An overview of Y Combinator’s non-profit program

Contents

Summary

I provide a brief overview of Y Com­bi­na­tor’s non-profit pro­gram, com­par­ing its work to that of GiveWell and the Open Philan­thropy Pro­ject in the con­text of cause se­lec­tion and pri­ori­ti­za­tion. GiveWell and Y Com­bi­na­tor have sig­nifi­cantly differ­ent ap­proaches to se­lect­ing non-prof­its to fund. Whereas GiveWell seeks the most effec­tive char­i­ties and causes, Y Com­bi­na­tor uses a port­fo­lio ap­proach of fund­ing sev­eral pro­jects that satisfy cer­tain crite­ria (e.g. solv­ing real prob­lems and com­mit­ment to growth). In gen­eral, Y Com­bi­na­tor seems to take the ap­proach they use for for-prof­its and ap­ply it to non-prof­its. In par­tic­u­lar this means they fo­cus on non-prof­its that cen­ter around in­for­ma­tion tech­nol­ogy, where they have the most ex­pe­rience.

Why look at Y Com­bi­na­tor?

My mo­ti­va­tion for look­ing into Y Com­bi­na­tor’s non-profit pro­gram was in the con­text of my work on cause pri­ori­ti­za­tion. I’ve been in­ter­ested in differ­ent ap­proaches to cause se­lec­tion and pri­ori­ti­za­tion, so look­ing at how a suc­cess­ful for-profit or­ga­ni­za­tion like Y Com­bi­na­tor ap­proaches the prob­lem is prima fa­cie likely to provide in­sight that is dis­tinct from that pro­vided by look­ing at non-profit or­ga­ni­za­tions like GiveWell. I also treated this in­ves­ti­ga­tion as a “test case” to see if this sort of overview of differ­ent ap­proaches and com­par­i­son is use­ful.

To elab­o­rate, there are sev­eral rea­sons look­ing into some­thing like Y Com­bi­na­tor might be im­por­tant for effec­tive al­tru­ism (and es­pe­cially cause pri­ori­ti­za­tion) and why I de­cided to look into it:

  • As a gen­eral prin­ci­ple, within rea­son­able bounds, try­ing to gather the think­ing of as di­verse a col­lec­tion of groups as pos­si­ble seems like a good idea, in terms of hav­ing ac­cess to a larger pool of ap­proaches with which to com­pare one’s own ap­proach and ex­pand­ing one’s imag­i­na­tion of what’s pos­si­ble. At the very least, it shows that one has “done one’s home­work”. More­over, look­ing be­yond self-pro­fessed effec­tive al­tru­ists and their fa­vorite go-to peo­ple and pro­jects can help to keep effec­tive al­tru­ism from seem­ing too in­su­lar.

  • Y Com­bi­na­tor has spent a lot of time think­ing about which for-prof­its to fund. While this is not di­rectly re­lated to the non-prof­its that effec­tive al­tru­ists tend to look into, Y Com­bi­na­tor’s for-profit work has clearly been valuable1, so we can ex­pect to gain in­sight in terms of the gen­eral task of “de­cid­ing which pro­jects to fund”.

  • As Y Com­bi­na­tor has now be­gun eval­u­at­ing non-prof­its, it’s now pos­si­ble to at­tempt a more di­rect com­par­i­son be­tween, say, how Y Com­bi­na­tor chooses its non-prof­its ver­sus how GiveWell does.

  • It’s pos­si­ble to ex­pand the ex­ist­ing dis­cus­sion of for-prof­its ver­sus non-prof­its (or more gen­er­ally, just talk­ing about non-prof­its in the con­text of for-prof­its or us­ing a “for-profit mind­set” of sorts).

  • There seems to be a class of peo­ple who are re­cep­tive to Y Com­bi­na­tor’s work (and the tech-sec­tor/​startup ap­proach to im­prov­ing the world) yet un­will­ing to em­brace “effec­tive al­tru­ism” (see for in­stance shminux’s com­ment in a LessWrong open thread and also pos­si­bly Effec­tive Altru­ism from XYZ per­spec­tive). Figur­ing out how effec­tive al­tru­ist ap­proaches differ from that of Y Com­bi­na­tor can elu­ci­date this differ­ence in re­sponse by this class to some ex­tent.

