Sanjay

• Sanjay 20 May 2020 22:29 UTC

  2 points

  on: Pri­ori­tiz­ing COVID-19 in­ter­ven­tions & in­di­vi­d­ual donations

  Thank you very much for writ­ing this. I was very in­ter­ested in your com­ments on Johns Hop­kins CHS, and found your cri­tiques very in­ter­est­ing. Those who found this in­ter­est­ing may wish to take a look at re­cent call notes from SoGive’s re­cent call with them, which can be found here: https://​​fo­rum.effec­tivealtru­ism.org/​​posts/​​d2LyLQZpoiqbCn­rBo/​​call-notes-with-johns-hop­kins-chs

  I also copy and paste be­low the ex­cerpt which speci­fi­cally tack­les the cri­tiques raised here.

  CHS ex­plic­itly recom­mended against wear­ing DIY masks in early March (a po­si­tion re­versed by the end of the month). When asked about this, CHS ob­served that there’s still not great data and ev­i­dence on masks. And that there’s a risk that us­ing it with­out proper train­ing could lead peo­ple to touch their face more. How­ever on bal­ance CHS has up­dated their opinions on this, and fur­ther ac­knowl­edges that masks are helpful for source con­trol.

  CHS were not dis­cour­ag­ing peo­ple from press­ing ahead with travel plans as late as 6th March. When asked about this, they ob­served that it was a change in em­pha­sis be­tween a com­monly-held ear­lier view, which was that travel bans don’t achieve much be­cause they ul­ti­mately don’t change the num­ber of peo­ple in­fected, they only de­lay the in­evitable. The change in view was be­cause they had pre­vi­ously un­der­es­ti­mated the value of de­lay­ing in­fec­tions.

  When dis­cussing the items referred to in the pre­vi­ous two para­graphs, CHS referred to the already es­tab­lished re­ceived opinion or “dogma”, and how re­cent ex­pe­rience has been up­turn­ing that dogma.

Call notes with Johns Hop­kins CHS

• Sanjay 22 Apr 2020 14:41 UTC

  14 points

  in reply to: Sanjay's comment on: The Case for Im­pact Pur­chase | Part 1

  I per­haps need to clar­ify what I meant by Im­pact Fi­nance. Given that this risks be­ing con­fused with Im­pact In­vest­ment, I’ll re­name it to Con­di­tional Im­pact Fi­nance.

  Con­di­tional Im­pact Fi­nance oc­curs when a Pro­ject re­ceives fund­ing in­volv­ing a Con­di­tional Donor and an Im­pact In­vestor as fol­lows:

  (1) The Con­di­tional Donor (who may be a pool of donors) agrees to fund the Pro­ject *on the con­di­tion* that they achieve a cer­tain outcome

  (2) The funds needed + a bonus go into es­crow. If the out­come is achieved the funds are paid out, if not they are re­turned to the donor.

  (3) In or­der that the Pro­ject re­ceives the funds it needs to pro­ceed with the work, it re­ceives fund­ing from an Im­pact In­vestor (or a pool of in­vestors).

  Fur­ther notes:

  (1) Ex­am­ples of con­di­tions (fic­tional):

  “The ABC in­no­va­tive ed­u­ca­tion Pro­ject will re­ceive fund­ing if a lon­gi­tu­di­nal RCT shows that em­ploy­ment rates among benefi­cia­ries are at least 15% higher than con­trol in 15 years’ time”

  “The DEF malaria Pro­ject will re­ceive fund­ing if the malaria-linked mor­tal­ity rate in Bu­sia, Kenya is bet­ter than a con­trol un­der a RCT con­ducted by LSHTM”

  (2) A bonus is needed pri­mar­ily to pay the in­ter­est on the debt (al­though per­haps a bonus for the staff might be a good idea too)

  (3) If the pro­ject fails, the im­pact in­vestor loses out. This means that all the in­cen­tives that cur­rently ap­ply to fi­nan­cial mar­kets now ap­ply to im­pact.

  I con­sider this bet­ter for any rea­son­ably sized pro­ject. If we are think­ing of just one per­son do­ing some work on their own, I could imag­ine Im­pact Pur­chase maybe hav­ing some value.

  I be­lieve this is what is be­ing done by al­ice.si

• Sanjay 21 Apr 2020 10:09 UTC

  8 points

  on: The Case for Im­pact Pur­chase | Part 1

  I’m con­fused why there are sev­eral peo­ple ap­par­ently ex­cited about this idea. It strikes me that im­pact fi­nance so ob­vi­ously out­performs im­pact pur­chases. So much so that I’m wor­ried that I must be mi­s­un­der­stand­ing im­pact pur­chases.

