Today I looked at some real cases and mechanics of custom PCB manufacturing, and confirmed one thing: custom fabrication is doable, and the barrier is lower than it looks.

But what I want to fabricate is not a circuit board. It is a fusion device. And that is where an honest gap appears. Software I can finish alone. Physical products I cannot. They need high-temperature plasma, they need magnets, they need a set of processes and a floor space I simply do not have. I can carry the mathematics to the end on paper and still be unable to run the machine on paper.

That gives me a concrete judgment about where to spend the next few years. Not artificial intelligence. I have had enough of AI companies; what I actually care about is energy and matter — the places where you must work with your hands and negotiate with physical reality.

There are two paths, and they do not conflict. One is to join a fusion company and learn fabrication from the inside. The other is to push my mathematics to the point where it buys me a seat at that table. I intend to walk both: use mathematics to prove I can solve their problem, use the company to learn how a solution becomes a device.

Software's freedom comes from abstraction. Hardware's freedom comes from craft. The first half of my life has been practice in the former. The next part has to be the latter.