Donor Lottery Debrief

Good news, I’ve fi­nally al­lo­cated the rest of the donor lot­tery funds from the 2016-2017 Donor Lot­tery (the first one in our com­mu­nity)! It took over 3 years but I’m ex­cited about the two pro­jects I funded. It prob­a­bly goes with­out say­ing, but this post is about an in­de­pen­dent pro­ject and does not rep­re­sent CFAR (where I work).

This post con­tains sev­eral up­dates re­lated to the donor lot­tery:


My pre­vi­ous com­ments on the origi­nal donor lot­tery post share the ba­sics of how the first $25k was used for CZEA (this was $5k more than I was origi­nally plan­ning to donate due to trans­fer effi­ciency con­sid­er­a­tions). Look­ing back now, I be­lieve that dona­tion likely had a strong im­pact on EA com­mu­nity build­ing.

My dona­tion was the largest that CZEA had re­ceived (I think they pre­vi­ously had re­ceived one other large dona­tion—about half the size) and it was enough for CZEA to tran­si­tion from a purely vol­un­teer or­ga­ni­za­tion into a par­tially-pro­fes­sional or­ga­ni­za­tion (1 FTE, plus vol­un­teers). Based on con­ver­sa­tions with Jan Kul­veit, I be­lieve it would have taken at least 8 more months for CZEA to pro­fes­sion­al­ize oth­er­wise. I be­lieve that in the time they bought with the dona­tion, they were able to more eas­ily se­cure sub­stan­tial fund­ing from CEA and other fun­ders, as well as scale up sev­eral com­pel­ling ini­ti­a­tives: co-or­ga­niz­ing Hu­man-al­igned AI Sum­mer School, AI Safety Re­search Pro­gram, and a Com­mu­nity Build­ing Re­treat (with CEA).

I also have been glad to see a hand­ful of peo­ple get in­volved with EA and Ra­tion­al­ity through CZEA, and I think the move­ment is stronger with them. To pick an ex­am­ple fa­mil­iar to me, sev­eral CZEA lead­ers were re­cently part of CFAR’s In­struc­tor Train­ing Pro­gram: Daniel Hynk (Co-founder of CZEA), Jan Kul­veit (Se­nior Re­search Scholar at FHI), Tomáš Gavenčiak (In­de­pen­dent Re­searcher who has been funded by EA Grants), and Irena Kotíková (Pres­i­dent of CZEA).

For more de­tail on CZEA’s early his­tory and the im­pact of the donor lot­tery funds (and other in­fluences), see this de­tailed ac­count.


In late April 2020, I heard about Epi­demic Fore­cast­ing—a pro­ject launched by peo­ple in the EA/​Ra­tion­al­ity com­mu­nity to in­form de­ci­sion mak­ers by com­bin­ing epi­demic mod­el­ing with fore­cast­ing. I learned of the fund­ing op­por­tu­nity through my col­league and friend, Eliz­a­beth Gar­rett.

The pitch was im­me­di­ately com­pel­ling to me as a 5-figure donor: A group of peo­ple I already be­lieved to be im­pres­sive and trust­wor­thy were launch­ing a pro­ject to use fore­cast­ing to help pow­er­ful peo­ple make bet­ter de­ci­sions about the pan­demic. Even though it seemed likely that noth­ing would come of it, it seemed like an ex­cel­lent gam­ble to make, based on the fol­low­ing pos­si­ble out­comes:

  • Prevent ill­ness, death, and eco­nomic dam­age by helping gov­ern­ments and other de­ci­sion mak­ers han­dle the pan­demic bet­ter, es­pe­cially gov­ern­ments that couldn’t oth­er­wise af­ford high-qual­ity fore­cast­ing services

  • High­light the power of—and test novel ap­pli­ca­tions of—an un­der­uti­lized tool: fore­cast­ing (see the book Su­perfore­cast­ing for back­ground on this)

  • Test and demon­strate the op­por­tu­nity for in­di­vi­d­u­als and in­sti­tu­tions to do more good for im­por­tant causes by think­ing care­fully (Ra­tion­al­ity/​EA) rather than rely­ing on stan­dard ex­perts and au­thor­i­ties alone

  • En­gage mem­bers of our com­mu­nity in an effort to change the world for the bet­ter, in a way that will give them some quick feed­back—thus lead­ing to deeper/​faster learning

  • Cross-pol­li­nate our com­mu­nity with pro­fes­sional fields en­gaged by EpiFor—pos­si­bly im­prov­ing both those ex­ter­nal fields and the EA/​Ra­tion­al­ist community

I de­cided to move for­ward as quickly as pos­si­ble; EpiFor was already mak­ing de­ci­sions that would go differ­ently based on whether they had se­cured fund­ing or not. In par­tic­u­lar, the timing of the fund­ing com­mit­ment af­fected how many su­perfore­cast­ers and soft­ware en­g­ineers they could af­ford to hire and on­board, and it en­abled them to make the tran­si­tion from a part-time/​vol­un­teer or­ga­ni­za­tion to a full-time salaried staff. It also seemed like some of the stan­dard in­sti­tu­tional donors in the com­mu­nity had pre-de­ter­mined fund­ing cy­cles that might take much longer to com­mit fund­ing—and that even ac­cel­er­at­ing the pro­ject by a week might be quite valuable, es­pe­cially in early days of the pan­demic.

