October 2017 Open Thread

This post aims to give blog readers and followers of the Open Philanthropy Project an opportunity to publicly raise comments or questions about the Open Philanthropy Project or related topics (in the comments section below). As always, you’re also welcome to email us at [email protected] if there’s feedback or questions you’d prefer to discuss privately. We’ll try to respond promptly to questions or comments.

You can see our previous open thread here.


To what extent does open phil plan to promote the idea of doing the most good possible to other large funders, perhaps in the future where capital may become open Phil’s limiting factor? Is this something that is important?

Are you aware of people in the EA community working on projects targeted at promoting the idea of thoughtfully seeking to do the most good profitable to large foundations, or aligning existing incentives of large foundations like status/respect, being able to create visible progress and results, legacy, etc with the goal of doing the most good possible?

Do you project that open Phil will be capital constrained, as opposed to opportunity/grantee constrained, within the next 7 years?

If you had 10 additional excellent full-time researchers / analysts, what research/reports would be highest priority for them to create?

I recall reading about the value of expressed interest in a cause.

Have you considered whether this may also apply at a more micro level, ie expressed interest in specific kinds of projects in all cause areas? (I suppose this is somewhat the case eg with criminal justice refom noting the strategies you look for and like.)

As a clarification, I suppose what I’m asking is a) is there some level of increased detail that could be published in terms of ‘what you want to fund’ in each given areas and b) if there is some increased level of granularity that is possible, what would make you/open phil decide to do that?

Have you / the team found any strategies particularly helpful when doing cause write-ups, particularly potentially non-obvious ones? Currently in my head doing a write-up involves reading all the high quality resources on a topic and talking to a number of experts. Anything else? Do you have a standard list of questions that people use when they are interviewing people? Any other tips? Thanks!

In areas with less opportunities but a high desire to fund / highly undervalued area, would you consider funding things that don’t require new organizations/grantees, eg perhaps executive coaches, exec recruiting services for COO type roles, or funding for chiefs-of-staff, exec assistants, or similar for key executives? Ie some set of things that may provide leverage to already great organizations working in great areas as a way to work around being constrained by a lack of excellent potential new grantees.

Or sales and marketing coaches, given how so many things involve sales (eg recruiting, advocacy, etc) and marketing.

Kind of like the A16Z vc firm with their service model, which seems to have been substantially helpful to some of their portfolio companies (eg a number seem to have gained millions of dollars in sales pipeline via A16z’s executive briefing center.)

Open Phil could have sales coaches, marketing advisors, exec recruiters and exec coaches available to some orgs. My expectation is it would be very helpful to the orgs, *if* they were eager to use them. You could validate potential effectiveness by talking with for profit companies that have used these services - my experience would be rave reviews for exec coaches and sales coaches from what I’ve seen.

What would make you/open phil investigate [funding ‘effectiveness increasing things that don’t require new organizations and can be provided for your existing most effective grantees, like exec coaches, sales coaches, exec recruiters on staff’] at a shallow level?

Would you consider funding “labs” designed to spin off multiple non-profits in a given area? Eg like Science Inc.

How are you thinking about founding non-profits that you want to fund, these days? I recall reading about that on the blog.
What is the limiting factor behind why open Phil hasn’t co-founded non profits you want to fund but that don’t exist, yet? Eg there aren’t ones that make sense for a given reason, time/investment required, uncertainty of if it would work, etc. Does the calculus change if it’s funding a separate org that is designed to spin off multiple non profits in an undervalued area? (Eg an org that recruits founding teams and creates important AI safety non profits, or important EA community non profits, etc)

Have many of the team members read Principles by Ray Dalio? It seems to be a generally good meta-guide to achieving goals.

Has open Phil thought about promoting “definite optimism” as a way to increase the stock of people working on making the world better, in the long term?

What might make you/open phil investigate [increasing ‘definite optimism’] at a shallow level?

How valuable do you believe it might be to increase “the number of people who care (as demonstrated by their actions) about doing good,” as opposed to specifically “the number of people who care (ditto) about doing the most good?” Eg is getting 100 people to care strongly about doing good more valuable than getting one person to care strongly about doing the most good they can?

