Risk

Almost every day, someone tells me they are worried about the risks of building my company around artificial intelligence. The concerns arrive in slightly different packaging but rarely vary in substance. Geopolitical fragility. The funding pressures on the frontier labs. The supposed plateau in model performance. The energy bill. The cultural resistance of healthcare to anything new. The list is delivered with the gravity of someone who believes they are saying something I have not heard, and is then offered as a kind of unanswered question, hanging in the air, expecting me to flinch.

I never do. Not because the concerns are silly. Some are quite serious. But because the framing assumes that risk is a property of this particular bet, when in fact risk is the price of admission to anything worth doing. To build a company at all is to accept that most companies fail. To work in healthcare is to accept that institutions move slowly and forgive nothing. To wake up in a city, board a plane, sign a lease, marry someone, raise a child, is to make a decision under uncertainty (I haven't quite yet reached the last two). The question is never whether risk exists. It exists. The question is whether it has been priced correctly.

Poker taught me to think this way before I knew I needed the lesson. The amateur at the table confuses outcome with decision, and walks home angry when a correct play loses to the river. The professional folds when the math demands a fold, raises when the math demands a raise, and accepts with something close to indifference that the cards will sometimes betray him. He is not fearless. He is calibrated. He knows the difference between variance, which is weather, and edge, which is the only thing he actually controls. Most arguments about risk in business are really arguments about variance dressed up as insight, delivered by people who have not done the arithmetic.

So what is the arithmetic here? On one side, an unusually compressed window in which software has begun, for the first time, to read, listen, speak, reason about, and act on the texture of clinical work. On the other, a healthcare system in which the front desk is quietly collapsing, calls go unanswered for hours, staff turn over twice a year, and patients drift to whichever practice picks up the phone. The status quo is not a neutral baseline. It is itself a position with miserable expected value, paid for in exhausted receptionists, lost patients, health complications, and revenue that walks out the door before anyone notices it leaving. To stand still in this environment is not caution. It is a different bet, made silently, and usually a worse one.

None of this excuses recklessness. We are running operations for clinics that trust us with the first impression a patient ever forms of their care. That responsibility is real, and it shapes every technical and clinical decision we make. But responsibility is not the same thing as paralysis, and the people who confuse the two will spend their careers admiring problems they could have solved.

Risk usually arrives married to opportunity, and the marriage is rarely amicable. You take the upside by accepting the volatility, or you take neither. The founders, clinicians, and investors I admire have all made peace with this in their own way. They have stopped asking whether the table is safe and started asking whether the game is worth playing.

It almost always is. And besides, life is more interesting this way.