Q&A with Marcus Davis

Marcus Davis is Co-founder and Co-executive Director at Rethink Priorities, where he leads research and strategy. He's also Co-founder of Charity Entrepreneurship and Charity Science Health, where he previously systematically analyzed global poverty interventions, helped manage partnerships, and implemented the technical aspects of the projects.

Below is a transcript of the Q&A, which we’ve lightly edited for clarity. You can also watch it on YouTube and read it on effectivealtruism.org.

The Talk

Bridget Williams (Moderator): Hi, everybody. My name is Bridget, and I'm the emcee for this session. I hope you're all having a great EAGxVirtual so far.

This session is a live Q&A with Marcus Davis. Marcus is a co-founder and co-executive Director at Rethink Priorities, where he leads research and strategy. He's also the co-founder of Charity Entrepreneurship and Charity Science Health, where he previously systematically analyzed global poverty interventions, helped manage partnerships, and implemented the technical aspects of the project.

Marcus has told me that he's very happy to answer a whole range of questions in this session, so this is your chance to throw some questions at the Co-executive Director of Rethink Priorities. [...] We'll try to get through as many of your questions as we can in the next hour.

First of all, Marcus, thanks very much for joining us at EAGxVirtual.

Marcus: Thank you. Thank you to everyone who is participating, and thank you very much for the kind introduction, Bridget. I'm happy to be here.

Bridget: Great. To start us off, could you perhaps [share] an overview of Rethink Priorities — a bit on the history of the organization, what your aims are, and what work you're currently doing toward those aims?

Marcus: Yes. Peter Hurford and I founded Rethink Priorities in 2018. We are a think tank. We broadly wanted to do research to help improve decision-making at EA organizations and generate research to try to make the world a better place.

We cover a number of causes. We're primarily focused on non-human animals, but we also do work on longtermism; we spent a good deal of 2019 researching [the risks posed by] nuclear weapons. That work was led by Luisa Rodriguez.

More broadly, our goal is to use research to investigate areas that haven't been well-studied — things that are potentially very impactful — and to try to guide the decision-making of large donors, aligned organizations, and policymakers. And I should note that while we began very humbly with two people, we now have a staff of 10.

Bridget: That’s impressive. You mentioned that Rethink Priorities covers a few different areas. How do you go about picking those areas and what to focus on? Also, a question has come in about your plans to do longtermism-related work: What factors go into deciding where to put your energies?

Marcus: Using first principles, we think through the areas where we think we can have the highest impact.

When we first started out, Peter and I were really trying to get a sense for whether we could do anything at all that would be useful to the community and others. As we experimented, we decided that [Rethink Priorities] was useful, and that it was worth expanding the project. At that point, we began thinking a lot about the areas in which we could be making a serious impact.

We've mentioned some already. In addition to our work on animals and longtermism, we’ve worked on movement building. We run the EA Survey [which collects data on many aspects of the EA community] every year.

I’ll give the short version of how we came to [work in these areas]: Basically, we looked around the animal research space and saw that there just aren't that many players. It's not true that there aren't any existing organizations doing this work. There's THL [The Humane League] Labs, for example, which does a lot of high-quality work, and many other organizations, including ACE [Animal Charity Evaluators], which is a meta organization doing research.

But broadly speaking — compared to, say, global health or even longtermism, there just aren't that many people working in [the animal welfare] space. We think it's a very high-impact area, given the lack of resources dedicated to it. We thought there was a strong chance that if we were to spend significant resources in this domain, we could get pretty good returns.

As for the other cause areas, it's a similar case. The EA Survey, if I'm not mistaken, was run by Rethink Charity before Rethink Priorities came to exist. Nevertheless, we think it remains a useful tool for the community. We have heard directly from the leaders of other organizations that they find [movement-building] tools useful, and they want us to continue to do this work.

As for longtermism, I probably don't need to relay the basic case for the long-term future mattering. It would be nice if we all didn't die at some point in the near future. As for why we focus on nuclear weapons, rather than AI risk, Peter and I are drawn to the type of quantitative work that fits well in this domain [quantitative work is easier to do in a data-rich area like nuclear risk than in a more speculative area like AI risk]. When significant data exists, we try to build models to better understand the inherent risks. In fact, the first project that Peter and I worked on was related to vaccines. There's a lot of data out there on vaccines, but not a lot of analysis from an EA perspective.

This is all high-level background from first principles [that we apply] to any individual case or project that we take on. Then, given all of this background, we [address] large-scale project prioritization. While we're shooting for targets on different causes, the individual projects we're working on at any given time are the result of what I was going to call a “painstaking” process, but project prioritization is actually pretty fun to me.

