Milan_Griffes

• Milan_Griffes 17 Nov 2019 14:52 UTC

  3 points

  on: [Link] Against “Why We Sleep” (Guzey)

  Some good com­ments on the LessWrong cross-post.

• Milan_Griffes 17 Nov 2019 14:23 UTC

  2 points

  on: UK Gen­eral Elec­tion: Where would you look for im­pact anal­y­sis of policy plat­forms?

  You could adapt kbog’s scor­ing sys­tem to the UK con­text.

• Milan_Griffes 17 Nov 2019 14:01 UTC

  5 points

  in reply to: riceissa's comment on: I’m Buck Sh­legeris, I do re­search and out­reach at MIRI, AMA

  Pub­lished to­day: “EA res­i­den­cies” as an out­reach activity

• Milan_Griffes 17 Nov 2019 13:56 UTC

  −1 points

  in reply to: Davidmanheim's comment on: Steel­man­ning the Case Against Un­quan­tifi­able Interventions

  I think we can drop the Bletch­ley park dis­cus­sion.

  
  

  Okay, I take it that you agree with my view.

  
  

  … fu­ture-fo­cused in­ter­ven­tions have a very differ­ent set of ques­tions than pre­sent-day non-quan­tifi­able interventions

  
  

  How are you sep­a­rat­ing out “fu­ture-fo­cused in­ter­ven­tions” from “pre­sent-day non-quan­tifi­able in­ter­ven­tions”?

  Plau­si­bly geo­eng­ineer­ing safety will be very rele­vant in 15-30 years. As­sum­ing that’s true, would you cat­e­go­rize geo­eng­ineer­ing safety re­search as fu­ture-fo­cused or pre­sent-day non-quan­tifi­able?

• Milan_Griffes 16 Nov 2019 17:08 UTC

  28 points

  on: EA Lead­ers Fo­rum: Sur­vey on EA pri­ori­ties (data and anal­y­sis)

  Were rep­re­sen­ta­tives from these groups in­vited to the Lead­ers Fo­rum?

  • GiveWell

  • BERI

  • Sur­vival & Flour­ish­ing Fund

  • LessWrong

  • CSET

  • Re­think Priorities

  • Char­ity Entrepreneurship

  • Founders Pledge

  • Hap­pier Lives Institute

  
  

  If not, why not?

• Milan_Griffes 16 Nov 2019 17:02 UTC

  2 points

  in reply to: JoanGass's comment on: Which Com­mu­nity Build­ing Pro­jects Get Funded?

  Oh right, thanks!

  Also just saw this good com­ment on the same topic.

• Milan_Griffes 16 Nov 2019 12:21 UTC

  2 points

  in reply to: riceissa's comment on: I’m Buck Sh­legeris, I do re­search and out­reach at MIRI, AMA

  See also this re­cent Qualia Com­put­ing post about the or­thog­o­nal­ity the­sis. (Qualia Com­put­ing is the blog of QRI’s re­search di­rec­tor.)

[Link] Against “Why We Sleep” (Guzey)

• Milan_Griffes 15 Nov 2019 19:54 UTC

  4 points

  in reply to: agdfoster's comment on: Which Com­mu­nity Build­ing Pro­jects Get Funded?

  First I want to quickly flag that we no longer do com­mu­nity build­ing grants due to their com­plex­ity and in­stead in­tend to fund CEA CBG.

  
  

  Wait, given Ni­cole’s re­cent post, does this mean that both the Meta Fund & CEA are mov­ing away from mak­ing com­mu­nity grants?

  (From Ni­cole’s post: “At this stage, I think it is fairly likely that EA Grants won’t con­tinue in its cur­rent form, and that we will in­stead en­courage in­di­vi­d­u­als to ap­ply to EA Funds.”)

• Milan_Griffes 15 Nov 2019 19:36 UTC

  8 points

  in reply to: Davidmanheim's comment on: Steel­man­ning the Case Against Un­quan­tifi­able Interventions

  I feel con­fused about whether there’s ac­tu­ally a dis­agree­ment here. Seems pos­si­ble that we’re just talk­ing past each other.

  • I agree that Bletch­ley Park wasn’t mostly fo­cused on crack­ing Enigma.

  • I don’t know enough about Bletch­ley’s his­tory to have an in­de­pen­dent view about whether it was un­der­funded or not. I’ll fol­low your view that it was well sup­ported.

  • It does seem like Tur­ing’s work on Enigma wasn’t highly pri­ori­tized when he started work­ing on it (”...be­cause no one else was do­ing any­thing about it and I could have it to my­self”), and this work turned out to be very im­pact­ful. I feel con­fi­dent claiming that Bletch­ley wasn’t pri­ori­tiz­ing Enigma highly enough be­fore Tur­ing de­cided to work on it. (Cu­ri­ous whether you dis­agree about this.)

  
  

  On the pre­sent-day stuff:

  • My claim is that circa 2010 AI al­ign­ment work was be­ing (dra­mat­i­cally) un­der­funded by in­sti­tu­tions, not that it wasn’t be­ing funded at all.

  • It wouldn’t sur­prise me if 20 years from now the con­sen­sus view was “Oh man, we to­tally should have been putting more effort to­wards figur­ing out what safe geo­eng­ineer­ing looks like back in 2019.”

