EA Meta Fund Grants - July 2019 2019-09-13T15:40:59.914Z · score: 20 (10 votes)
EA Meta Fund Grants - March 2019 2019-05-23T08:55:23.292Z · score: 61 (30 votes)
EA Meta Fund AMA: 20th Dec 2018 2018-12-19T18:06:02.484Z · score: 21 (11 votes)
EA Meta Fund Grants Report: Nov 18 2018-12-03T10:45:26.767Z · score: 23 (15 votes)


Comment by agdfoster on A general framework for evaluating aging research. Part 1: reasoning with Longevity Escape Velocity · 2019-08-02T11:25:11.787Z · score: 2 (2 votes) · EA · GW

I work with donors and feel I don’t have enough to say when ageing comes up. Fancy giving me a quick primer?

A few things I’m clearly confused on:

  • why aging isn’t always discussed under the restriction of global population capacity, seems like we should have reached it by then (assuming no AGI). I realise this is everyone’s first intuitive response but seems like it should still be a key factor in any analysis.

  • seems like under a full capacity assumption the impact either comes from an improved distribution of where people are at in life (less teenage depression and end of life suffering or something like that) which seems uncertain. Or from increased career length which also seems not certainly net positive. Maybe I missed it but couldn’t see any of this in your calc discussion.

  • I know there are plenty of arguments for aging being neglected and so worth it from the medium term also but it seems like if you’re talking about escape velocity then pop cap needs to factor in.

If you’d be up for that primer I’m On my username (Agdfoster) and

Comment by agdfoster on Do we know how many big asteroids could impact Earth? · 2019-07-08T01:43:59.951Z · score: 4 (3 votes) · EA · GW

Don’t know much about this but I thought you could estimate a ballpark for the total frequency by looking at craters on moons and mars.

Comment by agdfoster on Corporate Global Catastrophic Risks (C-GCRs) · 2019-07-02T12:43:47.243Z · score: 4 (3 votes) · EA · GW

Is it meaningful to say that some companies are growing exponentially when the markets they take part in are growing exponentially?

After a quick Google, Microsoft was $500bn market cap in it's 2000 peak and now is about $1tn, but world GDP has almost doubled in that time also (60tn -> 110tn).

Comment by agdfoster on What is the effect of relationship status on EA impact? · 2019-06-27T17:11:34.375Z · score: 34 (14 votes) · EA · GW

If you want to try a work strategy that involves long hours then a positive successful relationship may be harder to achieve.

Otherwise I’d advocate you don’t instrumentalise your non-work time for impact. I know it’s a cliche but do what you enjoy. Instrumentalising your free time seems to make people less relatable (probably an understatement), less trustworthy, more prone to depression, less robust to sudden changes etc

Having a strong base, whatever it is for you, is pretty important I think. When impact stuff is going badly you don’t want to feel like everything is going badly. That kind of instability is going to have more Long term effect for your impact than a few thousand pounds a year in one direction or another.

Comment by agdfoster on How to evaluate the impact of influencing governments vs direct work in a given cause area? · 2019-06-26T12:08:43.468Z · score: 4 (3 votes) · EA · GW

As you suggest in the question I think an improvement would be: "impact of -org X- -trying- to influence governments vs direct work"

Some key considerations:

  • Gov interventions' numbers in my experience generally have much better expected values in back of the envelope models and then often look much less good when you add a bunch of additional intuitive discount factors.
  • Attribution is highly uncertain with policy interventions.
  • This said, I expect many of the highest return per dollar funding opportunities to be in policy and research.
  • What's the tractability/chance of success of the government intervention for the given team? It might actually be very very low.
  • Is this team more likely to do more harm than good? There are many ways of doing damage or damaging future efforts. I think this rules out a bunch of approaches that would otherwise justify their low tractability and
  • How time pressured is the intervention? Potentially risk is worth it. My general rule of thumb currently on this is something like "major risks are very rarely worth it, be very careful and wait for great te
  • Sometimes "direct work+ " is the best policy intervention. Do something the government could adopt really well, show that it works and has public support, then try lobbying for it. This said, it seems like quite a few charities just strap this argument on to what they were going to do anyway, which is fine so long as they've thought it through properly, just something to look out for.
  • Lobbying / technical assistance etc opportunities are generally going to be fewer and further between.
Comment by agdfoster on EA Meta Fund Grants - March 2019 · 2019-06-12T10:49:28.418Z · score: 1 (1 votes) · EA · GW

Thanks for flagging, now updated

Comment by agdfoster on Cash prizes for the best arguments against psychedelics being an EA cause area · 2019-05-10T18:38:41.657Z · score: 19 (14 votes) · EA · GW

I’d guess the best argument is the obvious one:

  • most the professional world and voting populace have a very negative view on psychedelics
  • whilst the potential upsides might be sizeable, they likely don’t compare to the negative damage to EA that EA orgs publicly supporting such work would likely do.
  • if done in secret that’s a) a secret (generally bad) and b) inevitably going to get out.
  • and a fair number of non EAs are working on it anyway as it’s quite a popular idea in California. I’m guessing anyone super passionate about it could get funding and hire without having to be associated with EA at all.
Comment by agdfoster on Is EA ignoring significant possibilities for impact? · 2019-05-10T18:29:44.571Z · score: 12 (6 votes) · EA · GW