Bridget: Yes, I'm sure “painstaking” isn't the right word — I'm sure it is very fun — but I'm also sure it's a lot of work and takes a lot of time.

What's interesting to me about Rethink Priorities is that it's specifically focused on the EA community. As far as I'm aware, it seems to be one of the few organizations with that specific focus. Is that correct? And was that a tactical decision — to retain the EA community as a central audience?

Marcus: We’re not trying to [exclude] the broader community. But it is definitely the case that we deliberately thought about approaching decision-makers within the EA community, be they donors or EA organizations.

A lot of existing EA organizations expend huge amounts of resources or personnel on projects. And given that, there is more work to be done to inform their decisions [about which projects to work on], or which grants they choose to hand out. Take the case of animal welfare. Open Philanthropy has three people working on animals full-time. I think they're fantastic people, but it's just three people. If you're thinking about all possible research topics, even if you're working on overdrive, there are just not that many people relative to the scope of the problem.

Given [that reality], I think it's important to help. But [staff size isn’t the only factor]. Even if these organizations up the number of staff working on these topics, I still think much of what we do is useful. The broad, quantitative approach we take — where we use data as best we can to understand what's currently working or has worked in the past, and what is not working or has not worked in the past — is something that I think is very valuable, and that Peter and I definitely thought we could bring to the EA movement.

Bridget: Great. Focusing on your work in the animal space for a moment, has [your research] caused you to change your mind on anything? Do you think it has changed how you think EAs should prioritize?

Marcus: It's definitely true that a number of projects we've worked on have brought topics to my attention that I was just unaware of, or about which I didn't hold the positions I currently do. To take an extreme example, I didn't know very much about invertebrate sentience before we started to work on it. After many months working on that project alongside Jason Schukraft, Daniela Waldhorn, Max Carpendale (who was a contractor working with us), and Peter Hurford, I came to update my thinking more in the direction of “it's plausible that a wide range of these species are sentient,” simply because we examined the evidence.

More fundamentally, I'd say I came to this work already thinking that animal welfare was pretty important — and nothing I've seen in our research has deterred me from that position. If anything, I've probably come to believe that it’s even more important. The scope [of the problem] is larger than I would have guessed at the time, because I wasn't aware of the massive amount of shrimp farming that's going on. Right now, Daniela Waldhorn is working on investigating the scope, conditions, and possible interventions for shrimp, and she has found that something like 240 to 320 billion shrimp are killed per year. That's much, much more than land animals. I was completely unaware of that number.

So if you think about the [range] of animals we’re concerned about, it’s larger than we thought. And that doesn't even include wild animal welfare.

Bridget: Do you think there's any low-hanging fruit, or interventions that should be explored in those spaces, that we currently aren't in?

Marcus: In wild animal welfare or invertebrate welfare?

Bridget: Either.

Marcus: Wild animal welfare is an area where I feel extreme uncertainty. You can kind of categorize possible interventions [based on whether they] decrease or increase the population. You could take the position that interventions that decrease the population of wild animals [indicate] suffering. Or you could consider the category of interventions that reduce suffering without altering the population very much.

The latter category seems a lot safer to me. I'm not at all confident that any given category of wild animals need to live a life that's negative or positive overall — or that we should be trying to [implement] large-scale interventions that are potentially very hard to embrace. And given that constraint of trying not to alter populations that much, I imagine there are plausible interventions we could be doing that are relatively smaller in scale, but capable of reducing the amount of suffering that some animals are undergoing.

I don't know how these particular interventions will compare to other animal interventions, or to human interventions. But I'm fairly positive that there are cases out there.

All that said, I'm not sure whether we should be [focusing on that] right now, or on trying to understand the situation better to build up confidence, support, and knowledge for larger-scale interventions down the line.

As for invertebrates, again, I'm pretty uncertain. But there is one thing I can say confidently: We should stop boiling things alive. It's such a low bar, but given that there’s a reasonable chance of various creatures being sentient, it seems wildly irresponsible to boil them alive. This happens to a lot of crustaceans — crabs, shrimp, and lobsters — and a number of other creatures as well, like snails.

While I think the probability of sentience in a snail is much lower than it is in a chicken, it's not zero. It's much closer to something like 5%. If that’s the case, consider that you wouldn't take a one in 20 chance of setting a human on fire, when you have 100% confidence that humans are sentient. [That analogy] ignores a lot of complications with inter-species comparisons, but the point remains: I think it's wildly irresponsible. Whether we could get the public to support something like that invites many different questions, but it is clearly true that we should stop boiling animals alive.