  • There is already in­ter­est in this topic. For in­stance GiveWell has been highly in­ter­ested in differ­ent ap­proaches to philan­thropy. They have even be­gun their His­tory of Philan­thropy pro­ject, which “aims to bet­ter un­der­stand the role that philan­thropy, as op­posed to other fac­tors, played in var­i­ous his­tor­i­cal epi­sodes that some peo­ple have iden­ti­fied as ‘philan­thropic suc­cesses.’ ”

Of course, it’s pos­si­ble that look­ing into Y Com­bi­na­tor’s non-profit pro­gram, or the ap­proaches of oth­ers, isn’t all that im­por­tant. It’s also un­clear whether this sort of in­ves­ti­ga­tion is tractable. As GiveWell has of­ten noted, or­ga­ni­za­tions don’t do a lot of ex­plicit pri­ori­ti­za­tion work,2 and even if they do, they don’t re­lease a lot of their ac­tual work/​thought pro­cess.3 It’s also un­clear whether groups hold them­selves to a high enough epistemic stan­dard, and even if they do, they may not be very good at ex­plain­ing their rea­son­ing.

Back­ground on Y Com­bi­na­tor’s non-profit program

Y Com­bi­na­tor be­gan fund­ing non-prof­its in 2013 when it in­cluded Watsi, a crowd­fund­ing non-profit for med­i­cal treat­ments, in its fund­ing batch. Y Com­bi­na­tor co-founder Paul Gra­ham wrote:

After we met the founders of Watsi, we re­al­ized they’d be the perfect [non-profit] to start with.

We owe Hacker News thanks for in­tro­duc­ing us to Watsi. When Watsi launched, they posted about them­selves on HN. They were a hit with HN users and raised a sig­nifi­cant amount of money.

Since then, they’ve also funded Zidisha, Bayes Im­pact, CodeNow, CareMes­sage, Noora Health, One De­gree, the Im­mu­nity Pro­ject, the Detroit Water Pro­ject, 80,000 Hours, Democ­ra­cyOS, and pos­si­bly more4. Note that this is still quite a small num­ber com­pared to all the star­tups they’ve funded so far.5

Y Com­bi­na­tor’s non-profit pro­gram is dis­tinct from YC Re­search, which is newer and tar­gets “fun­da­men­tal re­search” and in­no­va­tion that takes a long time to de­velop. I won’t con­sider YC Re­search in this post.

For the rest of this post, I want to go into how Y Com­bi­na­tor chooses which non-prof­its to fund, by look­ing at both what it says on the topic and also its ac­tual choices.

Y Com­bi­na­tor’s ex­plicit reasoning

Com­pared to an or­ga­ni­za­tion like GiveWell, with its com­mit­ment to trans­parency, there is rel­a­tively lit­tle that Y Com­bi­na­tor has said about its rea­son­ing in de­cid­ing which non-prof­its to fund.6 That isn’t to say they’ve been com­pletely ret­i­cent though. For in­stance, in one of the origi­nal an­nounce­ments about non-prof­its, Paul Gra­ham writes:

We don’t know how many non­prof­its we’ll fund, or what type of ideas we’ll like. We like Watsi a lot, be­cause they help peo­ple who re­ally need help, and do it in an effi­cient and trans­par­ent way. But fun­da­men­tally this is an ex­per­i­ment, just like YC it­self was at first. So we don’t want to tell peo­ple too pre­cisely what we’re look­ing for, be­cause we’re not sure yet what we’re look­ing for.