  • I be­lieve Drexler had a hard time get­ting sup­port to work on nan­otech stuff (be­lieve he’s cur­rently work­ing mostly on AI al­ign­ment), but I don’t know the full story there. (I’m hold­ing Drexler as some­one who is qual­ified and al­igned with EA goals.)

• Milan_Griffes 14 Nov 2019 12:52 UTC

  3 points

  in reply to: Davidmanheim's comment on: Steel­man­ning the Case Against Un­quan­tifi­able Interventions

  Bletch­ley park was ex­actly the sort of in­ter­ven­tion that doesn’t need any push­ing. It was funded im­me­di­ately be­cause of how ob­vi­ous the benefit was.

  
  

  Pretty sure that’s not right, at least for Tur­ing’s work on Enigma:

  “Tur­ing de­cided to tackle the par­tic­u­larly difficult prob­lem of Ger­man naval Enigma ‘be­cause no one else was do­ing any­thing about it and I could have it to my­self’.”

  
  

  If you were to sug­gest some­thing similar now that were poli­ti­cally fea­si­ble and similarly im­por­tant to a coun­try, I’d be shocked if it wasn’t already hap­pen­ing. In­vest in AI and ad­vanced tech­nolo­gies?...

  
  

  What about AI al­ign­ment work circa 2010?

  Quick ex­am­ples from the pre­sent day: prepar­ing for risks from nan­otech­nol­ogy; work­ing on geo­eng­ineer­ing safety

• Milan_Griffes 14 Nov 2019 12:47 UTC

  2 points

  in reply to: Nicole_Ross's comment on: Up­date on CEA’s EA Grants Program

  Got it. What could staff ca­pac­ity trade off for here that feels higher pri­or­ity?

• Milan_Griffes 13 Nov 2019 13:13 UTC

  2 points

  in reply to: MichaelPlant's comment on: Steel­man­ning the Case Against Un­quan­tifi­able Interventions

  About pre­dict­ing effec­tive­ness, it seems your con­clu­sion should be one of epistemic mod­esty re­lat­ing to hard-to-quan­tify in­ter­ven­tions, not that we should never think they are bet­ter.

  
  

  This is where I’m at too – e.g. the im­pact of Bletch­ley Park would have been hard to quan­tify prospec­tively, and in ret­ro­spect was mas­sively pos­i­tive.

  Cu­ri­ous if OP is ac­tu­ally say­ing the other thing (that hard-to-quan­tify im­plies lower cost-effec­tive­ness).

• Milan_Griffes 13 Nov 2019 3:05 UTC

  3 points

  in reply to: Milan_Griffes's comment on: In­sti­tu­tions for Fu­ture Generations

  Also this Quora thread: https://​​www.quora.com/​​What-is-the-old­est-in­sti­tu­tion-or­ga­ni­za­tion-that-ex­ists-today

• Milan_Griffes 13 Nov 2019 3:04 UTC

  3 points

  in reply to: tylermjohn's comment on: In­sti­tu­tions for Fu­ture Generations

  Yeah, it’s a great ques­tion.

  For Catholic stuff, The Great Here­sies looks in­ter­est­ing, though old. (I haven’t read it.)

  I have thoughts about Ma­hayana Bud­dhist value trans­mis­sion. Prob­a­bly best to DM about that.

  I bet Leah Libresco would have good thoughts on Catholic value trans­mis­sion. Mes­sage me if an in­tro would be helpful.

• Milan_Griffes 13 Nov 2019 2:10 UTC

  6 points

  on: EA Lead­ers Fo­rum: Sur­vey on EA pri­ori­ties (data and anal­y­sis)

  In­ter­est­ing, thanks for post­ing.

  Can you share any­thing from the dis­cus­sion about what to do about the prob­lems?

  (Most of this post is “here’s what we think the prob­lems are” – would be su­per in­ter­est­ing to see ”...and here’s what we think we should do about it.”)

• Milan_Griffes 13 Nov 2019 1:12 UTC

  2 points

  on: In­sti­tu­tions for Fu­ture Generations

  In­sti­tu­tions for im­prov­ing sci­en­tific knowl­edge pro­duc­tion & syn­the­sis.

  Some leads here, un­der the “meta-sci­ence” header: https://​​jas­minew.me/​​post/​​progress/​​

• Milan_Griffes 13 Nov 2019 1:08 UTC

  4 points

  on: In­sti­tu­tions for Fu­ture Generations

  So fun. Would be in­ter­est­ing to see tractabil­ity as­sess­ments for all the ideas.

  Have you looked at Catholic in­sti­tu­tions at all? They’ve done a good job of main­tain­ing a co­hered wor­ld­view over a long times­pan (and also of grow­ing very rich).

• Milan_Griffes 9 Nov 2019 16:03 UTC

  31 points

  on: Up­date on CEA’s EA Grants Program

  No time to track down the cite now, but I re­call Oliver men­tion­ing some­where on the Fo­rum that he felt there was a lot of value in there be­ing mul­ti­ple in­de­pen­dent grant-mak­ers with over­lap­ping fo­cus ar­eas, to miti­gate the bi­ases of any one de­ci­sion-maker. (I agree with this.)

• Milan_Griffes 9 Nov 2019 16:01 UTC

  2 points

  in reply to: RyanCarey's comment on: What book(s) would you want a gifted teenager to come across?

  It’s good.