Back when I used to commute home from a client’s office, I would call a friend almost every day to keep me company during the long drive. For the entire hour, we would chat about nothing in particular, just aimless conversation, until I reached my destination.
If you asked me what I saw during that drive, I wouldn’t be able to tell you. Not the road. Not the traffic. Not even the turns I took.
That is exactly how vibe coding feels.
Vibe coding pulls you into a flow where you are chasing momentum instead of clarity. You type quickly, bounce between tabs, paste code, see it compile, and feel productive. But you are not really understanding what is happening. You might get something working, but you are left unable to explain how it works or why it works. Just like that commute, it is all a blur.
Last year during my Christmas break, I traveled with my family to one of my favorite places in the world, Estes Park, Colorado. My plan was simple: wake up early, go for short hikes, and enjoy the beauty of the mountains.
But my family had other plans. They stayed up late and slept in, often not waking until well into the afternoon. I did not want to hike alone, so I passed the time with some light coding. I had no real goal, just vibe coding.
In under an hour, I finished three apps. They were basic, but I used ChatGPT to write most of the code. I pasted it into my editor, made sure it ran, and moved on.
One app involved a fairly complex distance calculation algorithm. It worked perfectly, but I had not written a single line of it. ChatGPT did. And while I had a general sense of what the code was doing, I could not explain the details.
It felt like progress, but it also felt empty.
These were just simple apps. But in an enterprise setting with a complex domain, relying on surface-level understanding is risky. If you do not take the time to learn the business logic, the domain constraints, and the reasoning behind key decisions, you are doing yourself and your team a disservice.
After nearly 20 years as a software developer, I have come to realize something important: domain knowledge often matters more than technical skill. You can always find someone who can animate a screen, wire up a form, or build a splash screen. But developers who understand the business deeply, including the edge cases, the workflows, and the reasoning behind the system, are rare.
When technical skill is combined with domain expertise, that is when real value is created. That is when you become a developer who cannot easily be replaced. That is when you become a unicorn.
There is nothing wrong with using tools like ChatGPT. They are great for boosting productivity, learning new techniques, and solving problems. But when they replace your thinking, your curiosity, and your willingness to dig deep, you stop growing.
Real progress comes from getting stuck. From debugging confusing errors. From understanding how and why something works. It is slower. It is frustrating. But it lasts.
ChatGPT and other AI tools are powerful assistants. But if you rely on them for every answer without engaging your own mind, you are missing the point.
If you do not understand the code you write, you cannot trust it. And if you cannot trust it, you cannot own it.
Momentum without mastery is not progress. It is just motion.
Do not confuse the two.