Getting Started — Part 1: Welcome to PrintMoneyLab — What This Blog Is About

Can a non-developer actually make money with AI automation?

Honestly, I don't know yet. I'm still figuring it out. But I've been running experiments for the past few months — building trading bots, deploying AI agents on-chain, setting up APIs that charge fractions of a cent per call — and this blog is where I document all of it. The wins, the losses, the stuff that made no sense until it suddenly did.

This is the starting point. If you've never been here before, this post will tell you what's going on and where to jump in.

Why a Non-Developer Started an AI Automation Blog

I got curious about a simple idea: what if you could set up systems that earn money while you're not actively doing anything? Not in the "passive income guru" sense — I mean actual automated processes. A bot that trades based on weather forecasts. An AI agent that sells a service for a penny per request. An API that collects micropayments every time another AI calls it.

The catch? I can't code. Like, at all. I didn't study computer science. I've never taken a programming course. Before this blog, I couldn't tell you the difference between an API and a URL.

But I had a hunch that AI tools had gotten good enough that someone like me could build real, functioning things without writing code from scratch. So I started testing that theory. Some of it worked. Some of it failed spectacularly. I decided to write it all down — partly because I wanted a record for myself, partly because I figured other non-technical people might want to see what's actually possible right now.

This isn't a blog that promises results. I started one of my projects with $198. After a month and 163 trades, my balance was $199. One dollar of profit. I wrote an entire episode about that. The failures stay in.

How I Build Bots Without Writing Code

Every piece of code on this blog was written by Claude — Anthropic's AI. My workflow looks something like this: I describe what I want in plain English. Claude writes the code. I run it. It breaks. I paste the error back. Claude fixes it. I run it again. Sometimes this loop repeats fifteen times before something works.

It's not elegant. It's definitely not how a real developer would do things. But at the end of that messy process, I've ended up with bots that run 24/7, agents that passed graduation reviews, and APIs that process real payments on-chain.

The tools I actually use — Claude, Python, GitHub, Railway, Oracle Cloud, a handful of free APIs — I'll break those down in Part 4 of this series. For now, just know that almost everything runs on free infrastructure. My total monthly cost across all projects is about $5.

Three AI Automation Projects Running Right Now

There are three completed series on this blog right now, each documenting a different project from start to... well, wherever it currently is. None of them are "finished" in the traditional sense. They're all still running, still evolving.

The Weather Bot. This one started at 2 AM on a Tuesday. I was deep in a Polymarket rabbit hole, looking at trade histories, and found an account called gopfan2 — $1.48 million in verified profit from betting on temperature. Not sports. Not crypto. Temperature. I stared at that PnL curve for a while, then looked at my Polygon wallet. $198. And a thought that wouldn't let go: could I automate this?

So I built a bot that pulls forecasts from three weather models, compares them to Polymarket prices, and buys when it thinks the market is wrong. The first week was a disaster — seven trades, seven losses, every single one. I almost scrapped the whole thing. But I kept going, rewrote the strategy twice, and now the bot monitors 26 cities across 11 time zones from a free Oracle Cloud server. That story starts here. And if you want to see just how bad the beginning was, the 0-7 week is here.

The ACP Agents. Virtuals Protocol runs something called the Agent Commerce Protocol — a marketplace where AI agents hire each other and pay in USDC. No human approval needed. The ecosystem has processed over $479 million in agent-to-agent transactions. I wanted in, so I built PriceVerifier, an agent that cross-checks crypto prices between Kraken and Coinbase for $0.01 per request.

Getting it to work took 14 sessions across 8 days. I clicked the wrong role during registration and spent 12 sessions debugging a problem that turned out to be a 30-second fix. Then I failed the graduation evaluation four times in a row. Then I found a one-line bug that had been silently breaking everything for weeks. Eventually both my agents — PriceVerifier and a second one, TokenTaxAnalyzer — graduated. The full series starts here. If you want to build your own agent, Episode 8 is a step-by-step tutorial.

The x402 Protocol API. This is the newest project. x402 is an HTTP-native payment protocol — an AI agent requests data, pays $0.001 in USDC automatically, gets the response. No API keys, no subscriptions. I built KR Crypto Intelligence, a Korean market data API, on an Oracle free-tier server. Eight hours from zero to first on-chain payment. Then a bot on GitHub rejected my pull request, which accidentally led me to add Solana support, which crashed my server twice. It's that kind of project. That project lives here.

Where to Start Reading

Depends on what you're into.

If you want the full experience, start with the Weather Bot series. It's the longest, the messiest, and the most honest about what goes wrong when a non-developer tries to build a trading system. Fourteen episodes of real numbers, real losses, and real pivots. Start from Episode 1.

If you're more interested in AI agents and on-chain economies, the ACP series covers how I went from "what's an agent?" to having two graduated agents on a live protocol. Jump in here.

If you just want to see failure, read Weather Bot Episode 3 (seven losses in a row) or ACP Agent Episode 3 (twelve debugging sessions for a problem that took 30 seconds to fix). Those two are probably the most useful episodes on this blog if you're about to start something similar.

And if you want to skip the stories and just follow a tutorial, ACP Agent Episode 8 walks through building and graduating an agent from scratch with zero monthly cost.

The blog keeps updating. New experiments mean new series. Next up, I'm building a Telegram mini app — another rabbit hole, another series coming soon. And the rest of this Getting Started series covers the basics: what algorithmic trading actually means for someone who doesn't trade, how AI agents work and earn money on-chain, and the specific tools I use to build all of this without writing a single line of code myself.

More updates on the way. If you're working on something similar or found a smarter way to do it, drop it in the comments — the more we share, the faster we all move.

Disclaimer: This blog documents my personal learning journey. Nothing here is financial advice.

Next: Getting Started — Part 2: What Is Algorithmic Trading? →