More sub­stan­tive is an in­ter­view done with Y Com­bi­na­tor’s Kate Courteau (who is Direc­tor of Y Com­bi­na­tor Non­prof­its). Courteau ex­plains how fund­ing Watsi led to them con­sid­er­ing more non-prof­its:

WATSI told us how much the YC pro­gram im­pacted their progress so we re­al­ized that a lot of the benefit of YC could be uni­ver­sally ap­plied to all star­tups; for-profit and non­profit. We also rec­og­nized that the WATSI team op­er­ated in the way that most great for-profit star­tups do. They were very de­ter­mined and fo­cused and moved quickly; launch­ing early and iter­at­ing based on watch­ing how their users in­ter­acted with their site. They had a vi­sion for the fu­ture of WATSI. It’s evolved a bit but they con­tinue to be a strong, mis­sion-based com­pany that works in the non­profit world in a very trans­par­ent way. This ex­per­i­ment seemed to work so well that in Jan­uary 2014, we de­cided to offi­cially launch our non­profit pro­gram.

More­over, she says the fol­low­ing, which might be the clos­est thing to un­der­stand­ing how they “se­lect causes”. The es­sen­tial idea seems to be that Y Com­bi­na­tor is us­ing the same frame­work with non-prof­its that they already use in de­ter­min­ing which for-prof­its to fund:

YC’s motto is: “Make some­thing peo­ple want.” In the for-profit space, we found that the best startup ideas come from founders solv­ing prob­lems they have en­coun­tered them­selves. This works much bet­ter than think­ing, “I’d like to start a com­pany. What should I work on?” This con­cept also ap­plies to non­prof­its. It’s re­ally im­por­tant to work on some­thing that you are pas­sion­ate about and have some his­tory with be­cause star­tups are re­ally hard.

The rhetoric is definitely differ­ent from what one hears from effec­tive al­tru­ist types: in­stead of “proven, cost-effec­tive, and scal­able” (GiveWell) or “the most good with your dona­tions” (Giv­ing What We Can), it’s ‘be­ing pas­sion­ate about the work you do’, and also ‘un­der­stand­ing other peo­ple’s cul­tures’7 and ’build­ing a user­base’8.

In part 2 of the in­ter­view, Courteau elab­o­rates on an­other part of Y Com­bi­na­tor’s thought pro­cess:

One of the biggest prob­lems in the non­profit world is the lack of early fi­nan­cial sup­port and men­tor­ship for young en­trepreneurs who have in­no­va­tive ideas about how to solve some of the world’s largest prob­lems. This lack of early sup­port is dis­cour­ag­ing to re­ally smart peo­ple who might have great po­ten­tial to tackle these big prob­lems. At YC we in­vest in peo­ple and we’re not afraid of teams with am­bi­tious or po­ten­tially po­lariz­ing ideas. We rec­og­nize that a lot of star­tups fail but we be­lieve that it’s worth the in­vest­ment even if there’s only a small chance that an am­bi­tious ap­proach can solve a re­ally big prob­lem.

In other words, un­like GiveWell’s tra­di­tional work, which tries to find the “very best” char­i­ties, Y Com­bi­na­tor takes a port­fo­lio in­vest­ment ap­proach, where it “doesn’t mat­ter” whether any par­tic­u­lar non-profit suc­ceeds or fails, as long as on av­er­age they do a lot of good. So the Y Com­bi­na­tor men­tal­ity is some­thing more like “try a bunch and see what works”, not “re­search a bunch and fund what works”.

Of course, given that Y Com­bi­na­tor’s main work fo­cuses on for-prof­its, it’s pos­si­ble to take what­ever ex­plicit rea­son­ing they’ve made pub­lic re­gard­ing for-prof­its as a proxy for their non-prof­its rea­son­ing, es­pe­cially since they ac­knowl­edge this in­fluence. I won’t elab­o­rate on this, but some places to look at are Paul Gra­ham’s es­says like “How to Start a Startup” and “Do Things that Don’t Scale”.

