Let me be honest with you from the start: I have no idea what I'm doing.
I run a screen printing and embroidery company in Lithuania. We put logos on t-shirts, hoodies, tote bags — the merch companies order for team-building events that nobody actually wants to attend. I've been doing this for years. It pays the bills. It's fine. And "fine" is exactly the problem.
So a few weeks ago, I decided to build an iOS app — a tip tracker for bartenders and service workers called TipJar. Never opened Xcode before. Didn't know what SwiftUI was. The app is now live on the App Store, and almost nobody downloaded it. But before you feel sorry for me, let me tell you how this actually went — because every "I built an app with AI" post gives you the highlight reel. This is the behind-the-scenes footage.
Why I'm doing this
I could give you something inspirational about following my passion. But the real reason is simpler and less pretty: I'm afraid of spending the next 30 years pressing logos onto hoodies.
And look — screen printing is honest work. I built this business from nothing and I'm proud of it. But there's something that happens when you do the same thing every day for long enough. You drive to the same building. Fire up the same machines. Press the same corporate logos onto the same blank shirts. And slowly, without noticing, you start to feel like one of your own products. Branded. Functional. Replaceable. Probably available in bulk at a discount.
That feeling is what pushed me. Bitcoin played a role too — not the price, but the idea behind it. Sovereignty. The concept that nobody needs to give you permission to build your own future. I realized I believed in that idea but wasn't living it. I had a business, sure. But the business owned me more than I owned it.
So one evening I opened Claude, typed "how do I make an iPhone app," and kicked off the most humbling experience of my adult life.
What building with AI actually looks like
You've probably seen the tweets: "I built a SaaS in 4 hours with AI." Those people are either geniuses, liars, or building something that breaks the moment a real person uses it.
My experience was different — but not in the way you'd expect. The coding part was actually smooth. Shockingly smooth. Within days I had a working app. Real screens, real features, tip logging, shift tracking, hourly rate calculations. Claude ships genuinely good code. I was not prepared for how fast things would move.
The best way I can describe it is directing a movie. You've got this incredible crew — camera operators, lighting, sound — and they're all world-class. They'll film anything you point them at. Beautifully. But if you don't know what scene you're shooting, what story you're telling, what the audience should feel — you end up with gorgeous footage of nothing.
That was my problem. Claude would build whatever I described, and build it well. But I had to decide: what does this screen need to communicate? Is this feature useful or am I adding it because I can? Does this make sense to someone who isn't me? When you can build anything, you have to decide everything. And AI won't tell you your idea is bad. It'll just build it, perfectly, and let you find out later that nobody cares.
I showed the app to my girlfriend at some point. She said "that's nice" — same tone she uses when I show her a new font I found for a client's merch. But she was right. The app worked. It just didn't feel like anything yet. Making it feel right took longer than building it.
Then came the part nobody talks about: everything that isn't code. Apple requires a privacy policy. I embroider logos on hoodies for a living — my hoodies never needed a privacy policy. But my app that twelve people might use? Legal document required. Hosted on GitHub Pages. Like a real tech company, except the CEO also operates a heat press.
Apple rejected me three times. First: privacy issues. Second: a broken paywall I didn't even know was broken. Third: metadata. Each rejection took 24-48 hours to hear back. Each one felt like getting punched in the stomach while checking your email on the toilet.
RevenueCat for subscriptions. App Store screenshots that looked so bad I had to redo them from scratch. Export compliance. Content ratings. This boring, unglamorous, non-codeable stuff took more time than the actual app. No tweet will ever go viral about writing a privacy policy at midnight.
The harsh truths
The code is the easy part now. That sounds insane coming from someone who couldn't write Swift a month ago. But it's true. The bottleneck has shifted. It's no longer "can you build this?" It's "should you build this?" That question is harder than any syntax error.
I understood about 60% of the code in my own app. There are functions that work perfectly and I couldn't explain why. Shipped it anyway. That's just the reality of building this way right now.
Nobody wanted what I built. I added features nobody asked for and missed what people actually needed. I was so caught up in what I could build that I forgot to check if anyone wanted it. AI makes building easy. Building the right thing is still completely your problem.
The flop
TipJar launched. I promised transparency, so here are the numbers:
Downloads first week: 15
Paying subscribers: 0
Revenue: $0
That's it. That's the whole empire. I made more money last Tuesday printing hoodies for a dental conference.
But here's what I've come to terms with: your first app is supposed to flop. It's not a product — it's tuition. I paid about $200 and two weeks of evenings to learn how to ship software. That's cheaper than any bootcamp. And unlike a certificate that sits on LinkedIn where nobody reads it, I walked away with a real app in a real store. That skill doesn't expire.
For the curious, here's what I spent:
- Apple Developer Program: $99/year
- Claude Max: $90/month
- Everything else: $0
I've spent more on ink that turned out to be the wrong shade of blue.
What happens now
I'm not stopping. That's the whole point of this.
My goal is a portfolio of small apps that together bring in $10,000 a month. That's where I can step back from printing and build full-time. The math says 5-10 apps over 2-3 years. I'm fine with that.
Because I'd rather spend a few years building towards something I actually want than spend decades doing work that's comfortable but never quite mine.
App one flopped. App two starts next week. And I'm documenting all of it on X — every dollar, every download, every miss. Just a screen printer from Lithuania who got tired of putting other people's logos on things and decided to build something with his own name on it.
🧭
I'm Neri. I screen print merch and build apps that nobody downloads. Follow the journey on X: https://x.com/neritamo_ . TipJar is on the App Store — if you work for tips, it's for you. If you don't download it, you'll be in good company. https://apps.apple.com/us/app/tipjar-tip-shift-tracker/id6759197318
Top comments (0)