Over the last few days, my TL has been swarmed with people modding existing games into other games — or actually game taking control of machines.

It’s really fun and exciting to watch. And what’s even more awesome is that a lot of it comes from people with almost no technical background: just an AI subscription and an explorative mindset.

At the same time, runtimes are being migrated to Rust by a single person, GitHub is on the verge of exploding with more and more code being produced every minute, and CEOs are apparently rewriting core modules in Assembly.

This contrast made me think about something:

Once technical capacity is given to everyone, where does the value of creating something move?

A sneak peek into history

Often, when faced with something we think is entirely new to our time, looking back at history is more than advisable.

This isn’t the first time we’ve lived through a period where people suddenly gained access to capabilities in fields that previously required years of work just to become relevant.

(And yes, I can’t flex my weather app to my friends anymore, and I’m sad about it.)


Let’s talk music

How was music produced before we gained access to synthesizers, samplers and DAWs?

Well, mostly by people who could actually play music!

Today, I can spin up an app and try to become the next J Dilla — spoiler: it won’t work — whereas a few decades ago I might have needed a $tudio that would’ve financially ruined me.

But music isn’t “solved”, is it?

I could face Fred again.. in a sampling battle tomorrow. We could have access to the exact same samples, the exact same software and the exact same tools.

He would still destroy me.

Because once access to the tool is no longer the differentiator, the choices made with that tool become the differentiator.

What do you sample? What do you leave out? What sounds wrong but somehow feels right? When do you stop?

Those choices are what make his production his — and mine, well, mine.

The tool gives you capacity. It doesn’t give you taste.


Photography gives us another example.

To borrow an example I recently heard in a talk from someone in the Ruby ecosystem: photography dramatically lowered the barrier to representing reality. Then smartphones lowered the barrier to photography itself.

Today, almost everyone walks around with an incredibly capable camera in their pocket.

Everyone can take a snapshot of a moment.

But did everyone become a professional photographer?

Of course not.

Once pressing the shutter becomes trivial, the value moves somewhere else: what you decide to point the camera at, how you frame it, when you press the shutter, and why you thought that moment was worth capturing in the first place.

Back to technology

So nowadays, almost anyone can spin up an agent and create a website.

Cool.

The question becomes:

Where is the soul in it?

I’ve developed PTSD from what I call “Claude Design.”

OVH VPS dashboard with a blue gradient header, rounded cards, and resource statistics.

You know the one.

The gradients. The cards. The giant rounded corners. The carefully spaced sections. The technically clean but somehow completely interchangeable interface.

Whenever I see websites like this, my first thought isn’t “AI made this.”

It’s:

Did anyone actually look at the output?

Did anyone care enough to ask whether this was really what they wanted to make?

That’s an important distinction.

AI isn’t the problem here. Indifference is.

In my opinion, the real differentiation between average and great work over the next few decades will increasingly come down to two questions:

Do you actually have something to say?

And:

Do you actually care about your craft?

I don’t give a damn about the design of some internal CRUD tool used by twelve people twice a month. Generate the whole thing if you want.

But when what you’re building is meant to face actual customers, you have to take responsibility for what you’re creating.

Use the AI. Generate the code. Generate twenty designs. Ask five agents for alternatives.

But at some point, you have to make a choice.

You have to decide that this interaction feels better than that one. That this animation is unnecessary. That this weird idea is worth keeping even though the model keeps trying to “fix” it.

You have to inject a little bit of your soul into it.

For the better?

And that’s why I’m actually really excited about this era of building.

Our limits are increasingly becoming twenty bucks and our imagination.

That’s fucking awesome.

But it also removes one of our favorite excuses.

We can no longer hide mediocre ideas behind “expertise.”

When producing something technically impressive required years of specialized knowledge, technical execution itself could make the result impressive.

As that execution becomes cheaper, that’s no longer enough.

So iterate.

Iterate.

Iterate.

Explore things outside your field. Look at what architects, game designers, musicians, filmmakers, fashion designers and random people on the internet are creating.

Steal ideas from places your model wouldn’t immediately associate with your problem.

Develop taste.

Try weird shit.

Maybe it’ll suck.

Maybe you’ll accidentally create the next neobrutalist design trend that everyone will copy for two months.

But at least it’ll be yours.

Because if history tells us anything, making the tools accessible to everyone doesn’t kill creativity.

It just moves the bar.

And maybe the interesting question isn’t:

“What can you build now?”

Almost everyone will soon be able to answer that.

Maybe the question is:

“Now that you can build almost anything, what do you actually want to build — and why?”