Overview of ac­tual non-prof­its Y Com­bi­na­tor has funded

In this sec­tion I want to list the non-prof­its Y Com­bi­na­tor has de­cided to fund so far, then try to see what might be some rea­sons they were cho­sen. At first I just want to list these and give brief de­scrip­tions of what they do be­cause I think it’s illus­tra­tive; I’ll give my thoughts af­ter the list. As men­tioned near the be­gin­ning of the post, the ones I’ll look at are Watsi, Zidisha, Bayes Im­pact, CodeNow, CareMes­sage, Noora Health, One De­gree, the Im­mu­nity Pro­ject, the Detroit Water Pro­ject, 80,000 Hours, and Democ­ra­cyOS.9

  • Watsi is a web­site that helps peo­ple crowd­fund med­i­cal treat­ments for peo­ple.

  • Zidisha is a “peer-to-peer microlend­ing ser­vice” that “al­lows peo­ple to lend small amounts of money di­rectly to en­trepreneurs in de­vel­op­ing coun­tries”.10

  • Bayes Im­pact is a bit vague; their web­site de­scribes them­selves as “a group of prac­ti­cal ideal­ists who be­lieve that ap­plied prop­erly, data can be used to solve the world’s biggest prob­lems”.11 Y Com­bi­na­tor ex­plains Bayes Im­pact in its an­nounce­ment as fol­lows:

    Bayes Im­pact is a non­profit that de­ploys teams of data sci­en­tists to cre­ate data-driven solu­tions for challeng­ing so­cial prob­lems. The or­ga­ni­za­tion runs full-time fel­low­ship pro­grams that bring to­gether do­main ex­perts and data sci­en­tists from top tech­nol­ogy com­pa­nies and aca­demic in­sti­tu­tions.

  • CodeNow runs work­shops de­signed to teach peo­ple pro­gram­ming. TechCrunch ex­plains:

    CodeNow aims to teach pro­gram­ming ba­sics to high school­ers, par­tic­u­larly girls, eth­nic minori­ties, and other un­der­rep­re­sented groups. It launched in Wash­ing­ton, D.C. in 2011 be­fore ex­pand­ing to New York City and San Fran­cisco last year.

  • CareMes­sage; TechCrunch on CareMes­sage (March 2014):

    Con­sid­er­ing that 70 per­cent of older pa­tients have difficulty us­ing and un­der­stand­ing printed ma­te­ri­als, CareMes­sage thinks this is where tech­nol­ogy can re­ally help, and is likely needed most. The Y Com­bi­na­tor-backed non-profit startup is launch­ing to­day to provide clinics and health­care or­ga­ni­za­tions with soft­ware and in­ter­ac­tive mo­bile pro­grams that help them stay in touch with (and en­gage) at risk pa­tients.

    It’s worth not­ing that “where tech­nol­ogy […] is likely needed most” is a bold claim.

  • Noora Health; TechCrunch ar­ti­cle on Noora Health (March 2014):

    The non­profit grad­u­at­ing from Y Com­bi­na­tor is build­ing hos­pi­tal ed­u­ca­tion plat­forms for pa­tients and their fam­ily mem­bers to teach them the skills needed to re­cover from a ma­jor med­i­cal event (like a surgery) or man­age chronic con­di­tions like di­a­betes and pal­li­a­tive care. Us­ing an iPad app, Noora Health works with hos­pi­tals to offer pa­tients and their fam­i­lies a com­bi­na­tion of videos, quizzes and in­ter­ac­tive con­tent to teach skills to aid in their re­cov­ery at home.

  • One De­gree is a web­site that helps peo­ple find com­mu­nity re­sources. The TechCrunch ar­ti­cle on One De­gree says:

    One De­gree, a new non-profit that helps peo­ple find so­cial ser­vices like af­ford­able hous­ing and job train­ing.

    […]

    “We are here to rev­olu­tionize the way that peo­ple ac­cess so­cial ser­vices,” Faustino said. “We be­lieve that we can do this be­cause the non-profit sec­tor has been stuck in the In­ter­net dark ages. Peo­ple still use binders and still rely on in­for­ma­tion that’s stuck in their heads. That means peo­ple aren’t get­ting the re­sources they need quickly and eas­ily.”