A related question is to what extent to you believe the best way to grow the EA community is to take people who already care deeply about doing good and help them care more about effectiveness vs take people who aren’t actively focused on doing good straight to doing the most good vs a separate focus on taking people to an active focus on doing good, and then later converting them to EA/doing the most good?

Additionally to the other qs, what might make you/open phil investigate [increasing the number of people who demonstrate with their actions that they care about helping others, but not necessarily (yet) in the most effective way possible] at a shallow level?

How might you think about the value of creating (effective) interventions that could take people and make them care more about other people and doing good, but not necessarily most effectively - would you view something like this as spreading the EA community? I suppose the question is also to what extent does open Phil value spreading altruistic desires / a culture of “kindness” in parallel to promoting effective altruism?

Would open Phil consider evaluating which problems are most “up-stream” - eg criminal justice issues may be in some part a symptom of personal fear/personal issues, and this may also relate to anti-immigration population segments, and possibly others? I wonder if then seeking out grantees working on things to help people at a meta level have less fear/personal issues could have down stream benefits on a number of causes today, and other challenges that may come up in the future?

What might make you/open phil investigate [which problems are most up-stream of existing focus areas and could serve as long-term ways to reduce the emergence of other future problems, e.g. fear/personal issues/lack of belonging, and how that may contribute to anti-immigration, anti-criminal justice sentiment] at a shallow level?

(Similarly I recall reading something implying that many KKK and gang joiners are doing so in part because it provides them with a sense of belonging and shared purpose.
What might make you/open phil investigate [ways to create sense of belonging and shared purpose for non-religious people] at a shallow level?)

“While our funding on these issues could easily be spread out across the country, one possibility worth discussing is whether we should “go all in” on a few states to be able to make the most impact and get better feedback about how our grants are working on the ground.” From the 2014 criminal justice report.

How has Open Phil’s thinking updated on this, and if you’re able to briefly share, what evidence or hypotheses led to the update?

“30%–40% of healthcare expenditures in the USA go to help address issues that are closely tied to the excess consumption of sugar.” 30% * ~$3 trillion = $900 million/yr, 40% * ~$3 trillion = $1.2 trillion/yr.

Source: https://research-doc.credit-suisse.com/docView?language=ENG&source=ulg&format=PDF&document_id=102245…

It seems like [excess consumption of sugar] could fit in a very high category of estimated importance, e.g. the mid-level estimate of labor migration was $150bn/yr and the high-level was $3tr/yr in the back of the envelope calcs from a while back: https://docs.google.com/document/d/1DTl4TYaTPMAtwQTju9PZmxKhZTCh6nmi-Vh8cnSgYak/edit#, so [excess consumption of sugar] could be in that similar range.

Additionally, it seems like companies that make lots of money from sugar have been/are investing heavily in “hooking” the international/developing world, which will presumably lead to extremely high costs in the future. I wonder if these high future costs could be avoided more cost-effectively with investments today, before developing world obesity rates catch up to the US.

Also seems like a field where increased investment will come from other sources over the next ~10-15 years, and where smart initial investment and research from Open Phil could influence the spending of other governments/foundations in the future. (I have the sense that Open Phil’s approach is perhaps on average more thoughtful/reasoned in any given space than other foundations, and so I wonder if Open Phil becoming involved in the early stages of a space would positively influence the later funding that comes from other foundations, e.g. if Open Phil funds the organizations it believes are best, those organizations can develop further, and thus may be in a better position to be sought out from other foundations ~10 years later.)

Might Open Phil consider exploring this area [excess consumption of sugar]?

What might make you/open phil investigate [excess sugar consumption, domestically but also particularly world-wide where it seems high-sugar companies may be investing for future sugar consumption growth] at a shallow level?

Any ballpark ranges for how long past shallow-depth reports, and separately medium-depth, have taken to create? Perhaps separately in terms of days/weeks/months between beginning the report to the report being “done” for internal purposes, and also just in terms of collective number of hours of input/work from the report beginning to being done.

Which ‘showings of appreciation/gratitude’ have open phil team members received that were most appreciated / made the team members feel most good?

I would like to say a very big thank you to the entire team, as I think that it is wonderful to have all of you working to make the world/future better, and I am very grateful for all the time/effort/thought/resources you all put in.

