I'm currently working alongside an AI agency on a project.

And I'm watching something happen in real time that I think we're going to see A LOT more of.

AI has allowed an entry-level developer to take on what is really a senior-level architecture problem.

And technically...

they can build it.

That's the crazy part.

But maintaining it is a completely different story.

The best way I can describe what I'm watching is a kid trying to stack slippery blocks.

One issue appears.

They fix it.

Move to the next issue.

Then one of the previous blocks falls.

They fix that again.

Another block falls somewhere else.

And it just keeps going.

Not because AI can't write the code.

It can.

The problem is that experience teaches you to recognize problems you haven't encountered yet.

Architecture.

Dependencies.

Failure points.

Edge cases.

The things an entry-level developer doesn't even know they should be looking for.

I'm using AI on my side of the exact same project.

The difference is I'm using it to accelerate work I already knew how to do.

That's where I think people are misunderstanding AI.

AI can dramatically increase what you're capable of building.

But it can also dramatically increase the complexity you're capable of creating.

And if you don't have the experience to recognize that complexity...

you can vibe-code yourself into one hell of a mess.