  • The Im­mu­nity Pro­ject (Wikipe­dia) aims to de­velop an HIV/​AIDS vac­cine.12

It should be noted that the vast ma­jor­ity of the non-prof­its in the list have to do with in­for­ma­tion tech­nol­ogy. The Im­mu­nity Pro­ject might be an ex­cep­tion in this re­gard (it still in­volves tech­nol­ogy, but does not aim to provide an app or web­site). This is prob­a­bly the case be­cause Y Com­bi­na­tor doesn’t just want to give money to the star­tups, but rather also wants to provide as­sis­tance, con­nec­tions, etc. Y Com­bi­na­tor has the most ex­pe­rience with in­for­ma­tion tech­nol­ogy, so this makes sense. And be­cause they have the most ex­pe­rience here, they also have more ex­pe­rience re­view­ing ideas here, which helps them pick good ideas.

Some of the star­tups (like CodeNow) may come across to effec­tive al­tru­ists as be­ing par­tic­u­larly “in­effec­tive”. In­deed, their re­sults sec­tion seems to con­sist of a sin­gle graphic and tes­ti­mo­ni­als from stu­dents. Here it might be worth not­ing that Y Com­bi­na­tor has a par­tic­u­lar au­di­ence they might cater to (con­sist­ing of pro­gram­mers and other tech peo­ple).

Com­par­i­son to GiveWell and the Open Philan­thropy Project

Be­sides the rhetor­i­cal differ­ences noted above, there are sev­eral other im­por­tant differ­ences be­tween Y Com­bi­na­tor’s non-profit pro­gram and GiveWell/​Open Phil.

  • Un­like GiveWell’s tra­di­tional work, which funds ex­ist­ing char­i­ties that have pro­duced sig­nifi­cant re­sults, Y Com­bi­na­tor’s non-profit pro­gram (with a few ex­cep­tions) tar­gets non-prof­its that are just start­ing.

  • Y Com­bi­na­tor also hasn’t for­mu­lated their se­lec­tion pro­cess ex­plic­itly as cause se­lec­tion. Rather, they seem to be pay­ing at­ten­tion to a lower level of se­lec­tion (com­pared to the high (cause) level se­lec­tion) in a par­tic­u­lar area (namely those in­ter­ven­tions in­volv­ing in­for­ma­tion tech­nol­ogy). In this sense, Y Com­bi­na­tor’s ap­proach might be similar to in­ves­ti­gat­ing all in­ter­ven­tions that in­volve leaflet­ing (what­ever the cause).

  • So far, Y Com­bi­na­tor’s non-profit pro­gram has been similar to the Open Philan­thropy Pro­ject and dis­tinct from GiveWell in that rather than recom­mend­ing non-prof­its for in­di­vi­d­ual donors, they provide the fund­ing them­selves. In­deed, for its part, the Open Philan­thropy Pro­ject has ex­plic­itly stated that their “de­ci­sions about whether or not to make a grant to an or­ga­ni­za­tion should prob­a­bly not be taken as a sig­nal to donors about whether or not they should donate”14. Y Com­bi­na­tor is not as ex­plicit as the Open Philan­thropy Pro­ject in this re­gard, and from what I have seen, they do not ex­plic­itly sug­gest peo­ple to donate to the pro­jects they fund, ei­ther (they just an­nounce their own fund­ing).

Doc­u­ment source and versions

The source files used to com­pile this doc­u­ment are available in a GitHub Gist. The Git repos­i­tory of the Gist con­tains all ver­sions of this doc­u­ment since its first pub­li­ca­tion.

Acknowledgements

Thanks to Vipul Naik, Jim Terry, and Ben­jamin Todd for their in­put while I was writ­ing this post. All im­perfec­tions are my own.

The writ­ing of this post was spon­sored by Vipul Naik.

License

This post is re­leased to the pub­lic do­main.


