From Life Insurance to Full-Stack
My journey building in public (and why I think AI is just a really good calculator)
Look, I'm not gonna pretend I knew what I was doing when I started this thing.
Three years ago I was selling life insurance. Hated every second of it. Like the kind of hate where you're sitting in traffic going to work and you're actively trying to think of legitimate excuses to turn around. I wanted to build something, create something, but I had zero coding experience. Zero.
So I did what everyone does when they want to skip the hard part—bought into Go High Level. Thought drag-and-drop would be enough. Spoiler: it wasn't.
Then WordPress with page builders. Which was... better? But I still felt like I was fighting the tools instead of actually building anything. Like trying to assemble IKEA furniture but all the pieces are for different products and none of the instructions match. So I started writing my own PHP files. Editing themes. Breaking literally everything. Fixing it. Breaking it again. Learning very slowly that I had no idea how databases worked.
Every time I switched platforms people probably thought I was just indecisive. But it wasn't that. It was hitting walls and realizing "oh shit I'm not doing this as well as it could be done" and that bothered me way more than starting over. Which is probably a personality flaw but here we are.
The Project Nobody Will Ever See
About two years ago I built this real estate website for an agent in Las Vegas. Full WordPress setup with IDX Broker integration. Listing feeds, search functionality, all of it. That was the first time I had to connect an actual API and do backend work.
Except I didn't even know that's what it was called. I was just trying to make the damn listings show up on his site and Googling things like "how to make WordPress talk to other website" because I didn't know the terminology.
The agent was brand new. And if you know anything about Las Vegas real estate you know it's absolutely brutal. Like trying to break into Hollywood level of competitive. We couldn't get any traction. Project's basically dead now. I have nothing to show for it.
And that's the thing about client work when you're starting out right? Their business fails, your work disappears with it. No portfolio piece. No proof you built anything. Just you sitting there knowing you did all that work with literally nothing to point to.
And honestly? If I'd had a developer friend back then—someone I could ask "hey is this approach stupid" or "am I building this the hardest way possible"—I might've avoided some dead ends. But I was just figuring it out alone, Googling everything, hoping I was doing it right.
It's the worst.
Wait, I'm a What Now?
So I kept building stuff. Learning. Breaking things (I'm really good at that part). Using AI more and more because manually debugging was making me want to quit entirely.
Then like six months ago I'm talking to Claude about something code-related and I just randomly ask "hey how would you describe me and my ability to build websites" and Claude goes "you're a full-stack developer" and I'm just sitting there like... wait what?
So I Googled it. Full-stack developer. And I'm reading the definition and I'm realizing—oh. Oh shit. I've been doing this for over a year. APIs, databases, frontend stuff, authentication, deploying things. That's all full-stack. I just didn't know that's what you called it.
An AI had to tell me what I was before I knew it myself.
Which is kind of hilarious and kind of embarrassing but also perfectly sums up this whole journey.
This is why I think we need more people building in public and sharing what they learn. I literally didn't know what to call myself until an AI told me. How many other people are out there doing the work but have no idea what category they fit into? Or what they should learn next? We need to talk to each other more.
But the REAL aha moment? That happened last week.
I finally got AI Art Arena working to the point where I could actually show people. And I'm looking at everything I built—the auth system, the database setup, the real-time voting, the cron jobs that automatically cycle contests every week, the admin dashboard—and it hits me: I actually understand how all of this works now. Like really understand it. Not just "I copied this from Stack Overflow and it works" but genuine understanding of what I'm capable of building.
That's why I'm doing the whole building in public thing now. This is different from that Las Vegas site or any other client project that went nowhere. This is MINE. Even if the contest completely flops, I still have this functioning application I can point to and say "I built that."
Okay So About AI Though
Here's what I need you to understand and I know this might be controversial but whatever: Everything you see on this site was built with AI. Every line of code. Every image. Every word. And I'm not trying to hide it—I'm straight up celebrating it.
Because without AI? None of this exists.
But also—and this is important—I'm the architect. AI is just the builder following my blueprints. There's a huge difference.
Calculators
Didn't make math obsolete—they made it faster
Printers
Didn't make writing obsolete—they made copies faster
Computers
Didn't replace NASA's thinking—they computed faster
You ever see Hidden Figures? That movie about the Black women who worked at NASA doing all the calculations for the space program? I didn't realize until I watched that movie that "computers" used to be PEOPLE. Like that was their job title. "Computer." Because they computed things.
Then machines came along that could compute faster than humans and we started calling the machines "computers" instead of the people. But the humans were still doing all the thinking, all the architecture, all the decision-making. The machine just did the computation faster.
