entirely sure whether we talk enough about what that shift means.
Someone with no vision may happily accept version 1. A person with a strong idea may spend 50 iterations getting there, while still not matching the idea in her head.
Why?
Because AI generates probabilities.
Vision requires intention.
This got me thinking. What is the trade-off between “everyone can ship, fast” and our own creativity being replaced by predictable (and predicted) outputs of AI, all gravitating towards mediocrity?
If AI generates the average of what already exists, and most people accept the first decent answer - what happens to creativity? What happens to craftsmanship? Will it become the burial ground of ideas that existed in someone’s mind but were never quite reproduced by AI?
But on a more positive note, this is probably not happening the first time in history either.
The calculator did not eliminate mathematics.
Photography did not eliminate painting.
Digital music did not eliminate musicians.
Every major technology changes what humans spend their effort on.
The question is whether AI is simply shifting the effort, or whether it is quietly changing our standards as well.
If most outputs become “good enough”, do we slowly lose our appetite for excellence?
Or does excellence become even more valuable because it stands out more clearly against a sea of acceptable results?
Maybe AI literacy is not really about learning prompts.
Maybe it is about learning yourself.
Understanding what you value.
What quality looks like to you.
What compromises you are willing to make.
And which ones you are not.
Because the easier it becomes to generate answers, the more important it becomes to know which answer you are actually looking for.
With love,
Zuzana
[
](https://substackcdn.com/image/fetch/%24s_!Eay-!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fa317271e-aada-41e0-9577-202113d42603_3000x3000.png) **Buy me a coffee
I drank 4 coffees while writing this post - support my habit by scanning the QR code/click on the link :)
**Fear of AI? Try this.
https://stillyours.couragefor.pm/
**Need a fresh perspective?
Visit my personal website - looking forward to working together!