lukeprog

• lukeprog 8 Mar 2019 1:31 UTC

  20 points

  in reply to: lukeprog's comment on: After one year of ap­ply­ing for EA jobs: It is re­ally, re­ally hard to get hired by an EA organisation

  BTW my “re­flec­tions on the 2018 RA re­cruit­ing round” post is now up, here.

• lukeprog 4 Mar 2019 4:22 UTC

  31 points

  in reply to: anonymous_ea's comment on: After one year of ap­ply­ing for EA jobs: It is re­ally, re­ally hard to get hired by an EA organisation

  I agree that if it’s true that “many EAs feel that ei­ther they’re work­ing at a top EA org or they’re not con­tribut­ing much,” then that is much worse than any­thing about ap­pli­ca­tion time cost and ur­gently needs to be fixed. I’ve never felt that way about EA org work vs. al­ter­na­tives, so I may have just missed that this is a mes­sage many peo­ple are get­ting.

  E.g. Scott’s post also says:

  Should also ac­knowl­edge the pos­si­bil­ity that “tal­ent-con­strained” means the world needs more clean meat re­searchers, malaria vac­cine sci­en­tists, and AI pro­gram­mers, and not just generic high-qual­ifi­ca­tion peo­ple ap­ply­ing to EA or­ga­ni­za­tions. This wasn’t how I un­der­stood the term but it would make sense.

  …and my re­ply is “Yes, tal­ent-con­strained also means those other things, and it’s a big prob­lem if that was un­clear to a no­tice­able frac­tion of the com­mu­nity.”

  FWIW I sus­pect there’s also some­thing a bit more sub­tle go­ing on than overly nar­row mi­s­un­der­stand­ings of “tal­ent-con­strained,” e.g. some­thing like Max Daniel’s hy­poth­e­sis.

• lukeprog 2 Mar 2019 21:49 UTC

  10 points

  in reply to: EA applicant's comment on: After one year of ap­ply­ing for EA jobs: It is re­ally, re­ally hard to get hired by an EA organisation

  Thanks, that all makes sense to me. Will think more about this. Also still cu­ri­ous to hear replies from oth­ers here.

• lukeprog 2 Mar 2019 19:09 UTC

  9 points

  in reply to: anonymous_ea's comment on: After one year of ap­ply­ing for EA jobs: It is re­ally, re­ally hard to get hired by an EA organisation

  Thanks for +1ing the above com­ment. I’d be keen to hear your re­ply to this com­ment, too.

• lukeprog 2 Mar 2019 19:07 UTC

  19 points

  in reply to: anonymousthrowaway's comment on: After one year of ap­ply­ing for EA jobs: It is re­ally, re­ally hard to get hired by an EA organisation

  Thanks for shar­ing. As some­one who spends a lot of time try­ing to fill EA meta/​longter­mist tal­ent gaps — e.g. by man­ag­ing Open Phil RA re­cruit­ing, helping to match the strongest ap­pli­cants we don’t hire to other open­ings, and by work­ing on field-build­ing in AI strat­egy/​policy (e.g. CSET) — hear­ing sto­ries like yours is un­nerv­ing.

  What changes to the land­scape, or hiring pro­cesses, or what­ever, do you think would’ve made the most differ­ence in your case?

  I’m also cu­ri­ous to hear your re­ac­tion to my com­ment el­se­where about available paths:

  There are lots of ex­cit­ing things for EAs to do be­sides “ap­ply to one or more of the 20 most com­pet­i­tive jobs at ex­plic­itly EA-mo­ti­vated em­ploy­ers,” in­clud­ing “keep do­ing what you’re do­ing and en­gage with EA as an ex­cit­ing hobby” and “ap­ply to key po­si­tions in top-pri­or­ity cause ar­eas that are on the 80,000 Hours Job Board but aren’t at one of a hand­ful of ex­plic­itly EA-mo­ti­vated orgs” and “do earn to give for a while while gain­ing skills and then maybe tran­si­tion to more di­rect work later or maybe not,” as well as other paths that are spe­cific to par­tic­u­lar pri­or­ity causes, e.g. for AI strat­egy & policy I’d be ex­cited to see EAs (a) train up in ML, for later work in ei­ther AI safety or AI strat­egy/​policy, (b) fol­low these paths into a US AI policy ca­reer (esp. for US cit­i­zens, and esp. now that CSET ex­ists), and (c) train up as a cy­ber­se­cu­rity ex­pert (I hope to say more later about why this path should be es­pe­cially ex­cit­ing for AI-in­ter­ested EAs; also the worst that hap­pens is that you’ll be in ex­tremely high de­mand and highly paid).

