If The Build Is Free, What's Actually Worth Building?

Something weird is happening to software. And I say that as someone who builds it.

Last week, I shipped something in twenty minutes that would have taken me days two years ago. Not a prototype. A working thing. And the strangest part wasn’t the speed. It was realising I’d spent almost no time thinking about how to build it. Just what I wanted it to do.

That shift is bigger than it sounds.

The gap that defined us is closing

We’ve always had the bottleneck of building. Even good developers have it. The gap between idea and working software was measured in time, context-switching, and technical debt that quietly accumulated in the background. That gap defined entire careers. It’s why senior engineers got paid what they did. It’s why “execution” was the answer to every “what matters more, the idea or…” question.

That gap is closing so fast it’s making my head spin.

And it raises a question I can’t stop turning over: if the build is nearly free, what’s actually worth building?

When your competitor can replicate your feature set on a Tuesday afternoon, what’s left to differentiate a digital business? The answer used to be speed of execution. Ship faster, iterate faster, learn faster. But when everyone can ship in an afternoon, speed stops being a moat. It becomes table stakes.

“When everyone can ship in an afternoon, speed stops being a moat. It becomes table stakes.”

The value is moving underneath the software

I think the answer isn’t in the software layer anymore. It’s underneath it.

The data. The relationships. The domain expertise. Understanding why this problem exists in the first place, not just knowing how to build a solution for it.

Think about the businesses that have lasted through every previous wave of disruption. They’re rarely the ones with the best technology. They’re the ones who understood their customers so deeply that the technology was almost incidental. The software was just the delivery mechanism for insight they’d earned the hard way.

That’s always been true in theory. But when building was expensive, the technology itself provided a buffer. You had time to figure out the deeper value proposition because your competitors couldn’t catch up quickly. That buffer is gone now. If your only advantage is “we built it first,” you’ve got weeks, not years, before someone else builds it too.

The middleware squeeze

Here’s where this gets uncomfortable for a lot of companies.

The platforms that currently sit between technology and people. The middleware, the SaaS layer, the “we make this accessible” businesses. They’re about to have a very uncomfortable few years.

Because what does a platform that simplifies a complex task do when AI makes that task simple anyway? If your value proposition is “we save you from having to code this yourself,” and AI just learned to code it for the user directly… what exactly are you selling?

This doesn’t mean all SaaS dies overnight. Products with deep data moats, strong network effects, or genuine workflow lock-in will survive. But the “thin wrapper around a commodity” businesses? The ones that took something slightly technical and put a nice UI on it? That category is getting compressed hard and fast.

The winners will be the companies that own something AI can’t easily replicate. Proprietary data. Trusted relationships. Regulatory expertise. A community that comes for the product but stays for each other. Things that take years to build and can’t be prompted into existence.

“When anyone can build anything, the person who knows what to build has all the leverage.”

What this means if you’re building right now

If you’re a founder or a builder reading this, the practical question is: where do you put your energy?

Not into the build. The build will increasingly take care of itself. The energy needs to go into the stuff upstream of the build. The research. The customer conversations. The domain knowledge that tells you which of the hundred things you could build this week is the one that actually matters.

I’ve been feeling this shift in my own work. I used to spend 80% of my time building and 20% thinking about what to build. That ratio is inverting. The building part collapses into hours. The thinking, the talking to users, the understanding of the problem space… that’s where the weeks go now. And honestly? That feels right. That’s where the weeks probably should have been going all along.

The irony is that AI making the build cheaper makes the pre-build work more valuable, not less. When anyone can build anything, the person who knows what to build has all the leverage.

The question nobody’s asking

We’re all having conversations about AI replacing jobs, AI augmenting workers, AI changing this industry or that one. Big, macro, structural questions.

But the question I keep coming back to is smaller and more personal.

If you can build anything… do you know what’s worth building?

Not what’s technically possible. Not what’s trendy. Not what your competitor just shipped. What actually solves a problem that someone will pay for, that AI can’t trivially replicate next month, and that you specifically are positioned to understand?

That’s the hard question. And it has nothing to do with technology.