I have been building digital products for over a decade. Every two weeks, I share a single decision from my practice: what I built, what I left out, what I handed over to AI, and what I kept for myself. It comes complete with the reasoning, so if you face a similar call, you do not have to pay the same tuition.
One email every two weeks. Unsubscribe in one click.
Why only one?
Because anyone can send a newsletter full of links, but a real decision either happened or it did not. Focusing on a single topic is not a design detail; it is a strict rule. It protects you from spending ten minutes on something that fits into one paragraph, and it stops me from writing filler just because it is send day.
The approach
To me, sustainability is not a buzzword: it is being life-centered. In the AI era, the key question is no longer what we can build, but what is actually worth building. I focus on software decisions that put human time and focus ahead of an endless task list.
What is inside a letter?
The actual decision I made.
The context and constraints that shaped it.
The alternatives I rejected and why.
What it cost, or what it would have cost to get it wrong.
Three recent decisions
One on software development, one on managing my time, and one on pricing. Here is the essence below; inside the letters, I also show what each one cost and how it turned out.
AI in my own work
I let the machine build it. I do not let it decide what to build.
The more work I hand over to AI, the clearer the boundary becomes. I no longer write most of the code myself. But whether a feature makes sense in the first place, what order to follow, and what to drop remains my call. There is no undo button on these: a bad line of code takes ten seconds to replace, but a bad feature stays on your shoulders for six months.
What I rejected: letting development speed dictate the product scope. Just because the machine got faster does not mean my professional judgment became replaceable.
Sustainable building
I do not fill time saved by AI with more tasks.
I built a feature for Modulyo in a fraction of the time it would have taken a year ago. But instead of stepping away from the desk, I opened the next task, then another, and by Friday I had bought back zero time. Efficiency alone does not reduce workload; it increases it. That was true for coal in the nineteenth century, and it holds true for a workweek today.
What I rejected: the default hustle. Saved time only stays saved if I leave it empty, and I have to remake that decision every single week.
From services to products
I published my prices. Eight capabilities became four named packages.
I used to list my skills and write a custom proposal for every request. That setup protected me, not the client: it cost them a call and a week just to find out if we were in the same ballpark. Now all four packages have a fixed price, timeline, and name, and each addresses a specific problem, not a discipline.
What I rejected: default custom proposals. Anyone outside the budget finds out in five seconds without having to ask. Custom proposals are still available under specific circumstances, but making packages the standard made my business significantly more transparent.
Who writes this?
I am David Rozsa, a digital product architect. I bring over a decade of experience across business analysis, UX design, and AI-assisted development, working as a solo builder without a team. Every decision comes directly from my own work, not from someone else's case study. I share them because working alone does not mean everyone has to learn the same lessons the hard way.
