I'll build it myself

December 2025. He got on a call about turning PromptDNA into real software.

Somewhere in the middle of it, he heard his own voice say he would build most of it himself.

He was not qualified to say that. He is a computer science dropout. The last time he had written code with any seriousness he was failing the semester that ended that whole idea about himself, and the twenty years since had been twenty years of hiring the person who does the technical part.

He said it anyway. Then he sat with it for a beat and decided not to walk it back.

This is a thing he does, and it took him into his thirties to notice the pattern. He calls the shot out loud before any evidence exists, then spends the next stretch of his life making the sentence true. He did it on camera in February. He did it again here.

The next move is the one that actually mattered.

He had the money to hire it out. September had cleared $60K in a month at roughly 95 percent margins, and paying a developer to build his software was the obvious, sane, twenty-years-of-experience answer.

He set up coding coaching instead.

The coach did not sell him the shortcut, which is the only reason this entry exists. It would rob you if I built it for you, he said.

Rob. Not it would be cheaper if you learned it. Not you would understand it better. Rob — as in something would be taken from you that you would not get back.

He was right, and the proof is every company John had built before this one. All of them ran on other people's hands. When the hands left, the company left with them. The ceiling was never the ambition and it was never the ideas. The ceiling was that he could describe the machine in perfect detail and could not touch it.

December was the decision. Six weeks later he opened a terminal.