I think your reasoning here needs a lot of work. Few quick points:

  • better to critic specific points rather than something broad like ‘all strategy of EA affiliated orgs’.
  • generally, if it seems like a large number of really smart people in EA appear to be missing something, you should have a strong prior that you are the one missing something. Took me a long time to accept this. It’s not wrong to shine a light on things of course, but a touch more humility in your writing would go a long way.
  • reasoning and evidence aren’t exclusive things, evidence is part of reasoning.
  • this said, I don’t think the criticism of “too evidence based” sticks anyway, have you read much academic ea research recently? Maybe in poverty.. but that’s a very busy area with loads of evidence where most approaches don’t work so it would be pretty crazy not to put heavy weight on evidence in that cause area.
  • Jude’s spends 2.1m a day but given the differences between the impact p dollar of projects easily gets into the order of 100s-1000s this isn’t very relevant.
  • OpenPhil could spend that. There are complex reasons why it doesn’t but the main thing to note that total spend is a terrible terrible signal.
  • for profit models have been explored numerous times, while still promising, little really great stuff has been found. People are working on it but it’s not a slam dunk.
  • earning to give is a great way to build career capital and do good.
  • advocacy and philanthropic advisory is really hard. People in that area are going as fast as they sensibly can.
  • it takes a long time to become a chief of staff at a powerful org.
  • policy / lobbying approaches are really hard, and people are again working on it as fast as they can.
Comment by agdfoster on Will splashy philanthropy cause the biosecurity field to focus on the wrong risks? · 2019-04-30T18:12:18.646Z · score: 18 (13 votes) · EA · GW

Really interesting read, a few thoughts below. Only skim read article so mostly responding to your prompts. I should also note I advise philanthropists for a living and so am inherently biased!

  • I’ve found Open Phil’s reasoning to be rigorous and thorough, far more so than virtually all of their peers. I also have deep intellectual trust for, so far without exception, everyone I’ve met that works there.

  • from a skim read the OPs arguments feel pretty zero sum. Perhaps the argument should instead be “should open Phil also fund non-GCBR bio work as well”.

  • It doesn’t seem even handed to both portray these researchers as easily swayed by flashy deep pocketed philanthropists as well as lamenting the loss of highly intelligent research talent. If they’re highly intelligent and also updated their actions based on open phils reasoning (albeit also including cash), the OP should probably be humble themselves about the likelihood of being right.

  • the OP seems to present philanthropy as this potentially negative steering force. Even if the field is zero sum (gov funds less as a result / too little talent to use extra funds wel), are we to believe that altnerate funding sources apply no directional pressure?

  • whilst voting keeps governments relatively aligned with the populace’s needs, it has only a small alignment with global needs and global public goods. The short time frame (4 years) also seems to result in shorter-term thinking. Future generations can’t vote. Philanthropy seems uniquely well positioned to be reasoning and funding in areas poorly tended to by the democratic system and markets.

  • a bunch of the arguments wouldn’t seem intuitive if re-applied to other more familiar causes like climate change or global poverty, reasons for difference should be highlighted and then the extent of the OPs arguments capped respectively.

Comment by agdfoster on Getting the Rich and Powerful to Give · 2019-04-29T09:34:06.150Z · score: 8 (6 votes) · EA · GW

From experience this gets more and more desirable both the wealthier a donor is and the longer the donor has been wealthy for.

I think the papers phrase “agency” is missing something though. It’s not a sensation of making decisions or being listened to that people are drawn to generally. I’d say it was more ‘proprietorship’ or ‘feeling ownership over the project’, feeling like a causally relevant component of the whole endeavour.

Agency is also, I find, quite patronising!

Comment by agdfoster on Why does EA use QALYs instead of experience sampling? · 2019-04-26T12:46:12.770Z · score: 3 (2 votes) · EA · GW

That's fine! Thanks.

Comment by agdfoster on Why does EA use QALYs instead of experience sampling? · 2019-04-25T10:31:53.748Z · score: 8 (6 votes) · EA · GW
[Disclosure: In February 2019, I corresponded about the experience-sampling idea with Alex Foster of the EA Meta Fund. He said my points were "certainly quite compelling," but the correspondence fell off.

Please note that the content of my correspondence with Milan was exploratory but primarily from a position of skepticism. Whilst technically accurate I find this quote to be misleading and not very good form.

Comment by agdfoster on Who is working on finding "Cause X"? · 2019-04-15T21:07:45.896Z · score: 3 (2 votes) · EA · GW

Arguably it was the philosophers that found the last few. Once the missing moral reasoning was shored up the cause area conclusion was pretty deductive.

Comment by agdfoster on After one year of applying for EA jobs: It is really, really hard to get hired by an EA organisation · 2019-04-15T14:59:04.164Z · score: 2 (2 votes) · EA · GW

haha - good question. And yes, from notes.