Thanks Arik! We appreciate the kind words, as well as the thoughtful questions.

Would you be up for choosing 5-10 questions you’re particularly interested in responses to? The number you’ve posted would take more time than we generally aim/expect to take with these open threads.

My pleasure! And thanks for noting so, and thanks for asking - I’d be happy to do that, and I’m glad that you requested that (particularly vs the possible alternative of skipping over the questions because there were too many).

I’ll collect my questions, and then post a new question thread with the 5-10 that I think have the most potential to be useful.


This grant page gives a 65% probability that “CEA provides compelling evidence that standard marketing and outreach (e.g. online ads) results on average in a new GWWC pledge for less than $2,000”.* I find this surprising given that

1. Open Phil has previously claimed that it’s pessimistic about the cost-effectiveness of online ads for creating new vegetarians
2. Signing the GWWC pledge is a comparably significant life change to becoming vegetarian, so on priors we’d expect ad campaigns for each to be roughly similarly successful

What’s the reason for this discrepancy?

*sentence structure changed for clarity

The claims I’ve seen about online ads for vegetarianism (which you describe us as being skeptical of) generally estimate a “cost per significant behavior change” of far less than $2000. For example, the figures on this page imply something like $.01-$2.50 per pledge to go vegetarian, which is 1,000x-100,000x the implied cost-effectiveness for our prediction re: GWWC pledges. As a more minor point, I think that “going veg*n” is a significantly greater behavior change than taking the GWWC pledge; the latter is nonbinding (and we accordingly value it at less than its literal “face value”) and requests ~yearly action, whereas the former implies a change to daily behavior.

Thanks for the explanation, Holden. I want to drill into this a little more.

Ultimately, the question we care about is, should I fund this? So the specific decision I want to question is whether it’s a good idea to fund GWWC ads but not veg ads (or other similar consumer-side interventions like leafleting).

I expect that online ads cost (much?) more than $2.50 per actual vegetarian created. But creating a new vegetarian probably doesn’t cost more than creating a new GWWC-pledge-keeper (which is a subset of people who sign the GWWC pledge). I would value a GWWC-pledge-keeper more highly than a vegetarian, but not by such a margin that GWWC ads would look substantially more cost-effective than veg ads. (Although even if GWWC ads are way more effective, veg ads are still worth buying if they have a better expected value than GiveDirectly.)

Then we have to value GWWC pledge-keepers against vegetarians. To make this quick I’ll just make up some plausible numbers. A GWWC donor is worth about $5000/year in donations, and let’s say on average their donations do as well as money going to GiveDirectly, which IIRC produces one QALY for about $1000, so that’s 5 QALYs per year of donations. A vegetarian-year is worth about 3 land-animal-years plus some hard-to-estimate number of fish years; let’s say it’s 10 total after adjusting for uncertainty about fish sentience. I would value preventing 10 factory farming years over creating 5 human QALYs*; I imagine you don’t, but they should at least be on the same order of magnitude.

That means

1. on priors, we would expect GWWC pledge-keepers and vegetarians to cost roughly the same to make (easily within an order of magnitude);
2. some evidence suggests that creating vegetarians costs much less than creating pledge-keepers, but this evidence is very weak;
3. by a rough estimate, a GWWC pledge-keeper does on the same order of magnitude of good as a vegetarian.

So based on this naive estimate, if GWWC ads are worth funding, so are veg ads.

And, according to OpenPhil’s standard of funding anything better than GiveDirectly, even if the cost-effectiveness of GWWC/veg ads looked substantially worse, they’d still both be worth funding because they look a lot better than GiveDirectly according to current estimates.

*I would value one chicken year at maybe 0.5 human years, but also I think a factory farming year is substantially more bad than a normal year is good. (I personally would give up 2-10 years of normal life to avoid one year on a factory farm, although I’m pretty uncertain about this because obviously I’ve never lived on a factory farm.)

P.S. I love that there’s now an option to get notified when replies are posted to my comment specifically, rather than getting notified for every comment. Makes threads much easier to follow.

Forgot to mention and I can’t edit comments, but I wanted to thank you for taking the time to engage on this.

Oops, accidentally screwed up my notifications. Please reply to my first reply and not my other comment where I replied to myself.


Leave a comment