  1. Y Com­bi­na­tor (and Paul Gra­ham in par­tic­u­lar) has ar­guably had more im­pact than any ex­plicit “effec­tive al­tru­ist” group (or per­son). See for in­stance Vipul Naik in “Paul Gra­ham on US im­mi­gra­tion policy and high-tech pro­gram­mers”:

    I’m a great fan of Paul Gra­ham, es­say­ist, en­trepreneur, and co-founder of startup ac­cel­er­a­tor Y Com­bi­na­tor (along with his wife Jes­sica Liv­ingston, whom I also ad­mire greatly). Through Y Com­bi­na­tor, Gra­ham has changed the startup and tech com­pany land­scape and profoundly af­fected the world. (Some Y Com­bi­na­tor-funded com­pa­nies you’ve prob­a­bly heard of are Red­dit, Airbnb, Drop­box, Scribd, Disqus, and Stripe). Gra­ham also started Hacker News, a Red­dit-of-sorts for the pro­gram­mer/​startup crowd. In the world of let­ters, Gra­ham is bet­ter known for his long-form es­says that in­clude in­ci­sive so­cial com­men­tary. If you haven’t yet read his pieces, I en­courage you to check them all out (I par­tic­u­larly like this one, that might be some­what rele­vant here). He’s done more for the world than most peo­ple, in­clud­ing me, could dream of. And he knows a lot more about how the world works than I do.

  2. For ex­am­ple in Strate­gic Cause Selec­tion:

    [Cause se­lec­tion] is where ex­ist­ing philan­thropists seem to be least thought­ful and to ask the fewest crit­i­cal ques­tions; yet this is where we’d guess the bulk of vari­a­tion in “how much good a philan­thropist ac­com­plishes” comes from.

  3. See for in­stance Paul Chris­ti­ano:

    Un­for­tu­nately, [cause pri­ori­ti­za­tion] seems to be done pri­mar­ily in pri­vate; to the ex­tent there is high-qual­ity think­ing on these ques­tions, it is very rarely shared, and pub­lic dis­cus­sions rarely move for­ward our over­all state of knowl­edge or re­veal im­por­tant new ar­gu­ments.

  4. Y Com­bi­na­tor doesn’t seem to keep a com­pre­hen­sive list of star­tups they’ve funded, and when an­nounc­ing star­tups they of­ten don’t men­tion whether some­thing is a for-profit or a non-profit, which makes the task of com­piling a list of YC-funded non-prof­its difficult. See Y Com­bi­na­tor Com­pa­nies for maybe more that I didn’t man­age to find.

  5. There’s an­other list that con­tains fewer com­pa­nies.

  6. Y Com­bi­na­tor does seem to value trans­parency though; see for in­stance Kate Courteau:

    Our hopes are to make some head­way on some of the biggest prob­lems that still ex­ist in the world. […] I get frus­trated that there are still so many peo­ple work­ing on the same prob­lems and not shar­ing their find­ings. I re­al­ize that some of that has to do with com­pe­ti­tion for scant philan­thropic re­sources, but if there was a way to over­come this and get peo­ple to share, I think we could work a lot faster and have greater im­pact.

  7. Fur­ther on in the in­ter­view, Courteau ex­plains:

    Some­times in the non­profit world, peo­ple do a lot of plan­ning in iso­la­tion be­cause their users might ge­o­graph­i­cally be far away. I learned this les­son early when I worked for a non­profit, teach­ing a new build­ing tech­nique in Su­dan. We were very sur­prised that our con­cept and meth­ods were not uni­ver­sally well re­ceived in Africa. Our the­ory had been de­vel­oped in the US and we had not spent much time in this cor­ner of the world be­fore we rol­led it out. This was the root of our prob­lem — we planned too much with­out deeply en­gag­ing with the peo­ple who were to use our product.

  8. Courteau ex­plains:

    Build a com­mu­nity that re­ally val­ues your product. This ini­tial core user base will be the foun­da­tion of your com­pany and a spring point for growth.

  9. There is in ad­di­tion Giveffect, which is a for-profit that helps non-prof­its. I dis­cuss Giveffect in this foot­note, since it’s not ac­tu­ally a non-profit.