• lukeprog 1 Mar 2019 22:09 UTC

  63 points

  in reply to: Throwaway8.1's comment on: After one year of ap­ply­ing for EA jobs: It is re­ally, re­ally hard to get hired by an EA organisation

  Sorry to hear about your long, very difficult ex­pe­rience. I think part of what hap­pened is that it did in fact get a lot harder to get a job at lead­ing EA-mo­ti­vated em­ploy­ers in the past cou­ple years, but that wasn’t clear to many EAs (in­clud­ing me, to some ex­tent) un­til very re­cently, pos­si­bly as re­cently as this very post. So while it’s good news that the EA com­mu­nity has grown such that these par­tic­u­lar high-im­pact jobs can at­tract tal­ent suffi­cient for them to be so com­pet­i­tive, it’s un­for­tu­nate that this change wasn’t clearer sooner, and posts like this one help with that, albeit not soon enough to help miti­gate your own 1.5 years of suffer­ing.

  Also, the thing about some peo­ple not hav­ing run­way is true and im­por­tant, and is a ma­jor rea­son Open Phil pays peo­ple to take our re­mote work tests, and does quite a few things for peo­ple who do an in-per­son RA trial with us (e.g. salary, health benefits, mov­ing costs, sev­er­ance pay for those not made a sub­se­quent offer). We don’t want to miss out on great peo­ple just be­cause they don’t have enough run­way/​etc. to in­ter­act with our pro­cess.

  FWIW, I found some of your com­ments about “elite cul­ture” sur­pris­ing. For con­text: I grew up in ru­ral Min­nesota, then dropped out of coun­sel­ing psy­chol­ogy un­der­grad at the Univer­sity of Min­nesota, then worked at a 6-per­son com­puter re­pair shop in Glen­dale, CA. Only in the past few years have I be­gun to some­what reg­u­larly in­ter­act with many peo­ple from e.g. top schools and top tech com­pa­nies. There are as­pects of in­ter­act­ing with such “elites” that I’ve had to learn on the fly and to some de­gree am still not great at, but in my ex­pe­rience the cul­ture in those cir­cles is still pretty differ­ent from the cul­ture at ma­jor EA-mo­ti­vated em­ploy­ers, even though many of the staff at EA-mo­ti­vated em­ploy­ers are now peo­ple who e.g. grad­u­ated from schools like Oxford or Har­vard. For ex­am­ple, it’s not my ex­pe­rience that peo­ple at ma­jor EA or­ga­ni­za­tions are as effu­sively pos­i­tive as many peo­ple in non-EA “elite” cir­cles are. In fact, I would’ve de­scribed the cul­ture at the EA or­ga­ni­za­tions I in­ter­act with the most in sorta op­po­site terms, in that it’s hard to get them ex­cited about things. E.g. if you tell one of my Open Phil RA col­leagues about a new study in Na­ture on some topic they care about, a pretty com­mon re­ac­tion is to shrug and say “Yeah but who knows if it’s true; most of the time we dig into a top-jour­nal study, it com­pletely falls apart.” Or if you tell peo­ple at most EA orgs about a cool-sound­ing global health or poverty-re­duc­tion in­ter­ven­tion, they’ll prob­a­bly say “Could be in­ter­est­ing, but very low chance it’ll end up look­ing as cost-effec­tive as AMF or even GiveDirectly upon fur­ther in­ves­ti­ga­tion, so: meh.” Also, EA-mo­ti­vated em­ploy­ers are gen­er­ally not as “cre­den­tial­ist,” in my ex­pe­rience, as most “elite” em­ploy­ers (per­haps ex­cept for tech com­pa­nies).

