A few pieces you can actually read.

Concept is free. Execution is the edge. The concept is where the views are. The execution is where the effort is. So I do the part they cut, in the open, receipts on the table.

An afternoon. One operator. Four disciplines. In one sitting: a rental pro-forma built from raw escrow screenshots, a multi-year tax-lever campaign on an actual calendar, a four-path net-worth model, and a board-grade deck holding it together. Raw inputs in, decision-grade artifacts out. Ten years ago that's four specialists, weeks of coordination, and a five-figure bill. Every tax and entity call still routes to a CPA and an attorney to bless, because AI finds and models, the licensed human signs. That routing isn't a caveat. It's the spine. Illustrative numbers, real method. The output is the proof.

Does this thing CREATE value, or just CAPTURE it? I've watched the same axis play out at every scale — a deal, a team, a whole sector. The create side is almost always the better business, not the charity option. Bigger market, a loyalty moat instead of a lock-in moat, capital and talent that actually want to be there. Most people never name the axis. Naming it is half the edge. One portable question for work, money, people, and the world.

Healthcare admin is the purest extraction machine in the economy. Almost none of it touches a patient. I spent thirty years inside that back office. I helped build parts of the machine, and then I built the AI that started removing it. The care was always the create side. The admin is the tax, and I spent three decades learning exactly where it hides. Applied AI in revenue cycle and operations, the billing and admin machine. Not clinical decisions. I don't blur that line, and you should distrust anyone who does. Bring me the single most extractive process in your operation, the one that eats the most hours and touches the fewest patients. I'll show where AI actually removes it. A workbench, not a pitch.

I build my company with an AI chief of staff. When it screws up, I call it a muppet. It'll do the wrong thing exactly as confidently as the right one. The only thing that catches it is thirty years of knowing what doesn't smell right. That's not a knock on the tool. That's the entire case for how you actually run a business with AI. The AI got fast. The judgment got more valuable, not less. Catching the muppet isn't overhead. It's the job.

I ran four versions of my next stretch of life, each modeled out on one honest engine. The dollars weren't the point. When the required game is already won, when the floor is covered, every question quietly stops being "will I make it" and becomes "how much more, and what am I trading for it." Those are completely different questions, and most people never notice they've swapped between them. I build from optionality, on my pace. No race, no revenue pressure. Abundance, not scarcity. The dollars aren't the point. The framing is.