The First Bill Felt Real

There is a strange little moment when a project stops feeling like an experiment.

For me, it was not when the screen looked nice. It was not when I changed the name from QuantisAI to MyTradePilot. It was not even when I spent far too much time and money fussing over the logo, rejecting one version, asking for five more, choosing the paper-airplane-and-candlestick idea, then eventually uploading my own image anyway.

The moment was when I agreed to pay a monthly bill.

Not a one-time logo fee. Not a little design detour I could laugh off later. A recurring cost. Every month. For the market data and behind-the-scenes pieces needed to make the scanner run on something closer to real data.

That felt different.

Up to that point, I could still tell myself this was a free experiment. I am 50. I have no coding background. I am building this with an AI builder, one small request at a time. If it failed, I could shrug and say, “Well, I learned something.”

But a monthly bill does not feel like shrugging.

It feels like the project is quietly asking, “Are you serious?”

The funny thing is, by then I had already had plenty of reasons to be nervous.

Early on, the scanner button returned zero candidates. Nothing. I asked the AI what was wrong, and it explained that the problem was in something called mockMarket.ts, where a seeded random-number generator had a broken hash function returning a float instead of an integer.

I did not know what that meant.

I still barely know what that means.

All I knew was that I clicked the button and the thing that was supposed to find trading opportunities found none. The AI could explain the plumbing, but I was standing there with a dry sink.

Then there was the $124,000 phantom gain.

That one genuinely scared me. The app showed a profit and loss number that was not real. In a normal app, a bad number is embarrassing. In a trading-related app, a bad number feels dangerous. The whole point of MyTradePilot is trust in numbers. If the numbers lie, even by accident, the rest of it does not matter.

So when I got to the point of paying for data and infrastructure, I did not feel like a bold founder. I felt like a guy staring at a checkout screen, wondering if he was being responsible or just getting in deeper.

There is also something humbling about paying for tools when you do not fully understand the tools.

I have seen warnings with names like SUPA_anon_security_definer_function_executable and search_path_mutable. Those are not words that make a non-technical person feel calm. They sound like warning labels on a machine in a room I should not be allowed into.

I had to ask the AI to fix them. Then I had to trust that the fix was real.

That is the uncomfortable part of building this way. The AI can do things I cannot do. That is the whole reason this project exists. But it also means I sometimes have to make decisions before I fully understand every layer underneath them.

The first monthly bill made that tension obvious.

Before the bill, I was playing with a possibility. After the bill, I was maintaining something. Even if it was small. Even if nobody else cared yet. Even if the app still had rough edges and three different names scattered across the site at one point — MyTradePilot on the homepage, TradePilot Academy on the learn page, and old QuantisAI bits still hanging around like socks under a bed.

A bill has a way of cutting through the fantasy.

It made me ask better questions. Do I actually need this right now? What does it power? What breaks if I cancel it? Is this helping the scanner become more trustworthy, or am I just buying the feeling of progress?

That last one stung a little.

Because I know I can confuse motion with progress. The logo adventure proved that. I can polish, rename, tweak, and rearrange until I feel productive. But paying for the scanner to run on better data was different. It was not decoration. It was connected to the core promise of the app.

Find candidates. Show the math. Don’t make things up.

The tiny win came later, when the scanner finally returned real candidates and the P&L math checked out against my spreadsheet. That did not make me fearless. I am still worried that a real user could lose real money because of a bug I did not catch. That fear is different from “the app crashes.” It sits heavier.

But the bill did something useful. It made me stop treating MyTradePilot like a toy I could abandon without consequence.

I still believe in small honest steps. I would tell anyone starting today not to try to build the whole thing in one prompt. Ask for one piece. Test it. Question it. Fix it. Then move on.

And maybe pick the name first.

Definitely stop redesigning the logo sooner than I did.

The first bill did not make the project successful. It just made it real enough that I had to be honest with myself.

The takeaway: a recurring bill is not just an expense — it is a commitment.