  Fi­nally, re: “you never know for sure if it’s not just perfect mer­i­toc­racy cor­rectly fil­ter­ing [cer­tain peo­ple out].” I can’t speak to your case in par­tic­u­lar, but at least w.r.t. Open Phil’s RA re­cruit­ing efforts (which I’ve been man­ag­ing since early 2018), I think I am sure it’s not a perfect mer­i­toc­racy. We think our ap­pli­ca­tion pro­cess prob­a­bly has a high false nega­tive rate (i.e. re­ject­ing peo­ple who are ac­tu­ally strong fits, or would be with 3mo of train­ing), and it’s just very difficult to re­duce the false nega­tive rate with­out also greatly in­creas­ing the false pos­i­tive rate. Just to make this more con­crete: in our 2018 RA hiring round, if some­body scored re­ally well on our stage-3 work test, we typ­i­cally thought “Okay, de­cent chance this per­son is a good fit,” but when some­body scored medium/​low on it, we of­ten threw up our hands and said “No clue if this per­son is a good fit or not, there are lots of rea­sons they could’ve scored poorly with­out ac­tu­ally be­ing a poor fit, I guess we just don’t get to know ei­ther way with­out us and them pay­ing in­fea­si­bly huge time costs.” (So why not just im­prove that as­pect of our work tests? We’re try­ing, e.g. by con­tract­ing sev­eral “work test testers,” but it’s harder than one might think, at least for such ill-defined “gen­er­al­ist” roles.)

• lukeprog 28 Feb 2019 23:48 UTC

  12 points

  in reply to: Max_Daniel's comment on: After one year of ap­ply­ing for EA jobs: It is re­ally, re­ally hard to get hired by an EA organisation

  Sounds plau­si­ble. E.g. I’m pro “train up as a cy­ber­se­cu­rity ex­pert” but I know oth­ers have ad­vised against.

• lukeprog 28 Feb 2019 23:47 UTC

  44 points

  in reply to: EA applicant's comment on: After one year of ap­ply­ing for EA jobs: It is re­ally, re­ally hard to get hired by an EA organisation

  Thanks for shar­ing your com­ment about per­son­al­ized in­vi­ta­tions, that’s in­ter­est­ing. At Open Phil, al­most all our per­son­al­ized in­vi­ta­tions (even to peo­ple we already knew well) were only lightly per­son­al­ized. But per­haps a no­tice­able frac­tion of peo­ple mis­per­ceived that as “high chance you’ll get the job if you ap­ply,” or some­thing. The Open Phil RA hiring com­mit­tee is dis­cussing this is­sue now, so thanks for rais­ing it.

• lukeprog 28 Feb 2019 21:23 UTC

  26 points

  in reply to: Throwaway8's comment on: After one year of ap­ply­ing for EA jobs: It is re­ally, re­ally hard to get hired by an EA organisation

  Sorry to hear how much mis­ery you’ve ex­pe­rienced. I’m cu­ri­ous to ask a fol­low-up ques­tion, but feel free to ig­nore if you aren’t com­fortable an­swer­ing.

  In par­tic­u­lar, I’m won­der­ing whether “make [EA] my ca­reer” feels ~iden­ti­cal (to you) to “work at a hand­ful of ex­plic­itly EA-mo­ti­vated em­ploy­ers.” If it does, then maybe the mes­sag­ing or en­ergy or some­thing in the EA com­mu­nity is pretty far from what I think it should be, which is more like what I said in an­other com­ment:

  There are lots of ex­cit­ing things for EAs to do be­sides “ap­ply to one or more of the 20 most com­pet­i­tive jobs at ex­plic­itly EA-mo­ti­vated em­ploy­ers,” in­clud­ing “keep do­ing what you’re do­ing and en­gage with EA as an ex­cit­ing hobby” and “ap­ply to key po­si­tions in top-pri­or­ity cause ar­eas that are on the 80,000 Hours Job Board but aren’t at one of a hand­ful of ex­plic­itly EA-mo­ti­vated orgs” and “do earn to give for a while while gain­ing skills and then maybe tran­si­tion to more di­rect work later or maybe not,” as well as other paths that are spe­cific to par­tic­u­lar pri­or­ity causes, e.g. for AI strat­egy & policy I’d be ex­cited to see EAs (a) train up in ML, for later work in ei­ther AI safety or AI strat­egy/​policy, (b) fol­low these paths into a US AI policy ca­reer (esp. for US cit­i­zens, and esp. now that CSET ex­ists), and (c) train up as a cy­ber­se­cu­rity ex­pert (I hope to say more later about why this path should be es­pe­cially ex­cit­ing for AI-in­ter­ested EAs; also the worst that hap­pens is that you’ll be in ex­tremely high de­mand and highly paid).

