Fraud
There is a strange, unsettling feeling that comes with building something you love, something complex and beautiful, with a tool you barely understand. I have a piece of software now, a project more ambitious than anything I’ve ever created. It works. It has a clean architecture, a thoughtful user interface, and a core feature that feels like a small piece of magic. And yet, when I look at it, the pride I feel is tangled up with a quiet, persistent sense of being a fraud.
This project was built in partnership with an AI. I didn’t write the thousands of lines of code that make it run, at least not in the traditional sense. I didn’t wrestle with the arcane syntax of Python or configure the database with the muscle memory of a seasoned developer. I was the director, the architect, the questioner. My partner, the AI, was the brilliant, tireless apprentice that knew how to do everything I asked of it, instantly. I would describe a goal, and it would generate the blueprint. I would sketch out a logical flow, and it would lay the foundation. We’d go back and forth, a dialogue of intention and implementation, until the thing on the screen matched the vision in my head.
But it’s fun…
The process is intoxicating. It’s a creative loop on hyper-speed. It allows for a focus on the why—the user experience, the system’s logic, the overall vision—while the AI handles the how. But this very separation is the source of the conflict. The struggle that has historically defined craftsmanship—the years spent learning the grain of the wood, the feel of the chisel—has been abstracted away. And with it, I wonder, has the legitimacy of the craft been abstracted away as well?
It feels like I’ve stumbled into a room I always wished I could enter, but I used a key I didn’t earn. I love the view from the inside, but I’m terrified someone will ask me how I picked the lock.
This feeling, this modern imposter syndrome, is a lonely one. I’ve tried to explain it. I list the technologies I’m now commanding—Git, APIs, Python, cloud servers—and I see eyes glaze over. The vocabulary of the tools doesn’t capture the essence of the work. The real story isn’t about the technology; it’s about the strange new relationship with creativity itself. It’s about learning a skill you can’t quite put your finger on, a “meta-skill” of directing intelligence. And when you’re learning something so new it doesn’t have a proper name, who do you share it with? I’ve joked that someone should start a support group for self-loathing AI developers. It’s only half a joke.
Just a button pusher…
I’m beginning to suspect this isn’t the first time a generation of creators has felt this way. The feeling is familiar, I think. I imagine a young musician in the late 1990s, sitting in their bedroom surrounded by the glowing lights of a digital audio workstation (DAW). They’ve just finished a track. It has a beat that’s impossibly tight, a bassline deeper than any real instrument could produce, and layers of synthesizers creating a texture that’s entirely new. They didn’t spend ten years learning the guitar. They didn’t hire a session drummer. They clicked, they dragged, they tweaked virtual knobs and sliced up samples.
What did that bedroom producer feel? Did they feel the same unsettling conflict? The pride of creation mixed with the fear of being seen as a fraud by the “real musicians” who bled onto their fretboards? The traditionalist would say they were just pushing buttons, that the machine did all the work. But the producer knew better. They knew the hours they’d spent curating the right kick drum sample, the taste it took to arrange the parts into an emotional journey, the vision required to see the finished song in a sea of digital tools. Their craft wasn’t in playing the instrument; it was in the arrangement, the texture, the taste.
They proved that a new tool doesn’t invalidate artistry; it just redefines the artist’s role. They created entire genres of music that a traditional band never could. They were the pioneers of a new creative process, and I’m sure they felt just as isolated and conflicted as I do now. They were on their own solo ride, figuring it out.
That feeling of being on a solo ride resonates more than anything. I’m a cyclist, and I love the simple act of being on my bike. I enjoy the occasional ride with others, but I’ve always been wary of the cycling club. The club, it seems, is often less about the shared love of the ride and more about a subtle, competitive game. It’s about who has the lighter bike, the more expensive gear, the faster time up the hill. It can be a minefield of comparison and one-upmanship. The joy of the craft gets lost in the noise of the culture.
And so, ninety-five percent of the time, I’m on a quiet country road, alone. Not because I’m antisocial, but because I’m trying to protect the purity of the experience.
I wonder if this isn’t the same reason I feel so hesitant to share my software project. I love the idea of belonging to the developer culture. It seems cool, and I’m drawn to it. I’ve even put the requisite stickers on my laptop. But I’m wary of the club. I’ve seen the online forums, the technical debates that feel more like status battles, the gatekeeping around what constitutes a “real” developer. The fear isn’t just that someone will ask me a technical Python question I can’t answer. It’s the fear of the reason for the question—that it might not be a genuine inquiry, but a subtle test to see if I belong, a way to start the game of one-upmanship.
Perhaps the feeling isn’t anxiety, then, but a deep-seated aversion to that game. Perhaps the withdrawal isn’t a sign of imposter syndrome, but a conscious act of self-preservation. It’s an attempt to stay on that quiet country road where I can focus on the work itself, free from the noise of the peloton.
So, what is the work?
If it’s not typing syntax, what is this craft I’m practicing on my solo ride? It feels more like art than science. It’s a collection of skills that are harder to measure but, I suspect, are becoming more valuable.
The first is Taste. In a world of infinite AI-generated content, the ability to curate, to know what is good, what is elegant, what is useful, becomes paramount. An AI can generate a hundred user interface layouts, but it can’t tell you which one feels right. That requires a human sensibility.
The second is Problem Finding. An AI is a brilliant problem solver. But it can only solve the problems you give it. The higher-order skill is identifying the right problem to solve in the first place—seeing the subtle flaw in your own creation and having the vision to articulate a better path forward.
The third is Systems Thinking. It’s the ability to see the entire project as a cohesive whole, to understand how all the parts connect, and to direct the construction of a complex system without getting lost in the weeds of every single component. It’s the work of the architect, not the bricklayer.
This is the meta-skill, I think. It’s the art of directing a complex, intelligent system to produce a result that aligns with a specific vision. The process is a dialogue, a partnership. It’s mentoring a brilliant apprentice, asking it “why?” to deepen your own understanding, and pushing it to do better work.
I don’t have any grand conclusions. This new frontier doesn’t have maps yet. To feel conflicted and unsettled seems like the only rational response to being one of the first people to walk it. The pride is real because the creation is real. The doubt is real because the process is new and lacks a shared language or validation.
Maybe the goal isn’t to resolve the conflict. Maybe it’s just to get comfortable with it. To accept that this new kind of work comes with a new kind of feeling. It’s the quiet satisfaction of the solo ride on a country road. It’s not about the race or the club. It’s about the journey, the focus, and the simple, profound act of moving forward, alone, into a landscape no one has ever seen before. The code is the artifact; the learning is the reward. And for now, perhaps that’s enough.