  Im­pact fi­nance solves both of the stark prob­lems of im­pact pur­chases:

  (a) if the peo­ple do­ing the work only re­ceive money af­ter the event, how do they live?

  (b) what in­cen­tive do pur­chasers of im­pact cer­tifi­cates have to pur­chase?

• Sanjay 27 Mar 2020 23:16 UTC

  2 points

  in reply to: alexrjl's comment on: What’s the best plat­form/​app/​ap­proach for fundrais­ing for things that aren’t reg­istered non­prof­its?

  alexrjl is cor­rect, SoGive can do this.

  SoGive in­cludes:

  • a “progress bar” of how much has been raised and what the tar­get is

  • a sec­tion show­ing other peo­ple who have donated (and op­tions to anonymise name or amount or both)

  • cus­tomis­able sec­tions for the per­son run­ning the fundraiser, in­clud­ing abil­ity to cus­tomise images and text

  Here’s an ex­am­ple: https://​​app.so­give.org/​​#fundraiser/​​san­jay.DRHEz44p.ccdce9

  If you want min­i­mal clicks from see­ing the fundraiser to hav­ing donated, you could sim­ply in­clude one of these links and once you’re on that page you have a donate but­ton right there.

  If this is of in­ter­est to you let me know on san­jay@so­give.org. Also, might be worth check­ing where you are based. This was origi­nally de­signed for the UK, and then adapted for over­seas (e.g. USA)

• Sanjay 24 Mar 2020 9:58 UTC

  6 points

  in reply to: Aaron Gertler's comment on: What posts do you want some­one to write?

  This com­mented was pointed out to me by some­one who thought I may be ex­tremely well qual­ified to an­swer this.

  I have been on the trustee boards of about half a dozen char­i­ties and performed short term con­sult­ing stints (about a month at a time) for an­other 10(ish) char­i­ties globally, and have seen how each of those en­gages with its donors.

  I’m also know peo­ple peo­ple/​or­gani­sa­tions sur­vey­ing char­i­ties about ques­tions like how much more valuable is un­re­stricted than re­stricted fund­ing.

  I would be happy to put some­thing to­gether on this topic, how­ever I’m snowed un­der with other things for the time be­ing, but could add it to the list and tackle it later?

• Sanjay 21 Mar 2020 21:45 UTC

  13 points

  on: syn­thetic in­dices used to rate all char­i­ties: What kind of star rat­ings ex­ist?

  Roughly speak­ing, or­gani­sa­tions that rate/​as­sess char­i­ties fall into two cat­e­gories:

  Those that in­clude thought­ful, im­pact-ori­ented, cost-effec­tive­ness-aware anal­y­sis tend not to use a for­mu­laic or in­dex-based or scal­able ap­proach. Ex­am­ples of this sort of or­gani­sa­tion in­clude GiveWell. I cre­ated this chart some time back, since then Founders Pledge could prob­a­bly added to this cat­e­gory.

  In the op­po­site cor­ner is those or­gani­sa­tions whose method­ol­ogy is more for­mu­laic/​scal­able.

  At the time I cre­ated this chart, there was a gap for anal­y­sis that was scal­able, but also im­pact-ori­ented and cost-effec­tive­ness-aware.

  

  Since then I’ve set up SoGive to re­solve this gap. You can have a look on https://​​app.so­give.org/​​ and search for char­i­ties or cause ar­eas there. Some char­i­ties are as­sessed as Gold, Silver, or Bronze. We are still work­ing through the char­i­ties, so not all of them have a rat­ing. If you click through to a char­ity’s page, you’ll see a tab called anal­y­sis. Lots (but not all) ma­jor UK char­i­ties are cov­ered her. In the US the clos­est or­gani­sa­tion to SoGive (in terms of ap­proach) is Im­pact Mat­ters.

• Sanjay 21 Mar 2020 12:59 UTC

  4 points

  in reply to: Sean_o_h's comment on: The best places to donate for COVID-19

  Thanks Sean_o_h. In case it’s not com­ing across clearly in the write-up, some dona­tion op­por­tu­ni­ties con­sid­ered are di­rectly rele­vant to COVID-19, and some are less di­rectly rele­vant (e.g. they cover pan­demics gen­er­ally, which may be an ap­peal­ing dona­tion idea for those in­spired to donate be­cause of COVID-19).

