This topic is not really related to the things normally discussed here, but I think it's really important, and it might interest Less Wrongers, especially since many of us are interested in ethics and utility calculations that are essentially cost-benefit analyses. Bone marrow donation in the United States is managed by the National Marrow Donor Program. Because typing donors for matching purposes can be costly, they often require people signing up to donate to pay a registration fee, which probably prevents a lot of people from signing up. These costs are being covered until the end of the month by a corporate sponsor, which means that right now, all you need to do if you live in the US is go to http://marrow.org/Join/Join_Now/Join_Now.aspx and fill out a simple questionnaire. You will be sent a kit to collect a cheek swab, and then you will be entered into the donor database. Doing this does not require you to donate if a match comes up.
The reason I think this might interest Less Wrongers is that this is a really cheap way to improve the world. According to their website, about 1 in 500 potential donors are actually asked to donate, so registering doesn't actually make it all that likely that you will be asked to do anything more. If you ARE a match for someone who needs a donation, the cost to you is at most the temporary pain of marrow extraction (many donors are asked only for blood cells), whereas the other person’s chance to live is much improved. This looks like a huge net positive.
Unfortunately I only found out about this a few days ago, and it only occurred to me today that this might be a forum of people who would respond to the argument "you can make the world better at little cost to yourself." However, I ask that you go to the website and spend a few minutes signing up. This is like buying a 1 in 500 lottery ticket that SAVES SOMEONE’S LIFE. If the Singularity hits and an FAI can generate perfectly matched marrow for anyone who needs it from totipotency-induced cells, that will be wonderful, but this is a chance to make sure one more person gets there.
DISCLAIMER: This is awesome, a great source of fuzzies, and those who register are deserving of praise (I personally registered in the course of gathering data for the comments in this thread). The analysis below is done in the spirit of accurately understanding tradeoffs, and practice in thinking about do-gooding effectiveness. If the argument below convinces you not to register when you otherwise would have, please donate a few bucks to a more efficient charity, or a piggy bank until you think of something better, rather than simply cutting back on do-gooding.
Here's GiveWell's report on its top-rated charity in international health, Against Malaria Foundation:
You say that there is a 1 in 500 chance that one will be called on to donate bone marrow. ETA: the FAQ says it is 1 in 540. If one donates, it surely is not guaranteed to save the life of a young person with 40+ years of life (who would otherwise not get marrow, and would soon die with no other treatments working). A 10% (note: edited figure from 25% based on further Googling, and propagated changes) chance of saving such a life (or that expected value) seems reasonable, for a 1 in 5400 probability. Comparing in terms of direct life-saving, if it costs even 37 cents (in time and demands of registry, expected pain, expected donation hassle and recovery, additional testing, distraction, etc) one might do better by giving to AMF or some better charity.
According to the FAQ, benefits would be higher for US racial minorities (fewer donors to match against) and less for Americans of European ancestry.
Of course, saving the life of a rich person has other spin-off benefits (they may have more positive impact on the world thereafter than a potential malaria victim), and solidarity with other members of one's (rich) community is a perfectly understandable motivation.
Still, I am skeptical that this is near the efficiency frontier, even with the donors covering costs.
I reposted this article in Discussion with a small addendum. Since your arguments are all really relevant, please repost them there if you have the time/inclination, so that people can still see and think about them. I didn't modify the article myself because 1) I'm lazy and need to get to class and 2) I'm hoping people will sign up before they read any arguments against it, because I do think that this is a way for people to put time towards charity that they otherwise wouldn't. But I respect and am grateful for your deconstruction of the mathematics of charity going on here.