Bridget: Yes, it's a very good point.

With Rethink Priorities, where does the scope of your research end? Do you confine yourself to just looking at questions of which interventions, broadly speaking, are worthwhile or should be better prioritized? Or do you also go into the weeds of which organizations or groups might be well-positioned, or which practical pathways [exist] to make change?

Marcus: I'd say we are really interested in the scope of all research that would be useful, up to and including the possibility of very specific charity ideas that we have found to be plausible. We do some investigations of interventions that are currently happening. For example, Saulius Simcikas did an investigation of corporate campaigns currently under way to try to determine how cost-effective they were.

I can imagine a similar scenario where someone facilitates an intervention that they're not currently practicing and we investigate it thoroughly — not just to see how cost-effective it might be, but also how effective it is. If we found it to be very effective, we might decide that we’re very interested in it and want to help by providing connections to academics or people who are knowledgeable about the domain.

So I'd say we're pretty open to a wide range of research. For something more like which organization would be best-placed to do [a particular intervention], I'm less sure what that means. If it means something like “We think the government should do X,” we're definitely in the domain of trying to figure out what policies are popular or possible to get passed. But I'm not sure in the case of whether, for example, “the Centre for Effective Altruism should do Y.”

Bridget: It sounds like you offer a pretty broad range of research. I'll move on to a slightly different question [from the audience]. In which ways is Rethink Priorities different from academic research institutes, and what do you think academic research institutes could learn from you, and you from them?

Marcus: This is an interesting question. I’ll start with the latter question, and perhaps that will answer the former.

I would think that one thing we could benefit from is academic institutes’ years of experience in the domain. Obviously, we've only existed for two and a half years, and we’ve only had a significant number of staff for fewer than two years. Although we work very hard to make sure that everything we publish is well-informed, and we speak with experts, it's also clearly the case that we don't have years and years of expertise to draw upon.

Also, academic institutes tend to have an extensive staff. I don't have a pool of graduate students to call upon to do a lot of work that I would like to see done. I'm not necessarily saying that is always the best [approach], but it jumps out to me.

Fundamentally, I’d say we could learn how to “professionalize” our organization, which is a common [challenge for early-stage organizations like ours]. We’re trying to make the organization sound and put great policies in place for handling various scenarios. But I'd wager that, for the most part, academic institutions are pretty strong in this regard, simply because of their years of experience.

[In terms of how we conduct] research, it’s harder to generalize; it would depend on the domain.

Going in the other direction — i.e. what I think academic institutions could learn from us — I am perhaps more comfortable [than academic researchers are] saying, “I don't know," “We'll look into this for X amount of time,” or “While this is important, we don't have time to look into it further.” We may just share the best knowledge that we can given the time constraints, or research constraints, we’re working within. We’re comfortable saying, “This is basically all we could do here,” and moving on.

In academia, I don't think there's a lot of incentive to do work — not even marginally — where that kind of result [is acceptable, despite the fact that] if you want to have the highest possible impact, you’re not going to be able to cross every “T” and dot every “I” with everything you do. [Chasing down every last bit] of information clearly has diminishing marginal value.

I guess this could be more fundamentally about the way that we publish research. We publish on the internet; basically, everything that we do defaults to being public. I don't know if that's necessarily true of every academic institute.

Also, I would add that one way in which we're probably similar — and that might be unclear to outsiders — is that we definitely try to practice at least some type of informal peer review. We ask other experts to review our work, including people at other EA organizations and academics from many disciplines. Everything we put out has been deeply reviewed before we make it public.

Bridget: Yes. What I was going to ask next [concerns] your choice to publish most of your work on the EA Forum. Is it fair to say that's your primary means of publication? And what do you think about the benefits of that compared to more traditional publication options, like putting reports on your website or submitting articles to academic journals and using a more formal peer review process? Could you talk about what you think the pros and cons of those different options are, and why you've taken the path that you have?

Marcus: Yes. To the extent that we are trying to [appeal to] the EA community, or key decision makers in the EA community, it makes a lot of sense to publish on the EA Forum. If you're trying to get information out to that audience, the Forum is a good place to [house] that information. However much I would like it if everyone in the EA community read our website, we're much more likely to succeed at getting people's attention if we put our research on the EA Forum.

Fundamentally, we don't particularly care about copyrighting our work or making sure that we're getting [noticed more broadly]. We're not a for-profit company that generates hits from our website to make money. The important thing to us is to get the EA community to read it.

