When AI entered the programming world, I was excited. It promised faster development, less boring work, and better productivity. And honestly, it delivered — at least on the surface.
But after watching how teams actually use AI, I’ve started to feel uneasy.
What we now call “vibe coding” works great for demos. You describe what you want, the code runs, and everyone moves on. But real software doesn’t stop there. It scales. It changes. And eventually, someone has to maintain it. That’s where things begin to break.
A lot of AI-generated code works, but no one really understands how. Logic gets repeated. Abstractions stay shallow. Decisions are hidden behind prompts instead of being made explicitly in code. Over time, this creates a slop layer — software that functions today but becomes fragile tomorrow.
For junior developers, this is especially worrying. AI helps them finish tasks faster, but it also removes the struggle that used to teach us how systems actually work. Debugging, reading ugly code, fixing small bugs — those were the moments where learning happened. Now, many juniors are expected to handle advanced tasks from day one, without ever building that foundation.
Even senior developers aren’t immune. When answers are always one prompt away, there’s less incentive to explore deeply, read documentation, or experiment. The craft slowly turns into supervision — approving outputs instead of understanding systems.
The job market reflects this shift. Junior roles are disappearing. Senior roles are being paid less. Companies are increasingly comfortable replacing experience with a mix of AI tools and cheaper talent. Some are already regretting it, but much of the damage is happening quietly.
AI was never the problem.
The problem is using it as a replacement instead of a helper. Writing code in English isn’t wrong. Shipping code you don’t understand is. If we keep choosing speed over understanding, we’ll ship more software — and learn less about it every year.
And that’s the real question:
When AI writes the code, who learns the craft?