  An in­di­ca­tion of whether the pro­ject is di­rectly or in­di­rectly rele­vant is given in the table in sec­tion ‘0. Exec Sum­mary’ (see the table with the pur­ple header)

• Sanjay 20 Mar 2020 15:58 UTC

  3 points

  in reply to: Peter_Hurford's comment on: The best places to donate for COVID-19

  “I must ad­mit I am con­fused. I don’t know why you are mak­ing recom­men­da­tions for COVID-19 dona­tions if you’d also pri­ori­tize the ne­glected needs of the de­vel­op­ing world? ”

  This is to an­swer the ques­tion of where to donate if the donors *speci­fi­cally* wants to donate to some­thing COVID-19-rele­vant.

  “I am skep­ti­cal that any­thing in Cat­e­gories 1 and 2 are cost-com­pet­i­tive with ex­ist­ing EA work—and at the min­i­mum this cost-effec­tive­ness is still far from be­ing es­tab­lished. ”

  Lots of the or­gani­sa­tions in cat­e­gories 1 and 2 would be con­sid­ered to ac­tu­ally *be* ex­ist­ing EA work. The first-men­tioned or­gani­sa­tion (Johns Hop­kins CHS) is recom­mended by Founders Pledge.

  ″ What kind of crite­ria were you us­ing to gen­er­ate these recom­men­da­tions? ”

  The crite­ria are: where the or­gani­sa­tion falls into a cat­e­gory which is more pos­i­tively viewed *and* which already has some anal­y­sis to sup­port it.

  “1.) If the re­sponse is likely not enough, wouldn’t it be worth­while to help? ”

  Yes, it would be worth­while to help. The ques­tion isn’t whether it would achieve some­thing and not noth­ing. The con­tention is that the higher rated dona­tion op­por­tu­ni­ties would out­perform.

  “2.) I also very much doubt there won’t be an af­fect on Trump’s re-elec­tion chances, but I don’t think it is rele­vant. ”

  The rea­son why I con­sider it rele­vant is that it gives the Trump ad­minis­tra­tion a stronger in­cen­tive to help those af­fected by the COVID-19 crisis. This means that the cause of helping those in the US is un­likely to be ne­glected, and sub­stan­tially less ne­glected than sup­port­ing those in sub-Sa­haran Africa

The best places to donate for COVID-19

• Sanjay 20 Mar 2020 8:36 UTC

  5 points

  in reply to: Peter_Hurford's comment on: The best places to donate for COVID-19

  This ap­pears to be­long to Cat­e­gory 4: crowd­ing out the work of gov­ern­ment.

  GiveDirectly coun­ters that the gov­ern­ment re­sponse will likely not be enough. That may well be cor­rect, but it will prob­a­bly be enough to en­sure that it doesn’t dam­age Trump’s re-elec­tion chances.

  If ever there’s a time when fo­cus­ing on the ne­glected needs of the de­vel­op­ing world ap­plied, it’s now. I don’t know to what ex­tent Afri­can gov­ern­ments will be ad­vo­cat­ing so­cial dis­tanc­ing, but it will be very hard to im­ple­ment for peo­ple who don’t have sav­ings, and the risk that the cur­rent in­sta­bil­ity could spark in­ter-state con­flict is much higher in the de­vel­op­ing world than in the US.

  Giv­ing cash to peo­ple in the US would not be recom­mended.

• Sanjay 20 Mar 2020 8:27 UTC

  3 points

  in reply to: Peter_Hurford's comment on: The best places to donate for COVID-19

  I ac­tu­ally never looked at them un­til the COVID-19 situ­a­tion, when I started to re­fer to them. So that does seem to sup­port your suggestion

• Sanjay 12 Mar 2020 21:39 UTC

  5 points

  on: Is nan­otech­nol­ogy (such as APM) im­por­tant for EAs’ to work on?

  I asked a re­lated ques­tion on the EA face­book page a few months back. Here’s the ques­tion I asked: https://​​www.face­book.com/​​groups/​​effec­tive.al­tru­ists/​​perma­l­ink/​​2592010687521939/​​

  I copy be­low the ques­tion I asked, and the re­sponse which I thought was most use­ful. (I’ve tried to in­clude the image, but if it didn’t work, click through to the fb page)

  MY QUESTION

  The 80,000 hours ar­ti­cle on ex­tinc­tion risk in­cludes a table from an FHI sur­vey in 2008 (scroll down for source) (ed­ited to clar­ify not nec­es­sar­ily cur­rent views of 80k)
  As molec­u­lar nan­otech weapons top the list, this makes me won­der what is be­ing done about this?
  The only or­gani­sa­tion that I’m aware of in this space is the Fore­sight In­sti­tute (https://​​fore­sight.org/​​) fea­tur­ing Eric Drexler. I also found the Cen­ter for Re­spon­si­ble Nan­otech http://​​www.cr­nano.org, but I couldn’t tell if the web­site was out of date.
  I’m in­ter­ested in dis­cus­sion along the lines of:
  

  • Which orgs are work­ing on this?