All that said, there are still definitely times when we might want some work to be published in academic domains. People have come to me and said, “I very much would like to feature some of your work on our site, but it's not in a peer-reviewed journal, so I can't really do so. It would be helpful if you could do more of that.” There are clear trade-offs between pushing to get things published in academic journals and choosing not to. Costs include the time and effort it takes, but [a benefit is] you get access to a broader audience. Also, some journals won’t publish articles that appear anywhere else — not even if [that other place is an organization’s website rather than another journal] — although I think this might be changing. Of course, you're free to send papers to people, but that’s very different.

I don't think [not publishing in academic journals is] a huge drawback, but we could probably do more. Publishing on Google Scholar, such that if you searched for a related topic our work would come up, could be low-hanging fruit.

A higher bar would be something like trying to influence a certain policy-making community. We might think that in addition to the EA community seeing a conclusion we’ve reached, it would be important to get it into an academic journal [to successfully sway policymakers]. That's something we're actively exploring, depending on the project. And we are, in fact, in touch with a few academics who would like us to push more in that direction.

Bridget: You made some good points about trade-offs. I think there are a lot of problems with getting things published in academia, and it's nice to be able to avoid some of those. You have more flexibility working outside of it.

A related question is about how receptive other groups are to your research, and how you make sure your research is being seen and [applied] by the people who might be able to make the best use of it.

Marcus: I would say this is a high-variance category. There are definitely times when we work on something that we think is very important, and then I try to present it to someone who might not (at least not at first glance) understand why.

One thing that we do to avoid that, and to drastically improve [our chances of getting heard], is talk to people on a regular basis. That way, we get feedback on what they think is important and can incorporate [what we hear] into the project prioritization I talked about earlier. We think through the key parties and what they think is important to work on — for example, the key challenges they face in their network. Then, when we present our work to them, we might say, “Here's what we found looking into the issue that you thought was important.” They're much more likely to listen to us.

Another general truth that doesn't apply just to EA: The more you know the people you're trying to convince, the more [effective you can be when] you talk to them. When you network and get to know what people like or dislike, the barrier for convincing them to look at [your work] goes down.

As far as making sure our work is used, we [start by] making sure that it’s read by important parties. In addition to publishing it on the EA Forum, we'll directly reach out to those parties and say, “Hey, have you read this? Here are the main conclusions. We'd be happy to chat.” We give them an opportunity to directly speak with us about the topic.

Bridget: Are people generally receptive to when you reach out directly?

Marcus: Mostly. We don't have quantitative data on this, but maybe I will start to track it.

Bridget: Yes, I understand you did an impact evaluation — was it last year? What were the main findings of that? Was there anything in particular that was surprising to you?

Marcus: Yes, we ran an impact survey last year [in 2019]. We were trying to figure out whether people read our work, thought it was important, and founded charities or decided to stop or start supporting an intervention based on it. We polled about 100 people who we thought were particularly relevant.

The major takeaways were that a lot of people read a significant amount of our work, but there weren't many clear career changes, [or indications that people stopped or started charities or interventions], based on the work that we put out. Some of that may be due to the people we targeted, who were more likely to be in positions like grant-making, or higher up in organizations and less likely to change their careers.

So, I'd say there weren’t any clear, significant [conclusions along the lines of] “We caused X big change to happen.” But we do believe that if we're going to exist as a research organization, there’s probably going to be a lot of variance in the impact of our research. Some of it won’t lead to very big changes, but every now and again we hear of a really massive impact.

One way to think about our work’s [impact is to address] deep EA considerations like the distribution of resources. Other ways may be more along the margins — e.g., “Work is already happening [in this space], so should we shift resources slightly?” Or, we might say, “Here's a completely new intervention idea that no one has researched or explored at all; how does it compare to existing domains?”

While I think we did a lot of work last year that was worthwhile, we didn't produce anything along the lines of “Here's a charity that no one has looked into yet, and that could be funded for $2,000,000 a year.” A big goal of ours is to produce that type of work. But not every project falls into that category. Some projects are more within the scope of “It's not that plausible.” Our work on shrimp, which I mentioned earlier, falls into that domain.

[Team member] Jason Schukraft’s work on moral weight, on the other hand, is more in the realm of a big EA consideration, because it might be that the distribution of resources between humans and non-human animals should be significantly shifted. Or, it could be that [the distribution] is roughly okay. Before looking into that topic, it's very difficult to know.

Bridget: Yes, it seems like the Rethink Priorities team covers a really broad scope of questions, from fundamental research to implementation. That probably makes it hard to evaluate impact.

