I am curious about how best to model utilitarianism. Several possible models spring to mind:
One is that it is a signalling device: selfishness is bad, utilitarianism is ultimate unselfishness - and so it signals goodness.
Another is that it is a manipulation device. Some utilitarianism advocates run causes that benefit from donations.
Another point of interest is exactly how bad utilitarianism is for the individual. One might think - like many memetic hijckings - it would typically lead to sterility. However, the famous utilitarian Peter Singer - for example - is married, with three daughters and three grandchildren. Similarly, Yudkowsky's fastest possible escape route wound up landing him a girlfriend. What gives? Is talking about helping others actually just a thinly-veiled way of helping yourself?
Are you talking about "assigning everyone's welfare equal utility" utilitarianism or "models one's world-state preferences and subjective preferences using a utility function?" I think that you could be the latter and a complete egoist at the same time. (I am the latter, not the former.)
It seems to me that usually, when someone says "ethics" on lesswrong, ey usually means something along the lines of decision theory. When an average person says "ethics", ey is usually referring to a system of intuitions and social pressures designed to influence the behavior of members of a group. I think that a lot of the disagreement regarding ethics (i.e. consequentialism vs deontology) is rooted in a failure to properly distinguish between decision theory and what society pressures people to do. Most lesswrong users probably understand the distinction fairly clearly, but we only ever talk about decision theory. Why don't we talk about the social meaning of ethics?