This started life as a video I recorded in 2026, rewritten here because it is the clearest thing I have said about my own job in years. I used to spend my days deep in Magento. Five developers and six months to build an e-commerce site that today I could stand up in an afternoon. Then I stopped. Not only the videos, the work itself. Not because I quit, and not because I burned out, though there was a bit of that. Because the job changed, and I do not think enough experienced developers are saying out loud what it changed into.

The role I do now is blurred

I still build things. I have a whole platform, a CMS and commerce stack, that I will get into another time. But the day to day does not look like writing code used to. I am not really a developer in the old sense. I am not a project manager either, though I do a bit of that. I am not a designer by trade, yet I have made more design decisions in the last year than most designers I know. The lines between the roles on a team just do not sit where they used to.

AI did not replace developers. It replaced the job description.

Everyone has a take on AI and most of them are wrong. Here is mine. The thing that changed is not that the work disappeared. It is that the work moved. The tasks I used to spend three days on take about fifteen minutes now. Writing the boilerplate, wiring an integration, scaffolding a module. That part genuinely compressed.

But the things that actually matter got harder, not easier. Deciding the architecture. Knowing which corner you are allowed to cut and which one will cost you in eighteen months. Most of all, knowing what not to build. AI will happily help you build the wrong thing faster than ever. The judgement about whether it should exist at all is still yours, and it is now the bottleneck.

What this means if you run a digital team

If you are hiring or managing developers right now, the instinct is to measure output. Lines shipped, tickets closed, features delivered. That made some sense when typing was the constraint. It makes very little sense now that typing is the cheap part.

A concrete version of this. A client asks for a custom dashboard with twelve filters. The old answer was a quote and six weeks. The honest answer is usually that they will use two of the filters, so the right move is to ship those two next week and wait to see if anyone asks for the rest. Saying that out loud is worth more than any amount of fast typing.

The people worth their salary now are the ones who stop you building things. Who look at a feature request and ask whether it solves the real problem, or whether it just adds surface area you will maintain forever. That skill does not show up on a velocity chart, and it is the one thing AI cannot do for you, because it does not carry the consequences. You do.

Why I am talking about this

I think a lot of senior developers are quietly in the same place. The job they trained for is not the job they do, the title on the contract stopped describing the work a while ago, and nobody is being straight about it. That is the gap I want to fill. How a digital team should actually work now, how AI should actually be used, the reality rather than the hype. The stack I build will show up along the way, and so will the opinions. I have got plenty, and I have been doing this long enough to know which ones are right.