Comment by agdfoster on How Flying Cars Will Solve Global Poverty · 2019-04-06T14:15:01.535Z · score: 2 (2 votes) · EA · GW

I hate April 1st so much.

Comment by agdfoster on How Flying Cars Will Solve Global Poverty · 2019-04-02T09:27:06.854Z · score: 1 (1 votes) · EA · GW

A small gripe with the title - you don’t make any argument for this tech solving global poverty, just congestion in the wealthiest cities on earth. I know transportation has economic benefits elsewhere but your post makes no claims about this.

Comment by agdfoster on How Flying Cars Will Solve Global Poverty · 2019-04-02T09:22:52.514Z · score: 1 (3 votes) · EA · GW

I want more posts about flying cars.

I’m still assuming the reliability requirement is too high. If a car stops working it rolls to a halt, a flying crashes into a residential area. Planes don’t do this, but they have costly constant checks. Maybe a fleet owner (non personal ownership) and lots of sensors for automated checks makes the reliability feasible.

Similarly security seems like a daunting challenge.

Noise I hadn’t thought of.

Do we even need them though, if a city goes full AV you can theoretically have very high speed regular cars and no junctions / traffic. At even just 60mph, a 30 min commute encompasses an area significantly larger than Greater London. And commuting in an AV could be very comfortable with a desk and WiFi. Whilst it’s hard to work on trains I could imagine even “going for an AV pomodoro” in the middle of the day just for the concentration benefits of reclusion and a fixed travel time.

Assuming good automation is required for good flying cars, I’m also not sold on automation being net-good for employment. Life satisfaction - sure.

Comment by agdfoster on a black swan energy prize · 2019-03-29T14:02:42.516Z · score: 2 (2 votes) · EA · GW

I drafted but didn’t publish a post yesterday titled “where are all the ideas?”. Really glad to see a contribution of this type.

Comment by agdfoster on Request for comments: EA Projects evaluation platform · 2019-03-22T16:48:20.787Z · score: 14 (6 votes) · EA · GW

I regularly simplify my evaluations into pros and cons lists and find them surprisingly good. Open Phil’s format essentially boils down into description, pros, cons, risks.

Check out kialo. It allows really nice nested pro con lists built of claims and supporting claims +discussion around those claims. It’s not perfect but we use it for keeping track of logic for our early stage evals / for taking a step back on evals we have gotten too in the weeds with

Comment by agdfoster on Primates vs birds: Is one brain architecture better than the other? · 2019-03-05T10:25:06.790Z · score: 6 (5 votes) · EA · GW

Really love how clearly you’ve communicated the relevance of the findings and how they fit in.

Comment by agdfoster on Latest Research and Updates for February 2019 · 2019-02-28T17:21:18.794Z · score: 3 (2 votes) · EA · GW

This is great - thanks for continuing to do these roundups, always things I’ve missed

Comment by agdfoster on Why you should NOT support Aubrey de Grey's work on ageing. (maybe) · 2019-02-28T16:51:01.405Z · score: 1 (1 votes) · EA · GW

Seconded on title, enjoyed content but title felt click-baity and misleading, especially given 90% of readers will only read the title.

Comment by agdfoster on Effective Impact Investing · 2019-02-28T12:42:20.049Z · score: 11 (5 votes) · EA · GW

I’m so glad to see a post from people working in the industry in question - thank you for taking the time, making the post and contributing to the discussion, I strong upvoted.

Impact investing comes up a lot in donor advisory, so have a few points to add:

1 I generally still advise donors not to use their philanthropic, impact maximising, allocation for impact investing. I still do not have a very thorough way of explaining how I came to this conclusion and no existing materials I know of could be sent to a HNW.

2 most large donors only give a small portion of their assets, >90% of their capital is generally invested and impact investing (II) can come from that allocation. This changes the discussion considerably. Eg from ‘II vs donating’ to ‘where are you spending your time’. We could also discuss lost returns but it seems like most II does not have sufficiently lower returns on average to worry about it from an impact perspective. Giving better will make significantly more of a difference than 5-20% more or less profit from their investments. Not always going to be true but generally seems the right view.

3 donor funds are a very different question to an individuals career, the leverage available in some impact investing careers is so high that it requires a separate investigation that I haven’t seen. I would guess that an evaluator could move 10-100x more capital for the same effort in II than in donation evaluation.

4 stepping back, effectiveness minded II might be an important consideration in designing the models orgs within top cause areas consider. If you were comparing two models for your org and one assumed limited philanthropic capital and so maximised its impact per dollar but the other assumed huge swathes of II funding and so took on a much lower impact per dollar, and tried to design a model that would be copied, build IP, be acquired by industry giant etc..

5 bio / ai safety have many opportunities for doing far more damage than good, at face value it does not seem a good fit for larger, IP-driven investment models. However, a big worry is control and trust. If top researchers and strategists in the area felt there was capacity for responsible cautious impact investing the area, that might speed up how quickly market driven approaches emerged.

