> If your operations have a 90% survivability rate it could still be considered a success despite those 10% they end up killing, because 90% is a lot better than 0%.

> Meanwhile a tired pilot is more binary, it either can have 100% passenger survivability if things go well or 100% fatality if things go tits up, meaning the risk are too high to take chances.

This isn't how the math works. A tired pilot either kills or doesn't kill their passengers on a given flight the same way that a tired doctor either kills or doesn't kill their patient in a given operation. In both cases it's 100% survive or 0% survive.

Pilots aside, we also have laws about keeping truck drivers from driving too many hours, and accidents involving drowsy truck drivers are unlikely to have fatality counts measured in the hundreds.

The difference between doctors and pilots/truckers isn't in the amount of risk involved, it's that surgeries are expected to have a non-zero fatality rate. A doctor can do their job perfectly and still lose a patient, so it's harder to prove that the fatality rate would be lower if we gave surgeons more rest. When a truck driver falls asleep at the wheel and kills someone it's obvious because they did something provably illegal or unsafe right before the crash. When a surgeon fails it's a lot harder to prove it was preventable.