I build software like a toddler learning to walk
I know almost nothing about coding. My high school specialized in IT-Economics, and somehow I still walked out unable to write a single working line of code. Whatever I learned back then evaporated years ago.
For most of my career, I did what smart-sounding entrepreneurs do. I outsourced every piece of development. Someone else touched the code, I touched the strategy.
That felt fine until AI made it possible to build almost anything just by describing what you want, and the whole game changed under my feet.
To be fair, I haven’t touched SaaS or built a startup in over five years. My life looked completely different: low-key consulting, day trading, and grinding semi-pro poker, and semi-pro poker, mostly PLO cash games but occasionally major tournaments like the WPT. (My tournament results tracked are awful, by the way)
Anyway, 2026 is a real career change for me, and this blog is also where I track my own development, not just my companies', so I want to be honest about where I started and where I am now with coding, building, and shipping software.
When I first sat down to build something real, I had no fucking clue where to begin. None. I felt like a toddler standing up for the first time, wobbling forward, praying I wouldn't fall flat on my face and break the one thing I'd managed to build.
I poked around in Lovable and a few other vibe-coding platforms, and that fell apart fast. Even when I got a front-end MVP working, which really just meant some nice-looking packaging that didn't do anything underneath, I kept slamming into the same wall over and over.
So I forced myself to learn more, and to learn better. Going into 2026, I planned to become a full-time content creator. That plan quietly shifted the moment I unlocked something with Claude AI and shipped my first working beta product a month later.
Then came a handful of micro-apps, built, shipped, and now heading into distribution.
Read also: I built a free subscription tracker
Here's the part that still makes me laugh:
I did all of that inside the GitHub web editor. Not GitHub Desktop. Not even github.dev. The actual browser-based editor, the most basic fucking version there is.
My process was stupidly simple
I would explain to Claude what I needed. I'd adjust things where necessary and ask for a full file replacement. I'd paste that into github.com and hit commit.
Then I'd watch my Vercel deployment counts shrink because I kept hitting the daily limit on how many times I could deploy in 24 hours.
For a while, that was fine. It worked. But now I'm managing more than 20 repos on GitHub, and at some point continuing to work that way stopped being a quirky habit and started being genuinely ridiculous.
Finally touching an IDE
Until recently, I had never used an IDE, an Integrated Development Environment. Cursor, Windsurf, VS Code, the tools that anyone who builds software uses every day. I'd tried before and bounced off immediately.
The layout confused me. I didn't understand a single fucking thing. How to clone a repo locally or push changes through anything that wasn't just a text box on a website.
If I'm honest, I'm still not fully comfortable with it. But I'm getting there, one small step at a time.
But I finally took the step. I moved from the GitHub web editor to an IDE, and instead of pushing every tiny change the second I made it, I started waiting until I'd altered a few files before committing.
Small, but it's a real shift in how I think about my own work.
What I need to learn (too much obviously)
Building and learning alone is hard, and I'm not going to pretend otherwise. But I also can't keep lying to myself that I can build something ambitious using nothing but a web editor and good intentions.
That realization pushed me to commit real time every week to a few things:
understanding IDEs properly
learning how Supabase and similar platforms work under the hood
understanding how the stack you choose shapes your speed and your decisions later
figuring out how to properly back up my code and document my own progress as I go.
That list could go on forever, and that's exactly the point.
I was never a developer, a builder, or an indie hacker. I was an entrepreneur who lived in corporate frameworks, sales, and marketing, the kind of person who delegates, outsources, or hires an expert the moment something gets technical.
Only now am I realizing that not looking closely enough at how things get built was my biggest mistake.
The good news is that absorbing knowledge moves a lot faster with AI than it ever did before.
The part I'm watching for now is technical debt, the kind of shit that quietly piles up until you wake up buried in it, another founder in the startup graveyard who built without ever fucking understanding what they were building on.
Where this leaves me
I'm not writing this because I've figured it out. I haven't. I'm writing it because a month ago I couldn't imagine shipping anything, and now I have a beta product live and a handful of micro-apps out in the world.
That's not nothing, even if I still feel like I'm relearning how to walk every time I open a new repo.
Currently, I need to keep building without lying to myself about how much I fucking understand from all this.. That's the whole plan for now.