> In business, science, philosophy - the first objective is to define your goals. How many times have you seen someone actually write out their utility functio

> In business, science, philosophy - the first objective is to define your goals. How many times have you seen someone actually write out their utility function? Or agree on a single number on which the results of an A/B test will be based, regardless of other metrics? Defining your goals is an admirable objective, but it's not commonly done to the degree you're proposing. All else equal, we want to preserve the maximum number of human lives, prevent the maximum damage to the economy possible, and provide maximum human freedom. Quantitatively defining the tradeoffs there is basically impossible. What's much more common is proposing a few options, and listing out how each options affects the metrics you care about, then deciding an option. Said more explicitly, it's easier subjectively to pick a point on a rigorously found pareto frontier than to subjectively define a utility function to be rigorously optimized.