Is Software Development DEAD?

What a title, right?

Today I was on-site at a customer. One of the board members showed me a recruitment app he had built himself with Lovable. It replaces the SaaS tool they had been subscribing to. Two to three days of prompting and configuring, and the result fits their processes better than any off-the-shelf product could. It works. And the team is already using it productively.

I was genuinely impressed. Not only by the app, but by the mindset: a board member who doesn’t wait for a project, but simply tries it out.

On the way home, one question wouldn’t let go of me.

If a board member can rebuild a product in three days, what exactly were they paying for all these years?

Not the code, apparently. Maybe it was never the code and technologies. Maybe they were paying for someone else’s idea of how recruiting should work. A best practice, built for thousands of customers, and therefore for no one in particular. Now they have their own.

And here’s what struck me most: the app wasn’t the impressive part. The impressive part was that he knew exactly what to build. Every screen, every step, every rule reflected years of experience in how his company hires. The AI didn’t bring that. He did.

That’s something I keep seeing in my work with agents: the agents work. What decides the outcome is the system around them. Knowing what’s worth building. Turning that knowledge into clear specs. Making decisions when things are unclear. In this case, that system lived in one person’s head, and for one app, that was enough.

But what happens when a whole company starts building like this? Ten apps, twenty processes, many people prompting in parallel. Then domain knowledge stops being a personal talent and becomes the core of how a company creates software. And advice changes with it. Less “which tool should we buy?”, more “how should this process actually work, and how do we make that knowledge usable for agents?”

So, is software development dead? I don’t think so. But it’s moving. Away from writing code, towards understanding domains, shaping specs and designing the system that agents work in.

Which makes me wonder:

• What is a software company worth, if its product can be rebuilt over a weekend?
• And what is domain expertise worth, now that it can be turned into software in days?

I’d genuinely like to hear your thoughts.