Choosing ends

Systems optimize objectives; someone has to have set them, and setting them is the act of deciding what is worth wanting. Preferences can be inferred from behaviour and values can be argued about with a model, but a chosen end is one a person will live with, revise and be judged for. Delegating that choice is itself a choice, and it stays on the account of the one who made it. This entry is held not because machines cannot rank options, but because an end without someone who bears it is a parameter.