AI Is Writing the Code – But Are Juniors Still Learning?

Hello! It's been a while, a long while at that. I've been having conversations with a few colleagues and one of them stuck.

Something we've seen more and more of is junior engineers leaning heavily on AI tools to write code for them. Yes, AI is amazing (if used correctly). Yes, it can save time. But no, it’s not a magic wand that will turn you into a great developer overnight.

I’m not here to bash anyone. I’m here because I think this new mindset is holding juniors back.

Crutch, Not A Tool

“Why write it myself when ChatGPT can do it for me?”

I’ve heard this. I’ve seen this... But here's the problem: relying on AI without truly understanding what it's doing means you’re just copy pasting your way into confusion and most probably chaos. You get the code, it runs, you move on... but do you know why it works? Do you know what to do when it doesn't?

AI should support your learning, push you to understand concepts... Make it easier to grasp them with a so called "personal tutor", but not replace it.

Learning VS Skipping

Let me paint a picture: You’re new. You get a task. You feed it into AI. It spits out a nice looking function. You copy it, it compiles, you push. Cool, right?

Now fast forward a few weeks. A bug shows up in that same code. You’re staring at a screen, reading your own commit message, and realising you don’t even remember writing that logic. Because technically... you didn’t.

This is where things start to fall apart. Because programming isn’t just about writing code... It’s about understanding systems, debugging, improving, iterating. If you skip that learning curve, you’re skipping the core of the craft.

AI Is Not The Problem

Let me be clear... AI is a fantastic tool. I use it... You probably use it... And we should... to an extent.

But there’s a difference between:

“I wrote this logic, now let me ask AI how to clean it up a little,”

and

“I don’t know how to do this, let me ask AI to do all of it..." - while flooding private code to it HA!

Senior Engineers: Babysitting The AI

I’ve noticed a shift... PRs from junior devs used to be messy but honest!! You could see the thought process, the effort. Now? You get a pristine looking block of code that smells a little too clean. It’s syntactically correct, logically plausible, but you can’t shake the feeling: "Did they actually write this? Do they really know what's going on here?"

So you spend more time reading it. Line by line. Wondering if some obscure AI generated bug is waiting in there like a landmine.

And that time adds up.

Instead of fixing, working, you’re PR'ing some blind written machine code for errors (or security vulnerabilities) - Instead of working on more specific tasks.

You’re double-checking that they didn’t blindly paste in a solution that breaks something downstream. The safety net gets tighter, heavier, and more exhausting.

This isn’t sustainable.

Code review shouldn’t feel like reverse engineering ChatGPT.

That's all from me - just wanted to get my thoughts out there. Still no gin, though I've had a bottle of wine.

Rus