It is a helpful heuristic, but on some level, if you know how people are going to act, you should act assuming that they’d act that way.
It’s easier to notice large effects than small effects. It’s obvious that if everyone voted for a specific presidential candidate, that candidate would win. It’s less obvious that the marginal voter would do anything. Nonetheless, in order for the first case to be true, the average marginal voter has to be making a difference, so it might be better to look at it that way. But the votes aren’t uniformly distributed, and if you know the distribution, you can make a better case of whether or not to vote.
That being said, if you accept EDT or UDT, then you should assume p% of people will act like you, since p% of people do act like you, and will make whatever choice you make, for the same reasons.
— Louis C.K.
More robust solution would be to act in a way that, if p% of people acted that way, things would work out; for as low values of p as possible.
Because it is very unlikely that everyone will act in some way.
It’s a rephrasing of Kant’s categorical imperative.
It is a helpful heuristic, but on some level, if you know how people are going to act, you should act assuming that they’d act that way.
It’s easier to notice large effects than small effects. It’s obvious that if everyone voted for a specific presidential candidate, that candidate would win. It’s less obvious that the marginal voter would do anything. Nonetheless, in order for the first case to be true, the average marginal voter has to be making a difference, so it might be better to look at it that way. But the votes aren’t uniformly distributed, and if you know the distribution, you can make a better case of whether or not to vote.
That being said, if you accept EDT or UDT, then you should assume p% of people will act like you, since p% of people do act like you, and will make whatever choice you make, for the same reasons.
Seems like a really good way of being taken advantage of by a DefectBot.