Engineers and not programmers

I used to work with an owner in the insurance world. He had the expertise related to policies and I had a team so we designed a solution to carry your policy info in your pocket an small but very helpful app. We worked hard over a year with a very small team then, right before going to market he took a different path due to some personal matters and step away. A year later he started over with AI. And he was able to make the same progress we did to have the App running in prod within months.

He built a project close enough to what we had in mind but there were disconnections in the workflows and from the business model to the tech part there were some ideas but nothing in concrete. He then approached me asking for help because even when he was able to build there was a disconnection between the business model and the technical side the app was running but wasn’t right.

That’s the engineer gap. And that’s where we’re going to live.

Code is syntax. Engineering is decision-making.

For years the hard part looked like writing the code. It never was. The hard part is deciding what to build, making the pieces fit and being sure the thing is actually correct. Ai made the typing cheap. It didn’t make the deciding cheap.

AI can generate and even decide, but there’s a lot of ground it doesn’t cover on its own. This is why we use things like Spec-Driven Development. We design, define and plan before the agents build, not because it’s bureaucracy, but because you want to control the outcome and follow your own path. Building blindly leads to losses and problems. Seeing that, managing it, getting ahead of it.

When do you reach for a LiveView? When do you actually need a GenServer? Does this system need to be distributed, or are you adding complexity you’ll fight later? None of those are typing questions. They’re judgment. And judgment is exactly the part that stays scarce.

I don’t believe programmers are dead. I believe they are evolving but I won’t pretend evolution is comfortable or fair to everyone.

Now writing code is not enough. This era will be difficult to those who based their role on the ability they had to produce it. I see a more clear path to those thinkers and directors or machinery and even more important the ones who can control it.

At my last job our CTO leaned hard into AI for dev operations. Good call. Then claimed AI should make us 12x more productive. Making such a bold statement still give me chills.

AI is a tool a powerful tool that absolutely can enhance productivity but as humans AI is not a tool to mesure our capabilities. To me and my own team (here at Nawal) I see it as a something that assist and improve the performance of my engineers. Deadlines have been a thing we decide together not a number I impose just because I’m fast and AI made me faster.

Fundamentals will be more visible now because the people who understand them will take advantage of those tools. Here is where I remember KISS a methodology my first mentor taught me and that simplicity overall makes me see the multiplier force this tool can add into my workflows and my projects.

At the end of all there’s a line I want to draw clearly. Using AI to think with, to sharpen and idea, to pressure-test your own reasoning it not the same as using AI to speak for you.

The first makes you better the second replace the one thing that was already yours. Your judgement and your name on the outcome.

The future of software engineering looks different. However we shouldn’t allow prompts or people who only do prompts to decide for us. Opinion matters, ideas have their own weight and humans are very capable. Stop doing prompts use AI to assist you to challenge you to improve.

This new era is here and we should be working towards our new software engineers generation.