With that in mind it was an easy call for me to make, and I com­mit­ted the re­main­ing $23,500 from the dona­tion lot­tery, as well as some per­sonal funds on top of that. Notably, EpiFor is now con­duct­ing its next fund­ing round, and I con­tinue to sus­pect that more dona­tions may have a sub­stan­tial (though high var­i­ance) im­pact—par­tic­u­larly since fund­ing is cur­rently af­fect­ing which op­por­tu­ni­ties they pur­sue.

The pro­ject’s con­crete out­puts so far in­clude some re­search (cited in a Vox ar­ti­cle yes­ter­day) and be­ing short-listed by phar­ma­ceu­ti­cal com­pa­nies look­ing for help de­sign­ing vac­cine tri­als.

Look­ing Back on the Donor Lottery

Some ob­ser­va­tions from my ex­pe­rience:

  • Us­ing the donor lot­tery to turn a ~$5k dona­tion into a ~$46k dona­tion meant the differ­ence be­tween spend­ing a cou­ple hours de­cid­ing among or­ga­ni­za­tions I already knew, and ac­tu­ally look­ing for pro­jects that would be­have no­tice­ably differ­ently be­cause of my money.

  • The donor lot­tery ended up slow­ing down my dona­tion, be­cause at the larger level I was no longer satis­fied just giv­ing to well-known op­por­tu­ni­ties that seemed to be get­ting the money they need from Open Phil or other in­sti­tu­tional grants. It ul­ti­mately took a lit­tle over 3 years for me to dis­tribute the dona­tions. Depend­ing on dis­count rates, that de­lay might have been a sig­nifi­cant cost.

  • In the three years I con­tinued to hold (at least some of) the money the nom­i­nal value in­creased by al­most $3k, but I think it would have in­creased at least $5k more if I had been pay­ing more at­ten­tion (sim­ply putting it all in S&P 500 in­dex un­til Feb 2020, as I did with most of my ‘own’ sav­ings). This is the clear­est/​sim­plest mis­take I made.

  • Since I had only used about half of the win­nings af­ter two years I con­sid­ered plow­ing the rest of it back into the 2019 donor lot­tery, but ul­ti­mately I did not. In hind­sight I be­lieve it was good that I didn’t—since on ex­pec­ta­tion the money would have been less im­pact­ful than fund­ing EpiFor—but I don’t think the choice was ob­vi­ous at the time.

  • Even af­ter I found the CZEA op­por­tu­nity, I wasn’t sure how easy it would be for me to find at­trac­tive fund­ing op­por­tu­ni­ties that weren’t already suffi­ciently funded by in­sti­tu­tional donors with heav­ily over­lap­ping val­ues (and more time to re­search and re­cruit ap­pli­ca­tions) such as Open Phil, LTFF, andEA Grants.

  • I have come to be­lieve that liv­ing and work­ing in the EA/​Ra­tion­al­ity com­mu­nity in the Bay Area made it much more likely I would hear about at­trac­tive op­por­tu­ni­ties that weren’t yet funded by larger donors. I have also up­dated that there does seem to be more of a niche for 5-to-6-figure long-ter­mist/​EA donors in the com­mu­nity than I had origi­nally thought (per­haps only for peo­ple that are well-po­si­tioned to hear about the op­por­tu­ni­ties).

  • Now hav­ing found such op­por­tu­ni­ties twice in three years, I’m guess­ing I could find more of them at a rate of one per year or more, es­pe­cially if I did more to ad­ver­tise my de­sire to find 5-figure dona­tion op­por­tu­ni­ties. Speak­ing of which, while I’ve finished dis­tribut­ing the donor lot­tery funds, I’m con­tin­u­ing to seek 5-figure dona­tion op­por­tu­ni­ties for (at least) the rest of 2020. See de­tails be­low.

  • Both of the or­ga­ni­za­tions that re­ceived dona­tions did not know I was look­ing for dona­tion op­por­tu­ni­ties un­til a friend or I found them—I be­lieve that not do­ing more to ad­ver­tise my in­ter­est in mak­ing dona­tions was likely the biggest mis­take I made. Hence my fi­nal sec­tion:

Look­ing for more pro­jects like these

If you are launch­ing (or know about) a pro­ject that you be­lieve may have a strong EA im­pact and has room for more fund­ing, I’d be happy to hear about it. I’m most in­ter­ested in pro­jects that have:

  • One or more lead­ers with a track record of mak­ing good stuff happen

  • Ob­vi­ously low down­side risk (I think pro­jects that risk do­ing real harm are bet­ter suited for giv­ing mechanisms that in­clude a solid due dili­gence pro­cess and have mul­ti­ple re­view­ers, such as in­sti­tu­tional grant­mak­ers)

  • 501(c)3 sta­tus or spon­sor­ship (a way for me to get a U.S. tax de­duc­tion). I’m also open to non-tax-de­ductible op­por­tu­ni­ties that are suffi­ciently at­trac­tive.

  • Some rea­son to pre­fer an in­di­vi­d­ual donor like me rather than in­sti­tu­tional donors like Open Phil, LTFF, EA Grants, and GiveWell, such as high time-sen­si­tivity.