During EA Global San Francisco 2017, there was a panel discussion called "Celebrating Failed Projects." At one point, Nathan Labenz, the moderator, asks, "What are some projects that you guys are harboring in the backs of your respective minds that you'd love to see people undertake even if, and maybe especially where, the chance of ultimate success might be pretty low?" In response, Anna Salamon says, "There's a set of books that pretty often change people's lives, especially 18 year old type people's lives, hopefully in good directions. I think it would be lovely to make a list of five of those books and make a list of all the smart kids and mail the books to the smart kids. This has been on the list of obvious things to do for the last ten years but somehow nobody has ever done it. I didn't do it. I don't know. I really wish someone would do it. I think it would be really high impact."
It seems that the following five books are popular in the EA community:
1. Doing Good Better by William MacAskill
2. 80,000 Hours by Benjamin Todd and the 80,000 Hours Team
3. The Life You Can Save by Peter Singer
4. Animal Liberation by Peter Singer
5. Superintelligence by Nick Bostrom
However, I doubt that Salamon meant to limit the selection to books related to effective altruism. If you could choose five books on any topic, which five would you choose?
Back in ~2014, I remember doing a survey of top-contributing MIRI donors over the previous 3 years and a substantial fraction (1/4th?) had first encountered MIRI or EA or whatever through HPMoR. Malo might have the actual stats. It might even be in a MIRI blog post footnote somewhere.
But w.r.t. to research impact, someone could make a list of the 25 most useful EA researchers, or the 15 most useful "AI safety" researchers, or whatever kind of research you most care about, and find out what fraction of them were introduced to x-risk/EA/rationality/whatever through HPMoR.
I don't have a good sense for the what the net impact is.
Re top MIRI donors, there is a 2013 in review post that talks about a survey of "(nearly) every donor who gave more than $3,000 in 2013" with four out of approximately 35 coming into contact via HPMoR. (Not to imply that this is the survey mentioned above, as several details differ.)