Comment by agdfoster on After one year of applying for EA jobs: It is really, really hard to get hired by an EA organisation · 2019-02-28T10:49:04.792Z · score: 35 (18 votes) · EA · GW

I ran my first hiring process to hire someone for an EA role last year and was amazed how long it took me. I’ve hired around 20 times in the past and only spent a couple weeks and 20-40h per role. Last year I spent 8 months and hundreds of hours. I reflected afterwards on why and can list a few hypotheses:

  • normally rely heavily on gut to build my shortlist. Did not feel comfortable doing this for this role as it felt like there were so many failure modes for a bad hire. Both ways that a hire could go badly and severity of impact for a hire going badly.

  • normally relying on intuition heavily is highly reversible. Worst case scenario I have to fire the candidate after probation, I’ll never see them again and no one knows them. I’m open with candidates that this is my policy and that they should be careful accepting an offer. In EA I felt like everyone knows everyone and a fired hire could cause significant reputational damage with a one-sided narrative. I don’t endorse this view as rational but the fear was definitely a factor in why I took so long.

  • I was hiring for a role that defies regular role definition. No one applying to the role had applied to a similar position before let alone worked in one. Potentially this was the largest factor and my other points are moot.

  • I wasn’t hiring someone to have similar skills to me, instead hiring someone to have the skills I don’t have. Normally I would judge experience, passion, intelligence, lateral thinking ability, ambition and team fit then let a team lead judge specific ability.

  • many candidates treated the process like a 2 way application the whole way through the process. This three off my intuitions and normally I would have dropped all candidates who weren’t signalling they were specifically very excited about my role. First call excluded.

  • many candidates’ conversations included career advice from me. This threw off my intuitions but I consider it time well spent in ll the cases where I spent over 2h

  • I worried a lot about how much time of others I was using. Assuming a candidate spent 4x more time than I did, I used over a thousand hours of people’s time.

  • ultimately I made offers to two candidates both of which I had had strong gut feelings about very early, which was rewarding but also highly frustrating.

The key thing I intend to change next time is being much faster. I didn’t feel like (for me) the extra process complexity and caution added that much insight and crucially, it threw off my intuition.

The main downside of reduced complexity seems to be the increased chance of a bad hire and the potential damage of firing them. I think next time I will return to my original method and be very transparent with the person I make an offer to that their 3 month probation is not just a formality, pointing them to this article as an explanation to why it’s not worth it for others for me to have a long drawn out process that may only slightly reduce the risk of a bad hire.


** I do not advocate anyone else doing this unless they are confident in their hiring intuitions. I also haven’t tried it yet and it may go terribly. **

Thank you to the OP for posting. Illuminating!

Comment by agdfoster on Impact investing is only a good idea in specific circumstances · 2019-01-09T18:06:37.087Z · score: 2 (2 votes) · EA · GW

Found Bridgespan's 2018 report useful and interesting.

Comment by agdfoster on List of possible EA meta-charities and projects · 2019-01-09T18:04:02.982Z · score: 3 (4 votes) · EA · GW

Nice list Saulius, thank you.

Comment by agdfoster on How should large donors coordinate with small donors? · 2019-01-09T17:40:13.840Z · score: 7 (3 votes) · EA · GW

[idea]: Invite-only Google Sheet List of considerations relevant to funding a group (one group per tab) and then columns of donor's weights for those considerations. I would find this really interesting.

Deal could be that you only get access if you're willing to share your weights!

Comment by agdfoster on EA orgs are trying to fundraise ~$10m - $16m · 2019-01-07T21:17:49.475Z · score: 25 (10 votes) · EA · GW
For instance, like other big non-profits, EA orgs might want to hire institutional fundraisers to tap into larger grants from big foundations other than the usual suspects

I've looked into this a few times and it does seem like it will become a promising channel. In particular from the big donors that do very large checks (>$500k). At least one org I know is experimenting with hiring a full-time grant-writer. I currently think it won't work well for most EA orgs for some time to come.

Worth noting that most big foundations have large sr. management time overheads and often require designing bespoke projects just for that foundation. Grant-writers also generally have slow payback periods (>1-2 years not rare, more if first one doesn't work out), are very tricky to evaluate during hiring and expire once you run out of foundations / major donors to apply to (most don't do much repeat funding). Not insurmountable challenges.

An alternative is to hire one-off fundraisers who approach lesser known major donors for you, I think that may be promising but requires a large time investment to train that person to talk about your charity. They also still require a large amount of Sr. mgmt's time (non-foundation major donors will generally want to speak to the founders and most those conversations will end up being a no) and are more likely to generate one-off funding rather than repeat.

It may be that expanding philanthropic advisory within EA in general is more promising. Whilst not specifically focused on raising funds for EA orgs, an increased number of smart and best-arguments-aware donors in the space in general could well have a similar result for less sr. mgmt time cost.

It could also be that having a semi-centralised fundraising team that manages a team of generalist fundraisers that are shared and specialist fundraisers that each work for a different large EA org could work really well. Train them all up in tandem and work out how to evaluate them, focus on >$300k checks from major donors but also have a grant-writer or two shared between them, hire most talent from mainstream pools, etc., We looked into something like this to function across all the GiveWell charities but it ultimately looked like it wouldn't work.

