EA Organization Updates: December 2019

Th­ese monthly posts origi­nated as the “Up­dates” sec­tion of the EA Newslet­ter.

You can also see last month’s up­dates, or a repos­i­tory of past newslet­ters (in­clud­ing past or­ga­ni­za­tion up­dates).

Or­ga­ni­za­tion Updates

80,000 Hours

In De­cem­ber, 80,000 Hours added over 170 roles to their job board and re­leased five pieces of con­tent, in­clud­ing their most pop­u­lar pod­cast epi­sode to date, which fea­tures David Chalmers:

Fi­nally, they up­dated their guide to choos­ing a char­ity for 2020.

An­i­mal Char­ity Evaluators

An­i­mal Char­ity Eval­u­a­tors re­cently pub­lished up­dates from or­ga­ni­za­tions that re­ceived grants from the Effec­tive An­i­mal Ad­vo­cacy Fund. They also shared a blog post in which some ACE staff mem­bers dis­cuss the rea­son­ing be­hind their 2019 dona­tions.

Cen­ter for Hu­man-Com­pat­i­ble AI

Adam Gleave, Michael Den­nis, Neel Kant, Cody Wild, Sergey Lev­ine, and Stu­art Rus­sell had a new pa­per (“Ad­ver­sar­ial Poli­cies: At­tack­ing Deep Re­in­force­ment Learn­ing”) ac­cepted by ICLR.

Cen­tre for Effec­tive Altruism

Char­ity Entrepreneurship

CE pub­lished the Poli­ti­cal Change — Ap­proach Re­port. The re­port ex­am­ines two ideas for reach­ing out to gov­ern­ments to change laws that af­fect an­i­mal welfare (dis­solved oxy­gen for fish in Taiwan and feed for­tifi­ca­tion for egg-lay­ing hens in In­dia). Pa­trick Stadler, CE’s Char­ity Men­tor and Cur­ricu­lum Devel­oper, pub­lished a new ar­ti­cle: “Six Ways Your Char­ity Startup Might Fail — and How to Prevent That.”

Effec­tive Altru­ism Foun­da­tion and Foun­da­tional Re­search Institute

In 2016, EAF launched a bal­lot ini­ti­a­tive to dras­ti­cally in­crease Zurich’s de­vel­op­ment co­op­er­a­tion bud­get. They later co­or­di­nated a coun­ter­pro­posal with the city coun­cil that pre­served es­sen­tial fea­tures of their ini­ti­a­tive. This coun­ter­pro­posal passed at the bal­lot with a 70% ma­jor­ity. The de­vel­op­ment co­op­er­a­tion bud­get will thus in­crease from $3 mil­lion to $8 mil­lion per year and will have to be al­lo­cated “based on the available sci­en­tific re­search on effec­tive­ness and cost-effec­tive­ness.”

Other news:

Faunalytics

Fau­n­a­lyt­ics has mul­ti­ple stud­ies in progress and has just be­gun a new study on re­duc­ing the suffer­ing of farmed chicken and fish. They will in­ves­ti­gate the fre­quency and im­pli­ca­tions of nega­tive be­liefs held by the gen­eral pop­u­la­tion with re­spect to small-bod­ied an­i­mals, iden­ti­fy­ing the biggest bar­ri­ers for in­sti­tu­tional and in­di­vi­d­ual cam­paigns. They are ac­cept­ing dona­tions for all of their cur­rent re­search in progress.

They also pub­lished a new blog high­light­ing their most pop­u­lar re­search library en­tries of 2019. High-traf­fic en­tries fo­cused on so­cial is­sues re­lated to ve­g­anism and veg ad­vo­cacy, TNVR (Trap, Neuter, Vac­ci­nate, Re­turn) pro­grams, and wildlife trauma.

Fu­ture of Hu­man­ity Institute

Two pa­pers by FHI re­searchers have been ac­cepted to the As­so­ci­a­tion for the Ad­vance­ment of Ar­tifi­cial In­tel­li­gence (AAAI) work­shop on AI safety in Fe­bru­ary 2020:

A pa­per by Michael Co­hen, Badri Vel­lambi and Mar­cus Hut­ter, “Asymp­tot­i­cally Unam­bi­tious Ar­tifi­cial Gen­eral In­tel­li­gence,” has also been ac­cepted to the main AAAI con­fer­ence.

Ben Garfinkel and Allan Dafoe pub­lished an ar­ti­cle in War on the Rocks based on their pa­per “How Does the Offense-Defense Balance Scale?

GiveWell

GiveWell de­cided to al­lo­cate $2.6 mil­lion re­ceived in dis­cre­tionary fund­ing from the third quar­ter of 2019 to He­len Kel­ler In­ter­na­tional’s vi­tamin A sup­ple­men­ta­tion pro­gram. They also recom­mended a $355,000 grant from the Effec­tive Altru­ism Global Health and Devel­op­ment Fund to One for the World, and recom­mended a $214,425 exit grant from Open Philan­thropy to Char­ity Science Health.

Global Catas­trophic Risk Institute

GCRI has an­nounced a new call for ad­visees and col­lab­o­ra­tors for se­lect AI pro­jects. They wel­come in­quiries from stu­dents and pro­fes­sion­als at all lev­els who are in­ter­ested in seek­ing their ad­vice and/​or col­lab­o­rat­ing with them on AI pro­jects funded by a re­cent dona­tion for AI re­search and out­reach.

GCRI also re­cently hosted a ses­sion on global catas­trophic risk at the 2019 An­nual Meet­ing of the So­ciety for Risk Anal­y­sis.

