The sentence I lead with everywhere — on my Instagram bio, on the about page, in this post — is that I'm 29, I study theology, and I don't write code. For a long time I assumed that sentence was a disclaimer. An apology attached to the front of my work so nobody would be disappointed when they found out.
I've come around to thinking it's the most interesting thing about the projects, and not for motivational-poster reasons. The constraint genuinely changes what gets built.
I want to be precise here, because "anyone can build anything now" is both the popular version of this story and not quite true.
What used to stop me wasn't ideas or effort. It was a specific, brutal gap: the distance between knowing exactly what I wanted and producing the thousands of lines of exact syntax that make a computer do it. That gap took most people years of full-time study to close, and I was not going to spend those years — I was studying something else I cared about more.
AI tools collapsed that particular gap. Not to zero, and not to "describe it and receive a product." What they did was let me operate at the level where I'm actually competent — deciding what should exist, how it should feel, what happens when someone does the unexpected thing — while the syntax gets handled. I describe, I test, I reject, I describe again. That loop is my whole method.
The parts that did not get easier are worth naming honestly. Knowing whether something is actually finished. Noticing that a photo looks subtly wrong. Sitting with a bug for three hours because the file exists and the browser insists it doesn't. Deciding which feature to cut. None of that was ever about syntax, so none of it was solved by tools.
Everything I understand about how the web works, I learned in the wrong order — from something breaking, not from a course.
I learned what a URL actually is because every photo on this site was broken and the file was demonstrably right there on the server. I learned that phones store photos sideways with a note attached, and that transparent images turn black when saved as JPEG, because my crop tool produced sideways black rectangles. I learned what a crawler sees when it can't run JavaScript because Google rejected my site six times.
A curriculum would have taught me all three cleanly, in an hour, with diagrams — and I'd have forgotten them by the following week. Learning them at 11pm while something I made was broken in front of me means I will never forget them. It's a wildly inefficient way to build knowledge and an extremely effective way to retain it.
People treat "theology student who builds apps" as a contradiction, two unrelated facts about one person. It doesn't feel that way from the inside.
Studying theology is mostly close reading, sitting with a difficult text, and taking seriously questions that don't resolve into clean answers. It also trains a particular instinct: paying attention to what a thing does to people, not just whether it functions.
That instinct is why my projects look the way they do. The drawing wall exists because a room full of people making something together, briefly, is worth engineering for. GuGu is about habits kept with friends rather than alone, because that's the part that actually works. DANA came from watching Korean-Americans struggle to find someone who understands their hair. Hangeulmaru exists because Korean school teachers do enormous work with almost no support.
Not one of those started as "here's a technology, what could I do with it." They started as someone I could picture having a slightly better week.
The real-time drawing wall was built in three days for a church event. I've since been told, more than once, that real-time multi-user anything is not a beginner project. I'm certain that's true. I'm equally certain I only attempted it because I didn't know.
Experience is mostly a gift, but it comes with a filter that quietly removes ideas before you consciously consider them. Not knowing the difficulty rankings means I evaluate ideas on whether they'd be good, and find out about difficulty afterward. Sometimes that costs me a wasted week. Sometimes it produces a thing that gets built specifically because nobody qualified would have tried it on a three-day deadline.
There's a particular kind of person I'm writing all of this for: full of ideas, quietly convinced they're disqualified. Wrong degree, wrong age, wrong job, no team, no funding. I was that person and the conviction felt like realism rather than fear.
What I'd say now: pick the smallest version of your idea, give it a real deadline with actual people attached to it, and accept that the first version will be held together with tape. Ship it anyway. The gap between "person with an idea" and "person with something live" is now mostly persistence — and persistence, unlike a computer science degree, is something you can start today.
From egg to something. The rest of the process, bugs included, is in the build log and on Instagram.