That's what AI is. It's a calculator for code. A printer for words. A computer for computation.
AI doesn't make development obsolete. It just makes execution faster. But you still need to architect the whole system. Evaluate if the code is secure. Debug when things break (which is constantly). Decide if this is even the right approach in the first place.
AI is only as good as the person using it and how much work you're willing to put in.
The Work Nobody Sees
When I built AI Art Arena I didn't just type "build me a voting platform" and watch it magically appear like some people seem to think AI works.
I had to design the entire system first. Like sitting down with a notebook and figuring out: What tables does the database need? How do we prevent someone from voting 50 times? What happens when a contest ends? How do we handle people who are logged in versus anonymous voters? What security stuff do we need so we don't immediately get hacked?
I'm not asking "make a voting system." I'm asking stuff like "I need a Next.js API route that checks if a user voted in the last 24 hours by querying their last vote timestamp, validates the artwork_id actually exists in the database, hashes their IP address with SHA-256 before storing it for privacy, increments the vote count atomically in Supabase so we don't get race conditions, and returns proper error codes for rate limiting, duplicate votes, and invalid artwork IDs."
And even then? AI's first attempt is usually wrong. Or it's right but not optimal. Or it works but doesn't handle edge cases I didn't think to mention.
So I read through the code. Test it. Break it. Realize it's not handling something properly. Regenerate with more specific requirements. Test again. Refactor because it doesn't fit with the rest of my architecture. Repeat this like 10-20 times per feature.
Every single feature in this app went through that process. Every. Single. One.
And yeah, if you're reading this and thinking "he's doing X wrong"—PLEASE tell me. That's literally why I'm writing this. I'd rather know now than six months from now when I have to rebuild everything.
The mental work is all mine. The architecture, the decision-making, the vision, the hundreds of iterations, all the evaluation and refinement. AI just speeds up the actual typing part.
The tool changed. The work is still there.
What I Actually Built
The Tech Stack
- •Next.js 14 with TypeScript - my choice, not AI's suggestion
- •Supabase for backend - looked at like five different options and picked this one
- •NextAuth - GitHub, Google, and email magic links (designed the whole auth flow)
- •Real-time voting - 24-hour cooldowns (had to architect all that logic myself)
- •Automated weekly cycles - Vercel cron jobs (took me forever to figure out the scheduling)
- •Admin dashboard - role-based access control (designed the entire permissions system)
- •Security measures - rate limiting, IP hashing (went down a rabbit hole researching best practices)
For every feature I had to know:
WHAT to build
The requirements
WHY it needs to work this way
The logic
HOW to structure it
The architecture
WHEN AI's solution is wrong
Evaluation
Could I have built this without AI? Honestly no. The manual typing, the semicolon debugging, the endless documentation reading—it would have broken me. I would have given up.
But did AI "build" this? Also no.
I built this. AI was my tool.
Why I'm Doing This Publicly
You learn way more
If I'm building in private and making architectural mistakes I'll never know. Building in public means someone with more expertise can tell me "hey here's a better way to do that" and I actually want that feedback. I need it honestly.
It creates a portfolio as you go
I'm not waiting until everything is perfect to show people because perfect never happens. Every commit, every feature, every iteration is visible. That feels more valuable than a polished final product with zero context about how you got there.
It holds you accountable
When you publicly say "I'm building this" people kind of expect to see progress. Which keeps you moving forward instead of abandoning projects when they get hard (which I've definitely done before).
It shows the reality
If you're switching careers into development you need to see that it's not a straight line. That people rebuild stuff multiple times. That you don't need to understand everything before you start. Building in public shows the messy reality instead of just the highlight reel.
Building With People, Not Just Code
Look, I'm not trying to be salesy here or push anything on you. But I'm genuinely looking for people to build alongside. Not clients. Not followers. Just... people.
Developers
If you know more than me, I'd love to pick your brain sometimes. Review my architecture, share knowledge, collaborate.
Career Switchers
Earlier in the journey than me? Ask me anything about Next.js, Supabase, authentication, deployment, whatever.
Anyone Building
Just want to talk about what we're building? Share the journey? Let's connect.
Because honestly? Learning to code alone kind of sucks. Having people to share the journey with makes it way better.
Let's Build Something Together
This is just the beginning.
Come check it out. If you see something I'm doing wrong, tell me. If you're working on something similar and want to compare notes, let's talk. If you have questions about anything I mentioned, ask.
I'm not gatekeeping any of this.