the whole point of it is to lower the message-description-length of proposals that will improve overall utility
I thought the point was to help us model the things we care about more accurately and efficiently, which doesn’t require utilitarianism to be an appealing proximate goal (it just has to require caring about something which depends on objective reality).
I thought the point was to help us model the things we care about more accurately and efficiently, which doesn’t require utilitarianism to be an appealing proximate goal (it just has to require caring about something which depends on objective reality).