• lukeprog 28 Feb 2019 21:22 UTC

  10 points

  in reply to: Milan_Griffes's comment on: After one year of ap­ply­ing for EA jobs: It is re­ally, re­ally hard to get hired by an EA organisation

  Nearly all of Open Phil’s RA hiring pro­cess is fo­cused on as­sess­ing some­one’s im­me­di­ate fit for the kind of work we do (via the re­mote work tests), not (other types of) fit with the team and mis­sion.

• lukeprog 28 Feb 2019 17:47 UTC

  8 points

  in reply to: Ben_Kuhn's comment on: After one year of ap­ply­ing for EA jobs: It is re­ally, re­ally hard to get hired by an EA organisation

  Per Buck’s com­ment, I think iden­ti­fy­ing soft­ware en­g­ineer­ing tal­ent is a pretty differ­ent prob­lem than iden­ti­fy­ing e.g. some­one who is already a good fit for Open Phil gen­er­al­ist RA roles.

• lukeprog 28 Feb 2019 17:46 UTC

  26 points

  in reply to: EA applicant's comment on: After one year of ap­ply­ing for EA jobs: It is re­ally, re­ally hard to get hired by an EA organisation

  Thanks for men­tion­ing the thing about the con­ver­sa­tion notes test. It was sim­ply an over­sight to not ex­plic­itly say “Please don’t spend more than X hours on this work test,” and I’ve now added such a sen­tence to our lat­est draft of those work test in­struc­tions. We had ex­plicit time limits for our other two tests.

• lukeprog 28 Feb 2019 17:46 UTC

  47 points

  in reply to: Max_Daniel's comment on: After one year of ap­ply­ing for EA jobs: It is re­ally, re­ally hard to get hired by an EA organisation

  Not sure I fol­low the part about how the kind of thing de­scribed in the origi­nal post makes you “more re­luc­tant to in­tro­duce new peo­ple into the EA com­mu­nity.” There are lots of ex­cit­ing things for EAs to do be­sides “ap­ply to one or more of the 20 most com­pet­i­tive jobs at ex­plic­itly EA-mo­ti­vated em­ploy­ers,” in­clud­ing “keep do­ing what you’re do­ing and en­gage with EA as an ex­cit­ing hobby” and “ap­ply to key po­si­tions in top-pri­or­ity cause ar­eas that are on the 80,000 Hours Job Board but aren’t at one of a hand­ful of ex­plic­itly EA-mo­ti­vated orgs” and “do earn to give for a while while gain­ing skills and then maybe tran­si­tion to more di­rect work later or maybe not,” as well as other paths that are spe­cific to par­tic­u­lar pri­or­ity causes, e.g. for AI strat­egy & policy I’d be ex­cited to see EAs (a) train up in ML, for later work in ei­ther AI safety or AI strat­egy/​policy, (b) fol­low these paths into a US AI policy ca­reer (esp. for US cit­i­zens, and esp. now that CSET ex­ists), and (c) train up as a cy­ber­se­cu­rity ex­pert (I hope to say more later about why this path should be es­pe­cially ex­cit­ing for AI-in­ter­ested EAs; also the worst that hap­pens is that you’ll be in ex­tremely high de­mand and highly paid).

• lukeprog 28 Feb 2019 17:43 UTC

  29 points

  in reply to: Max_Daniel's comment on: After one year of ap­ply­ing for EA jobs: It is re­ally, re­ally hard to get hired by an EA organisation

  Yeah, this is one rea­son Open Phil pays peo­ple for do­ing our re­mote work tests, so that peo­ple who don’t hap­pen to have run­way/​similar can still go through our pro­cess. Pos­si­bly more EA orgs should do this if they aren’t already.

• lukeprog 28 Feb 2019 16:36 UTC

  59 points

  on: After one year of ap­ply­ing for EA jobs: It is re­ally, re­ally hard to get hired by an EA organisation

  Oof, 8 weeks of effort to get ⁰⁄₂₀ po­si­tions is pretty bru­tal. It’s easy to see how that would feel like your “Hey you!…” para­graph. And while I sus­pect you’re a bit of an out­lier in time spent and po­si­tions ap­plied for, I also think you’re point­ing at some­thing true about the cur­rent situ­a­tion re: job open­ings at EA-mo­ti­vated em­ploy­ers, as ev­i­denced by how many up­votes this post has got­ten, some of the com­ments on this page, and the data I’ve got as a re­sult of man­ag­ing Open Phil’s 2018 re­cruit­ment round of Re­search An­a­lysts, dur­ing which we had to say “no” to tons of ap­pli­cants with quite im­pres­sive re­sumes.

