The other thing is that the actual standard of the "quantitative case" put forward is hugely variable anyway. With sufficiently inventive fantasy numbers, theat

The other thing is that the actual standard of the "quantitative case" put forward is hugely variable anyway. With sufficiently inventive fantasy numbers, theatre enthusiasts could probably make an EA case of the form: "what if a person was so inspired by watching my amateur dramatic production of 2001: A Space Odyssey, that they solved all Unfriendly AI problems?". Sure, it's unlikely, but the rewards benefit an almost infinite number of people! In practice, I think people coming from a starting point of "I'm going to focus on solving friendly AI because it fits with my interests and computer science background and I think I'll be able to help the process along a bit" are being more honest with themselves than people dressing up their preferences in magical utilitarian numbers which conclude they're more efficient than merely feeding people or giving them goats...