Building Through the Noise

Every Tuesday feels like the end of software engineering

Wake up.

Open X.

”New AI model beats every benchmark.”

Refresh.

”Solo founder builds a unicorn in 48 hours with AI.”

Refresh again.

”Software engineers are cooked.”

If you’ve been on the internet long enough, you’ve probably seen all three before breakfast.

I laugh now.

Not because the technology isn’t impressive.

Because I’ve been hearing some version of this story for years.

Every new release feels like the one that’s finally going to replace us.

Every benchmark feels like proof that humans are becoming optional.

And if you’re anything like me, you’ve probably caught yourself wondering:

“Should I even bother mastering this anymore?”

It’s a fair question.

But after spending the last year building with AI instead of simply reading about it, I don’t think we’re asking the right question.

I use AI every single day

Probably more than most people.

• I use it to debug weird issues.

• I use it to write boilerplate.

• I use it to sanity-check architecture decisions.

Sometimes it points me in a direction I wouldn’t have considered.

Sometimes it confidently suggests something so wrong it makes me question whether we’re looking at the same codebase.

That’s the reality.

AI is incredibly useful.

It is also incredibly… human in the sense that it gets things wrong more often than the demos would have you believe.

And that’s okay.

a few hobby projects changed how I think about AI

A few months ago, I finished building Everly, a wedding website builder and planning platform.

Around 60% of the codebase came together with the help of AI.

That number surprises people.

Some assume it means the project took a few weekends.

It didn’t.

It took months.

The product decisions were still mine.

The architecture was still mine.

The countless moments where I threw away an approach because it just didn’t feel right? Mine too.

AI wasn’t building the product while I watched.

It became the best programming partner I’ve ever worked with.

It wasn’t about speed

Ironically, speed wasn’t even the biggest benefit.

People talk about AI like it’s some magical productivity multiplier.

For me, it was something much simpler.

It kept me in flow.

Before AI, my browser looked like every developer’s browser.

Twenty tabs open.

Documentation.

Stack Overflow.

GitHub issues.

A random blog from 2019.

A Reddit thread with an answer that almost matched my problem.

Back to my editor.

Back to Google.

Back to the docs.

Repeat.

The hardest part wasn’t writing code.

It was constantly breaking my concentration.

Today, instead of searching for snippets that are close enough, I explain the exact problem I’m solving.

The conversation happens in the context of my application.

My architecture.

My constraints.

I spend less time searching and more time thinking.

That’s a much better trade.

It feels like having a superpower

The best way I’ve been able to explain it to friends is this.

Imagine you’re already strong.

Now someone hands you an exoskeleton.

You’re still lifting the weight.

You’re still responsible for where you’re going.

You still need experience to know what not to lift.

But suddenly, you’re capable of doing far more than you could on your own.

That’s what AI feels like to me.

It didn’t replace my engineering.

It amplified it.

It removed friction.

It shortened the distance between an idea in my head and something people can actually use.

As someone who genuinely loves building products, that’s an incredible feeling.

I don’t think we’re competing with AI

I think we’re competing with engineers who know how to use it well.

That’s a very different problem.

When Git became mainstream, nobody said Git was replacing developers.

When Stack Overflow exploded, nobody stopped learning because answers were one search away.

When cloud platforms made infrastructure easier, companies didn’t suddenly stop hiring engineers.

The tools evolved.

The job evolved.

We evolved with them.

AI feels bigger, maybe it is but I still believe the same principle applies.

The engineers who embrace great tools usually outperform the ones who spend their time arguing against them.

One thing my career has taught me

I’ve worked across startups.

I’ve worked with enterprise teams.

Today I work remotely with incredibly talented people spread across different countries and time zones.

If there’s one thing I’ve learned, it’s this:

Nobody pays you because you know React or TypeScript, or the framework everyone will stop talking about next year.

They pay you because you solve problems.

Sometimes that means writing elegant code.

Sometimes it means convincing everyone not to build the feature they’re asking for.

Sometimes it’s catching an architectural issue months before it reaches production.

Sometimes it’s understanding people well enough to build what they actually need instead of what they asked for.

Those things don’t disappear because AI got better at writing functions; if anything, they become even more valuable.

The internet profits from panic

Nobody goes viral saying, “AI helped me finish a feature without opening thirty browser tabs.”

But tell everyone software engineering is dead?

Millions of impressions.

Fear scales much faster than perspective.

I’ve learned to be careful not to confuse engagement with reality.

The builders are usually quieter

One thing I’ve noticed recently is that the people building genuinely great products aren’t spending all day debating AI.

They’re too busy using it.

Learning it.

Questioning it.

Correcting it.

Shipping with it.

Meanwhile, everyone else is arguing about whether developers still have a future.

That contrast says everything.

Experience compounds

One thing AI can’t generate is your experience.

The production incident you stayed awake fixing.

The migration that almost failed.

The difficult client conversation.

The instinct that tells you something feels wrong before you can explain why.

Those lessons aren’t stored inside a model.

They’re earned.

And they quietly become your biggest advantage.

My advice?

Build anyway, seriously!

Learn the tools.

Experiment with them.

Delete the bad code.

Keep the good code.

Stay curious.

The worst thing you can do right now is freeze because the internet convinced you the future has already been decided.

It hasn’t.

If anything, this is one of the most exciting periods I’ve experienced as an engineer.

Not because AI writes code.

Because the distance between imagination and execution has never been smaller.

That’s a gift.

Build through the noise

Maybe our jobs look different five years from now.

Honestly, they probably will.

Mine already looks different than it did three years ago.

But after building a real product alongside AI, I walked away with a very different feeling than I expected.

Not fear.

Gratitude.

I genuinely enjoy building more today than I did before.

The friction is lower.

The feedback loop is shorter.

The possibilities feel bigger.

Every generation of engineers gets its defining moment, mine just happens to include AI.

And maybe one day we’ll look back and laugh at how much energy we spent arguing about whether AI would replace developers.

Because after the headlines disappear…

After today’s benchmark becomes tomorrow’s history…

After another model inevitably outperforms the current one…

Someone still has to imagine.

Someone still has to make decisions.

Someone still has to care about the people using the software.

Someone still has to build.

Why shouldn’t that someone be you?

P.S. Follow me on Twitter. I write about software engineering, AI, building products, and the occasional lesson that only shows up after shipping.