The Wrong Jurisdiction
The plan was technically correct. It was just valid in the wrong country.
About 25 years ago I spent summers on my father's ranch.
My uncle was there. Tractors. Horses. About a hundred deer in the fields. A lake we swam in multiple times a day. Sauna in the evenings. Berries you just picked. Planting hectares of trees together. The whole family. Late warm evenings where every day brought something new. The kind of air that does not exist near a city.
My uncle has lived in Norway for almost 20 years now. He works in construction. Heavy machines, 50 tonnes and more. He found something there that fits him. He is happy.
I visited once. Summer. Waterfalls. Green that does not look real. I remember thinking: maybe someday I leave the desk. Maybe someday I get back to something physical. Maybe someday I am there with him, and some version of that early 2000s sun comes back to my face.
It was not a serious plan. But it was a real one.
Late December 2025, I explained the idea to ChatGPT.
The goal was clear: get a tractor license, be useful in Norway, work with my uncle someday. I described the context. ChatGPT found a school in Latvia, outlined the steps, gave me a plan. It looked complete. I enrolled.
January 2026. Theory: late evenings, Zoom, 21:00 to 23:00 on working days. Sundays from 09:00 to 15:00. Around 60 hours. Road rules, safety regulations, maintenance checklists. I used AI again to test myself on the harder questions.
I drove to Talsi more than once. First for the practical sessions with the school owner. Then for the final exam with the VTUA inspector.
First time in the cab, the instructor pointed at the parking lot: back the trailer in.
Nothing in 60 hours had prepared me for what the machine actually does when you try that.
It went wrong. Several attempts. I got through eventually. My classmate did not pass.
I still think I was lucky.
On February 6th, I received my tractor license.
Then I found out that Latvian tractor licenses are not valid in Norway.
The plan ChatGPT built was technically correct about Latvia. It said nothing about Norway. I had explained the goal clearly at the start, and I followed the output without verifying the one fact that made the whole thing matter.
More than 500 euros. Around 60 hours. Not one step closer to the actual goal.
I am still happy. It was interesting. I learned things. But the gap between what the plan promised and where I actually ended up is not a small detail. It is the whole story.
Some time before this, during an MBA session, a professor made the familiar argument. Every generation fears the new technology. Electricity. Tractors. Computers. Now AI. The cycle repeats. The fear is always misplaced.
The pattern is not new. Electricity got rules. Tractors got licenses. Phones got regulations. AI will too. Not because of fear. Because that is how systems catch up.
What those rules will test exactly, I do not know. Whether theory or capability. Whether local compliance or actual readiness for the environment you operate in.
For now, the window is open. Nobody is checking your jurisdiction yet.
I got my tractor license. I used AI to build the plan. The plan was wrong where it mattered. I am still happy about the whole thing.
Theory gives you a license. It does not give you capability. And a license issued in the wrong jurisdiction means nothing where it actually matters.
Enjoy the unregulated moment.
It probably will not last.