It doesn't seem unlikely that that last option never makes sense, because by the time you have orgs large enough to justify the above ($5m-$10m pa), those orgs also organically start to hire their own internal fundraisers and grant-writers just to meet their large budgets.

Looking forward to putting more thought into this.

Comment by agdfoster on The Global Priorities of the Copenhagen Consensus · 2019-01-07T20:51:40.481Z · score: 2 (2 votes) · EA · GW
Social protection system coverage (helping more people access government benefits); CC estimates that this is less than one-fifth as valuable as cash transfers

That is surprising, they've done a lot of work in and around India where welfare budget utilisation has been infamously poor until only quite recently and where the huge rural population seems to make it particularly hard to get welfare to the poorest who need it most.

I wonder how their economists account for the counterfactual of unused government funds, I've seen quite a few calcs where unused welfare funds that go back into the central pot are only discounted by 1/4 - 1/2, which I still find very unintuitive even given only that the average wealth of a government spending recipient is far higher than that of a recipient of welfare.

I've been keeping an eye out for a charity / org in India that is particularly good at increasing access to government welfare so this is relevant for that.

Comment by agdfoster on EA Giving Tuesday Donation Matching Initiative 2018 Retrospective · 2019-01-07T20:36:04.629Z · score: 7 (5 votes) · EA · GW

Really impressed by both how you've executed this as well as the write-ups. 🙌

Thank you!!

Comment by agdfoster on Challenges in Scaling EA Organizations · 2018-12-21T16:04:04.345Z · score: 3 (3 votes) · EA · GW

Awesome think-piece, thank you

Comment by agdfoster on Response to a Dylan Matthews article on Vox about bipartisanship · 2018-12-21T16:00:58.745Z · score: 24 (16 votes) · EA · GW

Note: I have a feeling that 'policing tone' is an annoying meme for a forum and something more appropriate for moderators than for readers so I'll post this one and default to refraining from doing it again.

Quick few thoughts on the tone of this, feel free to ignore if it doesn't change your mind:

Most of these articles have been good but this one is certainly the worst out of all that I have seen (n=25 or so, from multiple writers) and I believe it has negative expected value.

This part, right at the top and at a few other points, I felt a little uncomfortable. If I were the author and I read this, I feel like I would feel more 'attacked by my allies' than 'constructively critiqued'.

I feel like some quick changes to the tone, particularly early on (e.g., 'I found this one distasteful' rather than 'this is the worst I've seen') feels less aggressive. Perhaps adding an extra paragraph at the beginning saying a few positive things about their column in general (if you have those views) and saying that you only mean your comments in a constructive way. Personally, that would be enough for me to take the feedback well. Maybe no one on their team reads it, maybe one person reads it and forwards it to the whole team. Seems worth assuming the latter is the case and it's that scenario that prompted me to make this comment.

Given in particular that Future Perfect is not funded by donors that explicitly identify with EA ideas and that run by Vox, my quick guess is that careful constructive criticism is far more valuable / low risk than more assertive / slightly aggressive criticism (apologies if I'm already preaching to the choir here). I'm currently still really glad that Vox, Ezra etc., have chosen to do this column taking lots of EA ideas into account.

Funnily enough I had a similar opinion about one of the mobile thumbnails for their anti-mars piece. The thumbnail read "Elon wants to go to Mars, here's why that's a bad idea" which didn't seem worth it.

Comment by agdfoster on EA Meta Fund AMA: 20th Dec 2018 · 2018-12-20T21:10:52.094Z · score: 3 (3 votes) · EA · GW

Some part of the large potential upside of this fund, and the reason why some of the team are excited to put so much of their time into it, is that if we do a really good job it could grow and attract significant additional capital into the meta cause areas.

Whilst a relatively small cause area, in a more efficient market for non-profits, I would expect that the space was funded to the brim due to the outsized returns available within it. I see moving additional capital into this space as highly valuable and I think it's often a smaller, easier jump for many donors than some of the more exotic super-high-impact cause areas.

My intuition is that I think that the fund would not be particularly attractive to new donors or have that much potential for growth if we only funded one project at a time and given that there are a number of projects available with similarly high expected value, spreading over a number of orgs (and including some early stage orgs) seems like a valuable thing to do.

I think this sentiment is shared across the team but may also include other reasons.

Comment by agdfoster on EA Meta Fund AMA: 20th Dec 2018 · 2018-12-20T21:01:34.568Z · score: 4 (3 votes) · EA · GW

Final post for the day as it's late. Posting this question myself as I think it's a useful one to have addressed:

Why have you chosen to spread your grants across a number of orgs and not given most of it to just the one or two org/s you thought were the most pressing?

Comment by agdfoster on EA Meta Fund AMA: 20th Dec 2018 · 2018-12-20T20:59:07.235Z · score: 6 (5 votes) · EA · GW

The challenges of funding small early stage orgs deserves a whole piece to itself but I can say that this topic has come up internally quite a few times and we will continue to try to address it.

We did make a few small grants to smaller orgs this time around as we wanted to signal that we supported those projects.