  • What are the gaps re­lat­ing to molec­u­lar nan­otech weapons?

  • Is there anal­y­sis on where to donate to tackle this?

  • Would a dona­tion to biorisk orgs already cover this? I’m think­ing of those recom­mended by Founders Pledge (https://​​founder­spledge.com/​​re­search/​​fp-ex­is­ten­tial-risk) or pos­si­bly those men­tioned by Open Phil (https://​​www.open­philan­thropy.org/​​fo­cus/​​global-catas­trophic-risks/​​biose­cu­rity)

  • Does any­one think this table over­states the risk from nan­otech weapons? (e.g. be­cause the nan­otech peo­ple sur­veyed are more bi­ased than other spe­cial­ists? Or re­cent de­vel­op­ments make the info out of date?)

  Notes:Here’s the 80k ar­ti­cle: https://​​80000hours.org/​​ar­ti­cles/​​ex­tinc­tion-risk/​​The ul­ti­mate source of the table is this sur­vey: https://​​www.fhi.ox.ac.uk/​​re­ports/​​2008-1.pdfFur­ther read­ing:https://​​en.wikipe­dia.org/​​wiki/​​Molec­u­lar_nan­otech­nol­ogy#Riskshttp://​​www.cr­nano.org/​​dan­gers.htmhttps://​​www.re­al­clear­science.com/​​lists/​​top_meth­ods_hu­man_ex­tinc­tion/​​molec­u­lar_nan­otech_weapons.htm­lhttps://​​www.cnbc.com/​​2017/​​03/​​17/​​mini-nukes-and-in­spect-bot-weapons-be­ing-primed-for-fu­ture-war­fare.html

  A PARTICULARLY USEFUL ANSWER (from Howie Lem­pel):

  [Dis­claimer: I was work­ing on GCRs at Open Phil when the be­low page was writ­ten but I did not write it and don’t have any in­ter­est­ing in­side in­for­ma­tion on Open Phil’s views on the topic. The views be­low are my own only.]
  

  Open Phil put out a (very) shal­low in­ves­ti­ga­tion of risks from atom­i­cally pre­cise man­u­fac­tur­ing (i.e. nan­otech) back in 2015. One thing they found is:
  

  “Un­less APM is de­vel­oped in a se­cret “Man­hat­tan Pro­ject”—and there is dis­agree­ment about how plau­si­ble that is —the peo­ple we spoke with be­lieve it would be ex­tremely un­likely for an ob­server closely watch­ing the field to be sur­prised by a sud­den in­crease in po­ten­tially dan­ger­ous APM ca­pa­bil­ities.”
  https://​​www.open­philan­thropy.org/​​.../​​atom­i­cally-pre­cise...
  

  This sug­gests we might ex­pect risk re­duc­tion work to nat­u­rally ramp up if/​when the tech starts de­vel­op­ing and that it may be more effi­cient for risk re­duc­tion work to be done once the field makes some progress.
  

  More on this:
  “Peo­ple who are watch­ing the field and know what to look for would be un­likely to be caught off guard even by rapid de­vel­op­ments in atom­i­cally pre­cise man­u­fac­tur­ing. While de­vel­op­ment could be sur­pris­ingly fast, it would be pos­si­ble to ob­serve the sub­stan­tial ad­vances in var­i­ous ca­pa­bil­ities of nanosys­tems (e.g., me­chan­i­cal stiff­ness of var­i­ous types of nanos­truc­tures, num­ber of mov­ing parts of me­chan­i­cal sys­tems, and lat­tice sizes of ma­te­ri­als used to build in­tri­cate sys­tems) that would come be­fore the tech­nol­ogy reaches its ma­ture form. Hy­po­thet­i­cally, sur­prise could come if there were a se­cret pro­ject aimed at de­vel­op­ing the tech­nol­ogy, but that would be im­plau­si­ble in the pre­sent cli­mate.”
  GiveWell’s non-ver­ba­tim sum­mary of a con­ver­sa­tion with Eric Drexler, Oc­to­ber 8, 2014.
  https://​​files.givewell.org/​​.../​​Eric%20Drexler%2010-08-14...
  

