I’m Becoming a Better Developer. I’m Just Not Sure How Much of It Is Me.

Lately, I’ve been feeling like I’m becoming a much better developer.

I’m handling things at work that would have scared me not too long ago. I can jump from a frontend issue to a backend API, then find myself checking the database, looking at server logs, figuring out a deployment issue, or trying to understand why something is behaving differently in production.

And somehow, I’m able to manage it.

Not always smoothly. There are still plenty of moments where I have absolutely no idea what’s going on. But I’ve become much more comfortable with that feeling. Instead of thinking, “I don’t know this,” and stopping there, I now think, “Okay, I don’t know this yet. Let me figure it out.”

AI has become a huge part of how I do that.

And that has made me think about something I honestly don’t have an answer to yet:

I know I’m becoming a better developer. I’m just not sure how much of it is me.

AI is just part of my work now

I use AI pretty much every day.

There’s no point pretending otherwise.

I use it when I’m debugging something weird. I use it when I’m working with a part of the codebase I don’t completely understand. I use it when I’m dealing with infrastructure, databases, APIs, deployments, or something I haven’t worked with much before.

Sometimes I’ll paste an error and ask what could be causing it. Other times I’ll explain the whole situation, what I’ve already tried, what I think is happening, and ask it to challenge my thinking.

And yes, sometimes I just ask really basic questions.

That’s one thing I love about AI. There’s no hesitation before asking something because I’m worried it might be a stupid question.

If I don’t understand something, I ask.

If the explanation still doesn’t make sense, I ask again.

“Explain this more simply.”

“Why do we actually need this?”

“What happens if I remove it?”

“Okay, but what is happening underneath?”

I can keep going until it finally clicks.

That has changed the way I learn.

I’m able to handle things much earlier than I expected

When I imagined my first years as a developer, I thought I’d mostly be given coding tasks.

Build this component. Fix this bug. Connect this API.

But real work became much bigger than that very quickly.

You build a feature, but then the API has an issue. You fix the API, but something isn’t right in the database. Then the deployment fails. Then something works locally but not in production. Then you’re looking at logs from a server you barely understood a few months ago.

At some point you realize that software engineering is much more than writing code.

And AI has made it easier for me to step into those unfamiliar areas.

Before, if I encountered something completely new, there was a bigger mental barrier. I’d think, “I’ve never worked with this before.”

Now I’m much more likely to just start.

I know I can read the documentation. I can inspect the code. I can ask AI questions. I can test things. I can break something locally, understand why it broke, fix it, and slowly build a mental picture of how everything works.

That confidence is real.

And I think that confidence has made me a better developer.

But then there are moments that make me question it

Sometimes I’ll be stuck on something for a while, ask AI, get a solution, try it…

and it works.

Amazing.

I move on to the next thing.

Then maybe ten minutes later I’ll think:

Wait. Why did that actually work?

That question bothers me more than the original bug.

Because fixing something and understanding something are not the same thing.

If AI tells me to change three lines and the problem disappears, technically I solved the task. The ticket can move. The feature works. Everyone is happy.

But did I learn anything?

If the same problem happens six months from now, would I recognize it?

Or would I open AI and ask the exact same question again?

That’s where I think the danger is for me.

Not that AI writes code.

Not that I use it too much.

The danger is that sometimes it becomes so easy to get the answer that I forget to understand the answer.

At the same time, I know I’m learning

And this is where things get confusing.

Because I definitely am learning.

A lot.

In fact, I think AI has accelerated my learning massively.

I’m touching areas of software engineering that I might have avoided before because they felt too complicated. When I don’t understand something, I can go deep immediately instead of bookmarking a 40-minute tutorial and telling myself I’ll watch it later.

I can ask why a database query is behaving a certain way. Then ask what an index actually changes. Then ask why adding indexes everywhere is a bad idea. Then connect that explanation back to the problem I’m actually working on.

That entire learning loop can happen while I’m solving a real problem. And that matters.

I’m not learning something because a tutorial told me I might need it someday. I’m learning it because something is broken in front of me and I need to understand it.

Those lessons stick differently.

I’ve had plenty of situations where AI helped me understand something once, and weeks later I encountered a similar problem and immediately knew where to look.

That knowledge became mine.

So when I ask myself, “Am I actually learning?”

The answer is clearly yes.

The harder question is whether I’m learning as deeply as I think I am.

I also write code differently now

This is probably the biggest change.

I don’t care as much about writing every single line myself anymore.

A couple of years ago, I probably would have felt weird saying that.

Now it seems normal.

Sometimes I know exactly what I want and let AI generate the boring part. Sometimes I ask it for a starting point and rewrite half of what it gives me. Sometimes it gives me a solution that looks great initially, but once I think through the edge cases, I realize it doesn’t fit the system at all.

And sometimes it gives me complete garbage.

That happens too.

The interesting part is that I’ve started thinking less about “Can I write this code?” and more about “Is this the right way to build this?”

Those are very different questions.

AI is incredibly good at producing code that looks convincing. That doesn’t mean the code belongs in production.

It doesn’t know every decision that was made before. It doesn’t always understand the business logic. It can miss security implications. It can create unnecessary complexity while confidently explaining why the complexity is necessary.

So even when AI writes the code, I’m still responsible for understanding what is going into the system.

If that code breaks at 2 PM on a working day, I can’t exactly tell everyone, “Well, AI wrote it.”

It’s my code once I accept it.

That realization has actually made me think more about architecture, maintainability and how different parts of a system affect each other.

AI has also made me more ambitious

This might be the part I appreciate the most.

I build more things now.

When I get an idea, I don’t immediately think about everything I don’t know.

I think, “I can probably figure this out.”

That is a huge shift.

There are things I’ve built, technologies I’ve explored, and problems I’ve attempted simply because the cost of being clueless at the beginning has become much lower.

AI doesn’t remove the difficulty. But it gives me somewhere to start.

And once I start, I usually learn something.

Maybe that’s why I find the whole “AI is making developers lazy” argument too simple.

AI can absolutely make me lazy.

I’ve caught myself accepting an answer because it worked and moving on without understanding it.

But AI can also make me ridiculously curious.

The same tool that can let me avoid thinking can also let me ask ten questions about something I would never have had the patience to research before.

The difference is how I use it.

I don’t want to stop using AI

That’s definitely not the conclusion I’m reaching here.

I’m not going to suddenly stop using AI and start writing everything from scratch just to prove that I can.

That feels pointless.

The way we build software is changing, and I want to change with it.

But I also don’t want to wake up a few years from now and realize that I became extremely good at producing software while becoming worse at understanding it.

That’s the balance I’m trying to find.

Use AI to move faster, but slow down when something matters.

Let it generate code, but read the code.

Ask it to fix the bug, but understand what caused the bug.

Use the answer, but ask why.

And sometimes, struggle with the problem for a little while before asking for help.

Because maybe some of that struggle is still important.

I’m definitely a better developer today than I was before I started using AI this heavily. I understand more, I’m comfortable taking on bigger problems, and I’m much less afraid of touching things outside what I already know.

AI deserves some credit for that.

But I also think I do.

Because AI can give me answers, but I’m still the one showing up every day, making mistakes, debugging them, asking questions, shipping things, breaking things and slowly understanding how all of it fits together.

Maybe becoming a good developer was never about knowing everything yourself.

Maybe it was always about being able to figure things out.

AI has just changed how quickly we can do that.

And right now, I’m okay with not knowing exactly how much of my growth is me and how much is AI.

I just want to make sure that as the tools keep getting better,

I’m getting better too.