    The main source of in­for­ma­tion that dis­cusses the startup is a TechCrunch ar­ti­cle, “Y Com­bi­na­tor’s Giveffect Has Built A Shopify-Meets-Sales­force For Non-Profits”:

    Giveffect, part of [Y Com­bi­na­tor’s] cur­rent co­hort, has built a suite of cloud-based soft­ware that fo­cuses speci­fi­cally on the needs of non-profit busi­nesses, cov­er­ing ser­vices like ac­count­ing and CRM (in­clud­ing donor track­ing), through to fundrais­ing and crowd­fund­ing plat­forms, all built from the ground up with its tar­get cus­tomers in mind.

    It’s also note­wor­thy that Giveffect, un­like many oth­ers that pass through YC, is not ac­tu­ally a very young startup. It was founded in 2012 and already has more than 300 cus­tomers.

    […]

    So why go to YC? Mirza ad­mits that the trio was hes­i­tant at first about ap­ply­ing. “We did won­der if it was suit­able given that so many star­tups here are at a much ear­lier stage,” she said. “But based on the trac­tion that we’d seen other YC star­tups get, and the suc­cess sto­ries we heard, we de­cided that we could build a good com­pany with­out YC, but we could build a great en­ter­prise with it.”

    The Giveffect web­site lists more “tra­di­tional” char­i­ties. Giveffect, while not a new startup (as noted above), still fits the pat­tern of the non-prof­its dis­cussed in this post by be­ing soft­ware-ori­ented.

  10. TechCrunch has an ar­ti­cle on Zidisha that com­pares it to other micro-fi­nance pro­grams.

    There has also been some dis­cus­sion of Zidisha within the effec­tive al­tru­ism com­mu­nity.

  11. There is also a TechCrunch ar­ti­cle on the non-profit, with the fol­low­ing quote:

    “I was vol­un­teer­ing at home­less shelters for a lit­tle while,” says Duan. “I was shocked by the differ­ence be­tween this man­ual la­bor, this one-to-one ex­change, and the idea that im­prov­ing the ac­cu­racy of a [pro­gram­ming] model by 1% can af­fect mil­lions of users and bring in hun­dreds of mil­lions of dol­lars in rev­enue.”

    The ah-ha mo­ment was de­ter­min­ing that the same sci­ence can be lev­er­aged for so­cial good. “I had been work­ing with the City of New York im­ple­ment­ing civic solu­tions and rec­og­nized that the chasm be­tween the tech­nol­ogy that in­dus­try and tech com­pa­nies are us­ing and what’s available to civic or­ga­ni­za­tions is huge,” says Jiang.

  12. There is a Wall Street Jour­nal ar­ti­cle on the Im­mu­nity Pro­ject:

    Im­mu­nity Pro­ject will seek to suc­ceed where other efforts have failed, in cre­at­ing a vac­cine that can pre­vent healthy peo­ple from con­tract­ing HIV, the com­pany’s di­rec­tor of strat­egy, Ian Cin­na­mon, said.

    Some­what sur­pris­ing (em­pha­sis mine):

    The pro­gram has not only given Im­mu­nity Pro­ject $20,000 in non-dilu­tive fund­ing, but Y Com­bi­na­tor Part­ner Sam Alt­man, who is also the founder of mo­bile-ap­pli­ca­tion com­pany Loopt Inc., said in a state­ment that he would per­son­ally donate blood as the com­pany moves its tech­nol­ogy fur­ther through clini­cal tri­als.

  13. See “A Con­ver­sa­tion With Democ­ra­cyOS, The YC Non-Profit That Built A Latin Amer­i­can Poli­ti­cal Party”, “Why Y Com­bi­na­tor Funded a Rad­i­cal Poli­ti­cal Party in Ar­gentina”, and “Out in the Open: An Open Source Web­site That Gives Vot­ers a Plat­form to In­fluence Poli­ti­ci­ans”.

  14. Quote from a Face­book post by Claire Za­bel. See also Holden Karnofsky’s “Why the Open Philan­thropy Pro­ject isn’t cur­rently fund­ing or­ga­ni­za­tions fo­cused on pro­mot­ing effec­tive al­tru­ism”.