Published: 2026-03-04 02:04

For a while now, I've been trying to understand the vibe behind the big push for "AI that writes all the code."

Not "AI helps you autocomplete a function." Not "AI explains a stack trace." I mean the full-on narrative:

"We won't need software engineers anymore."

And I couldn't quite name what bothered me about it—until it clicked.

Software engineers can do wizardry with words

Good software engineers have this strange, unfair advantage: we can make a computer do what we want using language.

Sure, it's a weird language. It looks archaic. Sometimes it's hostile. Sometimes it's beautiful.

But still—if you know what you're doing—you can sit down with a keyboard and turn words into:

• a product
• a workflow
• an automated business process
• a system that makes money while you sleep
• a tool that saves a team thousands of hours

That's real power. It's leverage.

And when people say "AI will replace software engineers," I don't think that's always a technical prediction.

I think sometimes it's a reaction to that leverage.

Part of this is jealousy (yes, I said it)

There's a specific kind of frustration that comes from watching someone do something you can't do—even if you understand it in theory.

Non-engineers can absolutely build things. They can be brilliant. They can run companies. They can design. They can sell. They can lead.

But software engineering has this unique "I can conjure a machine to do my bidding" quality to it.

And if you don't have that skill, the gap feels like:

• you're dependent on someone else to execute
• you can't move at the speed you want
• you have to ask for changes instead of just making them

So yeah… I think there's a jealousy factor sometimes. Not petty jealousy. More like:

"Why do they get to be the ones who can speak to the machine?"

But the real driver is simpler: bottlenecks

Even if you remove the psychology entirely, there's a very practical reason for the "replace the engineers" story:

Software engineers are a bottleneck.

Not because we're lazy. Not because we're gatekeeping. Because building real systems is hard, and the number of people who can reliably do it is limited.

If you're a company and you want to grow, software becomes the multiplier. And engineers are the ones holding the multiplier.

So of course everyone is trying to compress that constraint. Of course investors love the idea. Of course founders love the idea. Of course managers love the idea.

"AI writes the code" is basically the dream of infinite leverage without the human constraint.

What happens next: the gap gets bigger, not smaller

Here's the part I think a lot of people miss:

Even if AI gets dramatically better at generating code, that doesn't automatically flatten the playing field.

It can also widen it.

Because the best software engineers will use AI the way power tools get used by a master carpenter:

• faster iterations
• less time spent on boilerplate
• more time spent on architecture and tradeoffs
• more shots on goal
• more finished projects

The output increases, but the taste still matters. The judgment still matters. The ability to debug reality still matters.

And the people who are passionate about the craft—who love building—are going to use these tools to become terrifyingly effective.

Meanwhile, the folks who are in it purely for the paycheck (or who never really learned the fundamentals) are the ones most at risk. Not because AI is "replacing engineers," but because AI makes it easier to detect who can't actually drive the thing.

The outcome I'm betting on

I think the world is going to realize something boring and true:

Software engineering isn't going anywhere.

We're just changing what "doing the job" looks like.

The job shifts upward. The tools get better. The expectations increase. The baseline competence rises.

And then you get this weird moment where people look around and go:

"Huh. We didn't eliminate software engineers. We just made the good ones even better."

That's the chasm. And it's going to be uncomfortable.

But it's also an opportunity.

If you're an engineer, here's the play

Use the tools. Don't worship them. Don't ignore them.

Let AI handle the stuff you shouldn't be spending your life on:

• repetitive scaffolding
• "write the boring version of this"
• first drafts
• searching for the right API calls
• turning a concept into a starting point

Then do what engineers have always done:

• make it correct
• make it simple
• make it maintainable
• make it real

Because the magic isn't "typing code."

The magic is turning messy human intent into something a computer can execute—reliably—in the real world.

And that part still needs you.