Turning 40 and Starting Over as a Developer — I Was a Legacy Developer

This is the first post on this blog. Why I decided to start developing again, and how a legacy developer coming back after years away decided to study — originally two separate posts, merged into one here.

Why Development, Again

I worked as a developer through my 20s and into my early 30s, then moved into project management (PM) for the back half of my 30s. But turning 40 brought a lot of change into my life.

Back when I first started as a developer, in the early 2010s, being a solo developer was still viable. You could build a whole site by yourself and it would be enough.

By the late 2010s, technology had grown more advanced and complex. Backend, frontend, and design split into their own disciplines, the development process fragmented, and the world became a much harder place for a one-person team.

Around that time my role shifted from developer to project manager, and from 2021 on I stopped coding entirely.

Then came a break, and AI advanced enormously over the past year.

At first I thought, "how much could AI really do for development?"...

Then I saw a site built from a two-line prompt, and I was floored.

It felt like solo development was possible again — like the early 2010s all over again.

So I decided to give development another shot.

I've had ideas sitting around for a while but never built them, always blaming time and being busy. I wanted to finally make some of them real.

I'd like to make YouTube videos like everyone does these days, but I'm from the generation that's more comfortable with text... writing just feels natural to me.

This site itself was built with AI in a few hours — domain connected and all within about two hours.

From here on, I plan to write down, piece by piece, how I built each thing.

I Was a Legacy Developer

Diving into AI development this time, I quickly realized how much there is to learn.

CLAUDE.md setup, configuring AI agents for automation, and on and on... there's genuinely a lot to pick up.

When I write posts like this, some developers might say, "you're still using AI in such an outdated way? You should be doing it this way, that way..."

Starting around 2000, both Korea and the US assumed that giving kids more access to IT devices in education would mean more information access and better academic performance. But twenty years later, kids' academic performance has actually dropped. And so has their satisfaction with learning.

I think mastering the fundamentals leads to bigger gains

Which is probably why universities still teach foundational math and physics.

Build up your development muscles first

I've decided to first understand how AI actually works, then build things step by step, the way I used to before — the old-fashioned order.

Things like: set up a basic DB, build the site skeleton, wire up the UI...

I'm working through it by checking off each needed feature one at a time and iterating.

And watching AI debug things along the way, I catch myself thinking "ah, that would've been a better way to do it" or "oh, I didn't know you could do it like that" — and I like that I'm learning something too, right alongside it.

After writing these two posts, I actually went on to ship apps to app stores, run live web services, and write over 100 blog posts. The proof of that is the rest of this blog. → See all posts