Much less scary.

"Save the world" is a very high-level goal, and I don't know how to do it. Your procedure is straightforward. I just need to invoke it when I recognize there's a choice. Resisting temptation is much easier (not just simpler, easier) than deciding whether you're being tempted.

Also, that's not actually your goal. You don't rob banks.

Dunno - to me they sound almost equivalent (except that you have no other motivations at all, and I'm not sure I can honestly say that about myself).

In any case, I'm not sure what sounds scary. It's all the people who don't seem to want to improve the world in any way at all that scare me.

If I have a choice between actions, and one of them is more likely to save the world than the other, I will take the one that is more likely to save the world.

There are certain actions for me such that the impact that they have on the probability that the world will be saved is insignificant enough that such impact is overwhelmed by the amount of immediate fun that they will generate, so an action that generates lots of immediate fun may be more desirable than one which increases the chance that the world will be saved by a really-super-small amount. Are you saying that for EV_Eliezer, there is no increase small enough in the chance that the world will be saved such that a huge amount of immediate fun is of greater terminal utility?

Approximately what proportion of your actions (or time spent, if that's easier to compute) have a clear chance of contributing to saving the world?