Open Philan­thropy Project

Open Philan­thropy an­nounced grants in­clud­ing $9.1M to Cinc­in­nati Chil­dren’s Hospi­tal Med­i­cal Cen­ter, $4.2M to the Texas Or­ga­niz­ing Pro­ject, $1.7M to An­ima In­ter­na­tional, and $705K to the Berkeley Ex­is­ten­tial Risk Ini­ti­a­tive. They also an­nounced their al­lo­ca­tion of $54.6M to GiveWell top char­i­ties and pub­lished a set of sug­ges­tions from Open Philan­thropy staff for in­di­vi­d­ual donors.

Ought

Ought pub­lished a tech­ni­cal progress re­port (Eval­u­at­ing Ar­gu­ments One Step at a Time) de­scribing its re­cent fac­tored eval­u­a­tion ex­per­i­ments. Ought found that fac­tored eval­u­a­tion can in­deed dis­t­in­guish some valid from in­valid ar­gu­ments for claims about movie re­views. But a lot of var­i­ance ex­ists in the data, which makes the re­sults sen­si­tive to en­sem­bling meth­ods.

Ought is ex­cited about the foun­da­tion this work es­tab­lishes: the method­ol­ogy and suc­cess crite­ria defined over the course of this ex­per­i­ment should set clear frame­works and baselines for fu­ture work.

Rais­ing for Effec­tive Giving

REG’s Dou­bleUpDrive raised a to­tal of $3,421,524 from pro­fes­sional poker play­ers and other donors, in­clud­ing $640,000 across 80,000 Hours, the Cen­tre for Effec­tive Altru­ism, the Forethought Foun­da­tion, and the Long-Term Fu­ture Fund. They will pub­lish an anal­y­sis of the Drive’s coun­ter­fac­tual im­pact later this year.

Re­think Charity

Re­think Pri­ori­ties, a pro­ject of Re­think Char­ity, pub­lished an in­ter­ven­tion pro­file on bal­lot ini­ti­a­tives, and con­tinued to pub­lish their anal­y­sis of the 2019 EA Sur­vey with posts on ca­reers and skills and cause pri­ori­ti­za­tion among EAs.

The Good Food Institute

  • The dis­trict judge in GFI’s Arkansas la­bel cen­sor­ship law­suit granted the or­ga­ni­za­tion’s pre­limi­nary in­junc­tion, mean­ing that Arkansas can­not en­force its la­bel cen­sor­ship law against Tofurky while the pri­mary law­suit is pend­ing. The judge held that GFI, the Amer­i­can Civil Liber­ties Union (ACLU), the An­i­mal Le­gal Defense Fund (ALDF), and Tofurky are “likely to pre­vail on the mer­its of [our] First Amend­ment claim.”

  • GFI sub­mit­ted a pub­lic com­ment to the Na­tional Oceanic and At­mo­spheric Ad­minis­tra­tion (NOAA) on one of the agency’s four new blue tech­nol­ogy strate­gies.

  • GFI sup­ported the de­vel­op­ment of a new XPRIZE com­pe­ti­tion aimed at pro­duc­ing full-sized, whole-mus­cle chicken and seafood al­ter­na­tives. The prize com­pe­ti­tion will be sup­ported by $13M of fund­ing from the XPRIZE Foun­da­tion and is ex­pected to in­cen­tivize the found­ing of more than 10 new com­pa­nies in the al­ter­na­tive pro­teins sec­tor by 2022.

  • GFI Brazil or­ga­nized a se­ries of briefings with sci­en­tists from the Brazilian agri­cul­tural re­search agency Em­presa Brasileira de Pesquisa Agropecuária (Em­brapa). This re­la­tion­ship is already pay­ing div­i­dends in sci­en­tific in­ter­est in Brazil, as GFI has re­ceived more Com­pet­i­tive Re­search Grant Pro­gram pro­pos­als from Brazil than from any other coun­try be­sides the U.S.

The Life You Can Save

Since the re­launch of The Life You Can Save, The Life You Can Save has seen 10,000 new email sub­scrip­tions and over 15,000 book down­loads. They are de­vel­op­ing part­ner­ships with a va­ri­ety of or­ga­ni­za­tions, and en­courage read­ers to get in touch if they would like to help share the book.

The Giv­ing Games Pro­ject is cur­rently un­der­tak­ing an im­pact anal­y­sis based on 2019-2020 par­ti­ci­pant re­sults, and will make those re­sults available shortly.

Wild An­i­mal Initiative

Wild An­i­mal Ini­ti­a­tive had a pa­per ac­cepted for pub­li­ca­tion by the flag­ship jour­nal in the field of restora­tion ecol­ogy; wrapped up Sea­son 1 of the Wild­ness pod­cast; and re­leased new re­search on in­sec­ti­cides, long-term con­sid­er­a­tions for in­ter­ven­tion de­sign, and op­ti­mal pop­u­la­tion den­sity.

At EA Global: Lon­don 2019, WAI Re­search Fel­low Dr. Will Brad­shaw de­scribed how bio­mark­ers of ag­ing could serve as an ob­jec­tive and gen­er­al­iz­able met­ric of cu­mu­la­tive welfare. The record­ing is now available for view­ing.

Add your own update

If your or­ga­ni­za­tion isn’t rep­re­sented on this list, you’re wel­come to provide an up­date in a com­ment.

You can also email me if you’d like to be in­cluded on the list of or­ga­ni­za­tions I ask for up­dates each month; I can then add any up­dates you sub­mit to fu­ture posts. (I may not ac­cept all such re­quests; whether I in­clude an org de­pends on its size, age, fo­cus, track record, etc.)