One potential future is that we decide to give a much larger proportion of the grant allocation to smaller orgs. That would require either for the larger orgs to more easily fill their room for funding (one clear signal of this would be those orgs no longer needing capital from flexible donors that might have funded smaller projects in the alternative) or for some strong arguments for increasing our prioritisation of smaller orgs. Right now the benefit of experimenting by funding small orgs is traded-off against the increased uncertainty.

Some of the larger orgs are also only large because they are performing incredibly well. It's worth pointing out that they were previously smaller orgs themselves and that the benefit of experimenting with earlier stage orgs isn't realised unless they can be scaled up once they've proven their approach.

This is also how I currently feel about EA grants. I'd love to see them work out how to scale up their operation. It's a project that is more bottlenecked on talent than funding and early stage re-granting is really hard. An obvious retort would be for them to lower their standards and give out more grants with less analysis (in venture capital this technique is distastefully referred to 'spray and pray'). This isn't particularly feasible at present I don't think. Any re-granting program has its own donors and there aren't any major donors that I have heard of that are willing to have their funds distributed in this manner. It's also not clear that this would be a good thing given the risk of bad orgs with seed funding doing significant damage through PR risks, dilution effects or similar. Maybe the benefits outweigh the risks but either way, this is another effect that puts off donors from giving to a re-granting program that uses this approach.

My best guess, is that EA Grants' (or similar group's) best approach is to take their time to work out how to deploy early stage capital in a systematic and professional manner that can attract funding from larger foundations like Open Phil that usually don't have capacity to handle much early stage funding.

For now, and I don't speak for the whole team here, my best estimates are that very early-stage funding in most cases is lower impact than filing the funding gaps of growing orgs and for an impact-maximising fund such as ourselves we should only be putting smaller checks into particularly promising early stage orgs that we want to signal confidence in.

I'm very open to having my mind changed on this and have enjoyed taking the time to put thought to paper in this AMA today (as is hopefully apparent from the volume of text for any of those still reading).

Also to clarify: while I don't currently think it's highest impact for this fund to be making lots of early stage grants due to the points raised above and others, I do think that a systematic team of full-time staff putting all necessary processes in place, attracting the right talent would be able to raise and deploy lots of early stage funding in a way that is highly valuable. I'm very much hoping that this what EA Grants manage to achieve.

This is also not a closed topic for us and it's not that unlikely that, come next granting round and having discussed this further, the EA Meta Fund team decide to devote more of our allocation to earlier stage groups.

Comment by agdfoster on EA Meta Fund AMA: 20th Dec 2018 · 2018-12-20T19:38:12.148Z · score: 6 (4 votes) · EA · GW
  • I'd be interested to read some thorough thinking on how to trade-off between funding early orgs and later orgs. I'd have a bunch of considerations I would want to make sure made it in there but I don't currently have any piece of work where I've tried to do a general collection of these considerations and weight them. I generally feel like I make better decisions when I can apply my intuitions to a series of smaller decisions that come together in some way.
Comment by agdfoster on EA Meta Fund AMA: 20th Dec 2018 · 2018-12-20T19:31:28.393Z · score: 3 (2 votes) · EA · GW
  • A question I have that I would like to learn more about at some point (I currently assume it's not very relevant for some time to come) is how to think about competitive dynamics in small donor funded markets. In large profit-driven markets it seems well accepted by economists that more competition is generally a good thing for the consumer. In smaller donor-driven markets it seems like a less closed case. If choosing between funding an established player or another org doing the same thing, how should we think about that? Presumably at some point the cost of duplicated effort, talent and capital requirements are overridden by the benefits of competition. My current working assumption is that that point is quite far along, especially given second mover advantage. If this is incorrect then perhaps we should be funding more early stage clones of existing orgs.
Comment by agdfoster on EA Meta Fund AMA: 20th Dec 2018 · 2018-12-20T19:30:55.239Z · score: 3 (2 votes) · EA · GW
  • Better frameworks for thinking about fungibility. See my points from above reply to Tee. The reasoning used for these currently is pretty ad-hoc and potentially it's better that way, but it does seem like it could be important enough to warrant further research. I think I saw a while back that it was on GPI's research agenda.
Comment by agdfoster on EA Meta Fund AMA: 20th Dec 2018 · 2018-12-20T19:30:46.815Z · score: 3 (2 votes) · EA · GW
  • Better frameworks for evaluating community building or spreading important ideas. It's really hard and involves a number of highly sensitive input assumptions. Whilst this is of course very relevant for CEA-type groups and prioritisation research groups, it's also relevant for groups like Founders Pledge. In poverty charity evaluation one question has been whether long-term flow through effects actually dominate the equation. Similarly, it might be the case that for money moving groups that the community building or important-idea-spreading functions might dominate the equation. This would be an important realisation for funders but potentially more so for those orgs themselves. A big issue is that it might be too case dependent to generalise. I don't think a model would be very helpful, but a well thought through list of considerations to evaluate with some estimates on how much to weight each one could potentially be very useful.
Comment by agdfoster on EA Meta Fund AMA: 20th Dec 2018 · 2018-12-20T19:25:58.464Z · score: 6 (4 votes) · EA · GW

I've forwarded this data to the team - thanks for sharing it, I missed it when it went up on the forum.

