Things are moving fast. Too fast to stay in doubt for very long.

Yet I still meet people debating what has already been decided. I am genuinely curious.

Is anyone still in doubt that you can build production code with AI?

Is anyone still in doubt that the developer role, as we know it, won’t exist in a year or two?

Is anyone still in doubt that everyone on a team building digital products will write or prompt code in the future?

The answer should be no to all three. Let me take them one at a time.

Production code with AI

I have four products running. Knud is in the App Store. Super Yes runs on real users. I wrote the code with Claude Code. I am not a developer in the classic sense.

And I am far from the only one.

The debate about whether it’s possible is over. It is possible. It is happening right now. People who have never opened an editor before are shipping things that real people pay for. Next question.

The developer role

I am not saying every developer loses their job. I am saying the role changes.

The person who understands systems, architecture and complexity remains indispensable. The person who writes code for a defined problem? That role melts into everything else.

When one person can do the work that used to take five, the organization changes. Or it shuts down. There is no third scenario.

Let’s stop pretending it’s an open question.

Everyone writes code

Or rather: everyone prompts code. The designer prototypes. The product manager builds the small internal app that used to cost half a sprint. Support writes their own scripts for the data pulls they always wanted.

That is not a prediction. It has already started.

The interesting consequence is not that everyone becomes a developer. The interesting consequence is that the ability to write code stops being a differentiator.

So what does it mean?

The new bottleneck is not technology. It’s human disagreement.

Ten people with access to AI can produce ten different versions of the truth in an afternoon. All of them work. All of them can be demoed. All of them have their own champions. Someone has to facilitate what lives on and what dies.

That is not a technology problem. It’s a people problem.

So what probably won’t go out of fashion is:

  • Understanding the customers, the domain and the problem to be solved
  • Facilitating disagreements toward decisions
  • Daring to cut, saying no to 90% so the 10% can live
  • Judgment about what is good enough

It sounds banal. It is. It’s just not easy.

Good product managers have always done that work. In a year or two, it’s the only work left.