We live in a time when you only need an idea and AI will build it for you. AI already writes well enough that it is hard to tell whether a person or a machine wrote something, and it produces passable drawings and code too. Many people get fooled by that plausibility. Many others have digested it into their own tool: a personal assistant, a knowledge engine, a spring of inspiration. I am one of the people who got swept along by that spring.
I am a manufacturing engineer. My job, quite literally, is telling machines that cut metal where to go. I cannot code at all. Still, repetitive tasks come up at work, and for those I use programs that AI built for me. I just point them at the job and let the macro run. I cannot read a single line of the code, but that is how I use it.
I like making things with my hands. Even when a finished product exists and someone has already made it, I have always wanted one thing that is entirely my own. I have one hour a day and almost no money.
At first I wanted to make money
The first time I really used AI was for automated stock trading. I had AI do everything: the strategy and the code. I only supplied the ideas.
The method went like this. Take past trades whose outcome is already known, extract a formula for “this is what would have worked”, and apply it as is to the next trading day. Run it on historical data and it hits every time. It looked convincing.
I ran it five times, three months each. Three of those runs used real money, two million won each. Every time I adjusted the formula a little, and it seemed to get a little better, but the fundamental problem never went away. A formula that was right about the past knows nothing about tomorrow. Honestly, the process itself was fun. Watching the numbers line up is a pleasure. But no money came out of it. I would have been better off buying an ETF and leaving it alone.
It always comes back to what everyone already knows
I learned one thing here. Ask AI how to make money and at first you get a plausible answer. Dig deeper and it always comes back to what everyone already knows: diversify, hold for a long time. Not because AI is stupid, but because there is no such thing as a fast way to make money. The one fooled by plausibility was not the AI. It was me.
So I decided to build something
It did not make money, but one thing became clear. With an idea, you can build something even if you cannot code. Whatever is missing, AI can fill in.
This time I am building a device that finds people without a phone signal. You pair two of them and each shows the direction and distance to the other. Over forty thousand of these have been sold overseas for festivals. I am going to try building one myself.
There is a reason. A friend of mine recently had a baby. When I was a child, I once lost my younger sister. Back then there was nothing. I thought that if you clip one of these on, you could at least find someone nearby even when they have wandered off. That is why distance matters.
The parts are ordered and should arrive around October 22. From then on I will post the build process as it happens. What worries me is range. I hope it reaches as far as possible, but if it does not, I will say so.
What this blog promises
- AI writes the code. I do the soldering and the fumbling. I will not hide that.
- I do not copy and paste AI answers. I write in my own words.
- Failures go up as they are. If it does not work, I will say it did not work.
- Every part cost is published.
Stay tuned to see what I end up building. Thank you.