Priority can be mathematically stated as maximizing the minimum expected utility across a set of people's utility functions instead of maximizing the additive or average expected utility.
It still suffers from the utility monster who can make trivial inconveniences as numerically terrible as the worst life imaginable for other people, dragging the world down to a merely comfortable level for everyone else, which may frustrate the desire to thrive and grow in those other people. It caps the effect at not letting anyone else be worse off, which seems desirable over alternatives. It potentially leaves a lot of good on the table to avoid the risk of a lot of harm.
It sounds like a good initial optimization strategy until we figure out how to unify disparate agents' utility functions into a global optimization problem if that turns out to be possible.
It still suffers from the utility monster who can make trivial inconveniences as numerically terrible as the worst life imaginable for other people, dragging the world down to a merely comfortable level for everyone else, which may frustrate the desire to thrive and grow in those other people. It caps the effect at not letting anyone else be worse off, which seems desirable over alternatives. It potentially leaves a lot of good on the table to avoid the risk of a lot of harm.
It sounds like a good initial optimization strategy until we figure out how to unify disparate agents' utility functions into a global optimization problem if that turns out to be possible.