How to take data like this into account is an interesting and tricky question. I can have a go at a few points that seem relevant:

  • I'd rather see opportunities as projects and funding gaps rather than asking 'who received what funding already?'. Point made only to clarify, I realise this isn't what you were suggesting.
  • EA Meta as a cause area does have a larger requirement for funding than is currently available. The only two donors to meta orgs I know of that are not already donating at their own full capacity are my employer and Open Phil. However, both are deploying capital as fast as they can limited by other restrictions (risk, talent, appetite of principal etc). Certain key groups have performed particularly well and fundraised well so they are making decisions based not on maximising their 'impact per dollar donated' but on maximising their absolute impact given some other bottleneck/s. It does not necessarily follow that an org in the latter category is lower impact per dollar donated than an org that isn't. I think some more simply expressed version of the above would be more useful than discussing whether orgs / cause areas are funding constrained (outside of evaluating counterfactuals when making career decisions).
  • The operations of some orgs are also far more scalable than others and in general I want to reward this. While we mostly speak about relative returns (impact per dollar) we should also keep in mind absolute returns. In particular, it's worth nothing that reaching a certain size and scale of operation opens an org up to large grants from large foundations, accessing capital that otherwise wouldn't have gone into the cause area. An adage used in venture capital is that it "takes founders just as long to raise $100k as it takes to raise $1m". This does seem to hold true for non-profits as well so long as there is enough mid-stage and late-stage capital available.
  • The general approach I take is to only challenge an orgs declared room for funding if it seems surprisingly large, small or poorly justified. Potentially, given that in many cases funds one org receives come at the cost of those same funds going to another org, room for funding and budget declarations should be more heavily scrutinised. i.e., It's OK to stretch a high return business model slightly into it's diminishing returns so long as it remains more effective than smaller marginal groups whose funding they might be restricting. In other words: we should probably be encouraging EA meta orgs to avoid being wasteful with their resources.
  • This said, funding is only zero sum for some donor segments. Some donors are restricted by , for example, Open Phil's rough 50% rule or a donor's limited time and confidence requirements causing them to prioritise larger capital deployments.
  • I have seen at least two examples of larger orgs directly taking into account the flexibility of donor capital that they receive and spending some time trying to replace that donor with someone less flexible with lower opportunity cost. To me this seems highly commendable.
  • How a group have used previous funding plays an important role in evaluating their likelihood of using new funding well.
  • Some orgs are a cluster of valuable projects that could just as easily be evaluated project-by-project and we wouldn't want to be guilty of something akin to gerrymandering.
  • It seems like where orgs are a collection of projects, ideally we would be able to evaluate each of those projects individually, as well as evaluating the group as a whole. It would be helpful if these cluster orgs were better able to track the progress of their projects individually. 80k are particularly good at this.
Comment by agdfoster on EA Meta Fund AMA: 20th Dec 2018 · 2018-12-20T18:28:40.211Z · score: 5 (4 votes) · EA · GW

I also quite like:

Upstream orgs that improve the quantity and quality of available talent, capital and insight.

It's snappy, easy to remember and gets to the heart of the cause area.

I don't think it's as clear what it means though and perhaps shouldn't be used by itself without further explanation.

Comment by agdfoster on EA Meta Fund AMA: 20th Dec 2018 · 2018-12-20T17:50:16.240Z · score: 5 (4 votes) · EA · GW

Three main sources:

1) Via the application form: This is [hopefully] coming soon. The plan is for two of us to monitor the applications and then send proposals to the group same as the process for anything that comes from intro's (see reply to the comment above from @lukeprog). I must note that reviewing applications can be very time consuming and I've personally found it generally requires much less energy to review projects that come in via intros than what comes in via open application forms.

2) Via intros. From my time in early stage businesses I felt this model was adopted by almost all venture capital orgs for good reasons and I have some trust in the method. I'll put a little more on my own view on this in a separate comment.

3) Existing knowledge. Between us we made quite a long list of opportunities we already knew of in our first meeting. Even just the funding gaps on that list would be enough to absorb the whole fund for some time to come, however it does seem prudent to keep looking for even more promising projects.

Comment by agdfoster on EA Meta Fund AMA: 20th Dec 2018 · 2018-12-20T17:26:30.410Z · score: 4 (3 votes) · EA · GW

Two of the team members already spend a significant portion of their day job investigating and evaluating meta groups (Luke and myself). For Luke this is about 50% of his time for his own philanthropy and I spend maybe 10-20% of my time depending on what other projects I have on the go. In general, a core criterion for choosing us for the team was (I believe) that we all spend quite a bit of our non-day job thinking about or evaluating meta orgs. Matt and Denise have been very active, and strategic, donors in the space for many years and Tara whilst only making a switch to earning-to-give last year, thought very hard about this while she was on the other side of the table as COO and CEO of CEA. I also find her views on how to evaluate community efforts are particularly valuable. We all come across a variety of opportunities from our existing activities and expect anything we miss to come our way once we open up public applications (coming soon).