  I’ve been writ­ing up some re­flec­tions on that re­cruit­ing round, which I hope to share soon. One of my take­aways is some­thing like “The base of tal­ent out there is strong, and Open Phil’s cur­rent abil­ity to de­ploy it is weak.” In that way we might be an ex­treme op­po­site of Teach for Amer­ica, and I sus­pect many other EA-mo­ti­vated orgs are as well.

  Any­way, I plan to say more on these top­ics when I share my “re­flec­tions” post, but in the mean­time I just want to say I’m sorry that you spent so much time ap­ply­ing to EA orgs and got no offers. Also, set­ting the time in­vest­ment aside, it’s also just emo­tion­ally difficult to get an “Un­for­tu­nately, we’ve de­cided…” email, let alone re­ceive 20 of them in a row.

  A cou­ple other ran­dom notes for now:
  - A col­league of mine has heard some EAs — per­haps mo­ti­vated by con­sid­er­a­tions like those in this post — say­ing stuff like “maybe I shouldn’t even try to ap­ply be­cause I don’t want to waste orgs’ time.” In case fu­ture po­ten­tial Open Phil ap­pli­cants end up read­ing this com­ment, let it be known that we don’t think it’s a waste of our time to pro­cess ap­pli­ca­tions. If we don’t have the staff ca­pac­ity to pro­cess all the ap­pli­ca­tions we re­ceive, we can always just drop a larger frac­tion of ap­pli­cants at each stage. But if some­one never ap­plies, we have no op­por­tu­nity at all to figure out how good a fit they might be. Also, what we’re look­ing for is pretty un­clear (es­pe­cially to po­ten­tial ap­pli­cants), and so e.g. some of our re­cent hires are peo­ple who told us they prob­a­bly wouldn’t have both­ered ap­ply­ing if we hadn’t proac­tively en­couraged them to ap­ply. Of course, an ap­pli­cant could be wor­ried about whether ap­ply­ing is worth their time, and that’s a differ­ent mat­ter.
  - I think it would’ve been good to men­tion that some of these or­ga­ni­za­tions pay ap­pli­cants for some/​all of the time they spend on the ap­pli­ca­tion pro­cess. (Hope­fully Open Phil isn’t the only one?)

• lukeprog 20 Dec 2018 13:59 UTC

  16 points

  on: EA Meta Fund AMA: 20th Dec 2018

  Roughly how much time per month/​year does each of the fund man­agers cur­rently ex­pect to spend on in­ves­ti­gat­ing, dis­cussing, and de­cid­ing on grant op­por­tu­ni­ties?

• lukeprog 20 Dec 2018 13:58 UTC

  20 points

  on: Long-Term Fu­ture Fund AMA

  Roughly how much time per month/​year does each of the fund man­agers cur­rently ex­pect to spend on in­ves­ti­gat­ing, dis­cussing, and de­cid­ing on grant op­por­tu­ni­ties?

• lukeprog 21 Nov 2018 22:33 UTC

  1 point

  on: In­som­nia: a promis­ing cure

  I dis­agree about the strength of ev­i­dence for CBT-I effec­tive­ness.

• lukeprog 16 Oct 2018 4:51 UTC

  11 points

  in reply to: jpaddison's comment on: Reliev­ing ex­treme phys­i­cal pain in hu­mans – an op­por­tu­nity for effec­tive funding

  Open Phil grant on chronic pain.

• lukeprog 23 May 2018 21:01 UTC

  1 point

  in reply to: jpaddison's comment on: Hi, I’m Holden Karnofsky. AMA about jobs at Open Philanthropy

  Some­body asked the fol­low­ing ques­tion:

  I un­der­stand that OP’s prefer­ence is to have promis­ing can­di­dates at­tend an in-per­son work trial rather than do an ad­di­tional re­mote work test; would this prefer­ence still stand if the can­di­date in ques­tion has to ob­tain a US work visa spon­sor­ship in or­der to at­tend the in-per­son trial?

  Our re­ply is: Yes, that prefer­ence stands re­gard­less of cur­rent work au­tho­riza­tion sta­tus, though of course in some cases there won’t be any way for us to help an ap­pli­cant get US work au­tho­riza­tion, de­pend­ing on their situ­a­tion.