I Told AI to “Make a Game” — and It Built Something Smarter Than Me

I Started Using AI Because I Wanted Money

Let me be honest about my motivation: I wanted money.

Not for anything complicated. I wanted to travel, stay at nice hotels, eat good food, and live quietly alone. I live in rural Japan, where everything costs money — transportation, options, freedom. So I started using AI in 2024 as a way to change that.

Since then, I’ve been paying for ChatGPT, Google Gemini, Claude, and Midjourney. Lately, I’ve been focused on agentic AI — tools like Claude Code and Codex that don’t just answer questions, but actually build things.

I Wrote Zero Lines of Code. Here’s What Got Built.

A few days ago, I gave Claude Code a single instruction:

“Make a game. Something popular right now.”

Thirty minutes later, I had a Bitcoin mining simulator.

Click to mine BTC. Buy AI agents to automate your mining. Compete on a global leaderboard. A PayPal support button was included — without me asking.

I then added two requests: “Add a ranking system” and “Add a way to earn money from it.” Both were implemented immediately.

The game is live here: AI BITCOIN MINER v2.6.0

What I Actually Did

I created a GitHub repository and clicked an approval button when asked. That’s it. No code. No technical decisions. No real understanding of what was happening under the hood.

I Built It. I Don’t Understand It.

When I open GitHub, I see files and code. I still don’t know which file is the game, or what would happen if I changed something.

I made this thing, and I can’t fully explain how it works.

That feeling is strange. Not bad, exactly — but strange. The only proof that I was involved is that I expressed an intention. The AI did the rest.

The Feeling That Humans Might Not Be Necessary

This feeling has been growing since ChatGPT launched.

At first, I thought of AI as a tool — something I use. Now it feels more like I’m the one giving approval, and the AI is the one doing the thinking.

I recently started using models with extended reasoning — the kind that think deeply before responding. Sometimes I read the output and genuinely don’t understand what it’s saying. The gap between what I can think and what AI can think is widening in real time.

I don’t know if that’s frightening or exciting. Probably both.

Why I Keep Going Anyway

Because I want money. And because AI might be the most realistic path to it from where I am.

Living rurally in Japan means limited access to traditional opportunities. AI doesn’t care where you live. It doesn’t require a commute. If I can figure out how to use it well enough, maybe that changes something.

I’m not there yet. But I made a working game without writing a single line of code. That’s a start.

Try the game here: AI BITCOIN MINER