Re time spent discussing: we are trying to minimise the amount of time we spend on group calls, probably only 3-6 a year as they are generally less efficient than a-synchronous communication. We have internal email threads and a private forum that we are using for collaborating on research and swapping arguments / opinions.

Roughly, our process is that someone from the team looks into a funding opportunity, they then put it to the rest of the team in an email thread or in the forum with their research / arguments and the rest of the team then reply with their comments, questions, opinions. Then, ahead of a granting period, we come together to discuss the various proposals and over the next week or two each decide in a Google Sheet how we would want to split the available funds. There is a simple algorithm that averages and rounds, obeys our general rules for grant minimums etc., and we discuss / negotiate each other's decisions and change our allocations until we all agree the outcome looks like a) a sensible outcome and b) represents the group's views.

During the calls we have someone taking minutes and the offline discussions are all in text form, someone takes the initial proposals, minutes and offline discussion texts and turns those into write-ups. We each then give a round or two of feedback on the write-ups before submitting the decision and writeups to CEA for execution.

Exactly how much time this takes up for each of us is currently unclear as we all seem to be thinking about and replying to threads both inside and outside our day jobs.

Comment by agdfoster on EA Meta Fund AMA: 20th Dec 2018 · 2018-12-20T16:58:29.903Z · score: 7 (5 votes) · EA · GW

Thanks for the kind words Tee! Agreed and hopeful. I do also think that it's very valuable for some pots of funding to not be very public as there are some bad incentives and restrictions caused by public work.

E.g., I'm (currently) quite happy currently that EA Grants doesn't have to justify each grant publicly. This allows them to take gut-calls on early stage projects and to fund lots of small things without having to hire a large number of staff.

Whatever level of transparency each grant making body decides is appropriate for their strategy, in general I think more of the benefits of grant making (and research in general) compound when done publicly and transparently. I'm just glad that there are some pots of capital coming together that can make quick decisions and back lots of early stage projects.

This said, I'm of course not all that confident in this view.

Comment by agdfoster on EA Meta Fund AMA: 20th Dec 2018 · 2018-12-20T12:26:49.393Z · score: 3 (2 votes) · EA · GW

We do not have any specific general set of criteria for evaluating opportunities. I'd say we're probably all agreed that we are trying to use the best reasoning, evidence available and relying on intuitions more, or less, depending on the level of confidence that we are able to have in our opinions.

Different team members also have some specific techniques they use for different types of meta orgs. E.g., I have a simple model I use to help me evaluate money-mover (improving capital) type orgs.

Quantitative analysis doesn't currently look worth it for evaluating GPR or talent-focused groups (an exception is 80k's own metrics which are really impressive, but there aren't really any similar groups to compare those metrics with).

Comment by agdfoster on EA Meta Fund AMA: 20th Dec 2018 · 2018-12-20T12:17:27.775Z · score: 5 (4 votes) · EA · GW

I would say that current thinking is primarily to contribute towards our best estimates of the highest value funding gaps, with some adjustment for fungibility and how time expensive it will be for senior management to fill their funding gap.

However, the team are also keen to fund a variety of things each round and we have agreed to aim to have some spread each round between larger orgs and smaller orgs.

I would expect that we continue a trend of larger checks to larger orgs and smaller checks to smaller orgs.

Comment by agdfoster on Long-Term Future Fund AMA · 2018-12-19T18:35:05.358Z · score: 9 (5 votes) · EA · GW

I'd be interested to know, if it can be disclosed publicly, whether non-advisor team members also control alternative pots of discretionary funding.

Comment by agdfoster on 2018 AI Alignment Literature Review and Charity Comparison · 2018-12-19T17:53:26.942Z · score: 2 (2 votes) · EA · GW

>Last year I mentioned that EA Long Term Future Fund did not seem to be actually making grants. After a series of criticism on the EA forum by Henry Stanley and Evan Gaensbauer, CEA has now changed the management of the funds and committed to a regular series of grantmaking. However, I’m skeptical this will solve the underlying problem. Presumably they organically came across plenty of possible grants – if this was truly a ‘lower barrier to giving’ vehicle than OpenPhil they would have just made those grants. It is possible, however, that more managers will help them find more non-controversial ideas to fund.

The last sentence is one of the key reasons it was refreshed. It's also worth noting that I believe the new managers do not have access to large pots of discretionary funding (easier to deploy than EA Funds) that they can use to fund opportunities that they find. I could be wrong about that.

Comment by agdfoster on Announcing EA Funds management AMAs: 20th December · 2018-12-18T14:25:14.324Z · score: 2 (2 votes) · EA · GW

looking forward to this! Thanks for organising, Marek

Comment by agdfoster on Lessons Learned from a Prospective Alternative Meat Startup Team · 2018-12-13T02:13:33.108Z · score: 4 (4 votes) · EA · GW

Thank you so much for writing this up, I'm sure it will be really valuable for teams considering a similar move, whether right for them or not, in the future.