I think there's also a near/far thing going on. I can't find it now, but somewhere in the rationalist diaspora someone discussed a study showing that people will donate more to help a smaller number of injured birds. That's one reason why charity adds focus on 1 person or family's story, rather than faceless statistics.

Combining this with what you pointed out, maybe a fun place to take the discussion would be to suggest that we start with a specific one of our friends. "Exactly. Let's start with Bob. Alice next, then you. I'll volunteer to go last. After all, I wouldn't want you guys to have to suffer through the loss of all your friends, one by one. No need to thank me, it is it's own reward."

EDIT: I was thinking of scope insensitivity, but couldn't remember the name. It's not just a LW concept, but also an empirically studied bias with a Wikipedia page and everything.

However, I mis-remembered it above. It's true that I could cherry pick numbers and say that donations went down with scope in one case, but I'm guessing that's probably not statistically significant. People are probably willing to donate a little more, not less, to have an impact a hundred times as large. Perhaps there are effects from misleading vividness at a small scale, as I imply. However, on a large scale, the slope is likely largely positive, even if just barely.