  See also:
  https://​​files.givewell.org/​​.../​​Adam%20Mar­ble­stone%208-27...
  

  [Note—I haven’t thought about the is­sue re­cently enough or deeply enough to have my own view on how valuable work on APM risk might be to­day.]

• Sanjay 7 Mar 2020 20:20 UTC

  4 points

  in reply to: Denkenberger's comment on: Quan­tify­ing lives saved by in­di­vi­d­ual ac­tions against COVID-19

  Pre­sum­ably the benefit comes from flat­ten­ing the curve. I.e. if we don’t in­tro­duce con­trol mea­sures, the de­mands on the health­care sys­tem will be un­man­age­able, whereas if we spread them out, then the health­care sys­tem can cope with the de­mand.

  I don’t know how to add images in com­ments, so here’s a link to a rele­vant image: https://​​imgur.com/​​ckt0ujv

  And this is the ar­ti­cle the image came from: https://​​www.vox.com/​​sci­ence-and-health/​​2020/​​3/​​6/​​21161234/​​coro­n­avirus-covid-19-sci­ence-out­break-ends-en­demic-vac­cine

Con­flict and poverty (or should we tackle poverty in nu­clear con­texts more?)

• Sanjay 5 Mar 2020 12:43 UTC

  1 point

  in reply to: eca's comment on: Are there good EA pro­jects for helping with COVID-19?

  I have a maths back­ground, qual­ified as an ac­tu­ary in half the in­dus­try av­er­age time, and am com­fortable with stochas­tic mod­els, in­clud­ing markov chains and monte carlo meth­ods. Are you able to provide more in­for­ma­tion? In case you don’t want to do so in pub­lic, I have sent you a di­rect mes­sage via the fo­rum.

• Sanjay 26 Feb 2020 23:06 UTC

  7 points

  on: Against anti-na­tal­ism; or: why cli­mate change should not be a sig­nifi­cant fac­tor in your de­ci­sion to have children

  I ac­tu­ally think that a true mea­sure of the cli­mate im­pact of hav­ing a child should not just fac­tor in the ex­tra car­bon they will be re­spon­si­ble for over their lives, but also the very small prob­a­bil­ity that they will be re­spon­si­ble for do­ing some­thing awe­some (e.g. “solv­ing cli­mate change”), which may be enough to offset the ex­pected car­bon foot­print.

  (Of course, this only makes sense if you think that the tail pos­i­tive risk of them do­ing some­thing amaz­ingly pos­i­tive out­weighs the tiny prob­a­bil­ity that they will do some­thing stun­ningly nega­tive!)

• Sanjay 26 Feb 2020 22:53 UTC

  3 points

  on: Bi­ases in our es­ti­mates of Scale, Ne­glect­ed­ness and Solv­abil­ity?

  Thank you for giv­ing this some thought. The im­pres­sion I have from this is that the con­cerns you raise re­late to ei­ther spu­ri­ous lev­els of ac­cu­racy or edge cases.

  But maybe my im­pres­sion is wrong. If you could provide some spe­cific ex­am­ples where ei­ther 80k or an­other org has made a wrong de­ci­sion be­cause of these con­sid­er­a­tions, I would find this piece much more in­ter­est­ing.

Micro­cre­dit may some­times be effec­tive, but per­haps shouldn’t be funded by donations

• Sanjay 16 Feb 2020 18:32 UTC

  5 points

  in reply to: Aaron Gertler's comment on: Clean cook­stoves may be com­pet­i­tive with GiveWell-recom­mended charities

  Thanks for the en­courage­ment. I think that aiming for a “perfect” write-up has been a bar­rier to pub­lish­ing con­tent, so I in­tend for us to pub­lish more shal­low re­views to ad­dress this.

  To an­swer your ques­tion, I think the best fo­cus ar­eas would be the six bul­let points high­lighted near the start of the ar­ti­cle, with a par­tic­u­lar fo­cus on the first two (are the stoves ac­tu­ally used, and are they ac­tu­ally clean?) and the last (what is the best way to fund this work?).

  Also, we would fur­ther in­ves­ti­gate the very use­ful com­ments made by MatthewDahlhausen (which seemed very use­ful and was up­voted by me) and look fur­ther at the GiveWell anal­y­sis as men­tioned by cole_haus (I was aware of this, but had not had the ca­pac­ity to re­view it prop­erly)