Re: trolley problems and torture:

I seem to remember reading somewhere, I think it was something Daniel Dennett said, about the value of having philosophers willing to explore ideas that are (and maybe should be) taboo for ordinary people.

Take Peter Singer, for example. I don't buy the whole standard consequentialist package in ethics, but I really like Peter Singer. And he says things that are really shocking to many people, for example arguing that infanticide is often morally OK. But I suspect being willing to consider shocking ideas like that may be a prerequisite for being able to make progress on certain really important topics (see Singer's ideas about animal rights, charity, and some areas of medical ethics). Not everyone needs to be Peter Singer, but having a few Peter Singers - even a whole blog community of them - seems really valuable.

A couple other points: on torture, I don't think it's exactly being taken lightly. Rather, I suspect the reason it's used as an example is precisely because it an archetypal example of a really horrible thing.

As for seeming un-empathic, I don't think it's just rationality signaling. There's an issue that when you're making decisions that effect huge numbers of people, being too driven by your feelings about one case can lead to decisions that are really bad for the other people involved and that you wouldn't make if you really thought about it.