There are a few questions around the role of Rethink Priorities in the EA landscape of organizations. I think we’re in a pretty cool place at the moment, in that there are quite a few different organizations doing research. So, where do you think Rethink Priorities fits in? How should people think of you in comparison to an organization like GiveWell, the Sentience Institute, the Forethought Foundation, or the Global Priorities Institute? Also, do you coordinate with different groups in the EA space to ensure that you're not working on the same questions, or to work out who may be best-placed to work on certain questions?

Marcus: The second question is much easier than the first; the answer is yes. We definitely talk to other organizations. We figure out what they're working on before deciding what we want to do. Sometimes that means deferring to other organizations, sometimes it means collaborating, and sometimes it means we think it's important for us to take different approaches — to both work through the issue and get feedback on each other's work as it proceeds. That applies to individual projects more than broad causes.

As for how Rethink Priorities compares to other research organizations, I can say that we’re not in the global poverty charity evaluation space. We defer to GiveWell on that and trust that they produce really high-quality information.

To a lesser degree, some organizations, like GPI [Global Priorities Institute], are doing fairly foundational research. They help address questions that fall into the category of “big EA considerations that could affect how you [channel] your time and resources.” Some of our work falls into that category as well, but a lot of it does not; the specific intervention research we're doing certainly doesn’t. And while there are questions we both try to answer, like “What's the capacity of welfare for a different species?”, for the most part our agendas don't overlap. We probably should communicate more with them, but for the most part, their work just isn’t that applicable to ours.

There are definitely other organizations whose work is closer to ours. The Humane League Labs does a lot of high-quality work in animal welfare, writing relevant studies or doing literature reviews and other quantitative work in our domain. We interact with them a lot. I spoke with Jacob Peacock, who runs THL Labs, on Thursday. I know what they're up to, and we keep in good touch. We frequently collaborate with them and others doing work that we both think is important.

Bridget: Cool. Audience members have a few questions about the organization itself. One of them is “What do you think is the biggest bottleneck for Rethink Priorities’ growth?” Also, if you were able to overcome that bottleneck, what would you be doing? Where would you put your additional capacity?

Marcus: Structurally, we could use additional staff — and that comes down to having the funds to hire additional people. Right now, we have a backlog of surveys we want to run in collaboration with various aligned organizations. There are several of those on animal welfare. There are studies we ran last year that we want to write up. There's work in the longtermism space. In order for that to happen, though, we need someone to run the data analysis and someone to write these things up. We're just limited by the number of hours that our existing staff has.

Also, every time we prioritize projects or talk to other organizations, we [create a new] list of projects we and others think are high-value. We would like to investigate those as well, but we can't because there's not enough time. These projects include interventions that other people are currently working on, where they're trying to maximize the impact and we could contribute some research to help them develop their resources more effectively. Some of these are more fundamental projects.

I mentioned our moral weight project, which is in its initial scoping phase. A proper job may be very extensive — it may take many people years. And of course, it's not always the case that if you could do something involving two or three years of work, it's worth doing. But if you look across the number of projects that we and others think are worthwhile, we definitely don't have enough staff members to execute them all simultaneously.

Bridget: So, is it a question of funding, managerial capacity, or something else?

Marcus: Funding.

Bridget: Okay. Here’s another question about Rethink Priorities. It feels almost a bit rude to ask about money, but I think it's fine for us to do that here. Where does most of your funding come from?

Marcus: Most of our funding comes from large aligned foundations. A large chunk of our money comes from grants from Open Philanthropy and the EA Funds. All of those grants are public; you can look them up. Combined, they make up a large fraction of our support.

After that, a good chunk comes from individual donors. They give significantly less than those very large foundations, but large amounts of money. And perhaps after that, there's a smaller basket of more normal-sized donations. Those are still an important part of our budget. That's basically the breakdown.

Bridget: Okay. And given that you're specifically focusing on informing the EA community, it makes sense that most of your funding comes from within the EA space.

Marcus: I should follow-up and say that I very much would like our funding to be more diversified. One of the things I am personally working on is making sure that's the case.

Bridget: Why is that?

Marcus: If you're running an organization, it's very valuable to have a wider range of donors to work with. As much as I love the work that Open Philanthropy and the EA Funds have done, it's just the case that organizations [benefit from being less dependent on] a few donors. I would guess that the EA Funds and Open Philanthropy would agree; they wouldn't want grantees to be completely reliant upon them, either. In fact, Open Philanthropy's stated policy is not to support more than 50% of an organization, which means that it’s important for an inwardly-facing EA organization like ours to draw money from other organizations.

Bridget: Yes. Do you think you need to change the way you're doing things in order to attract different sorts of funding and achieve that diversification?

Marcus: I don’t think we need to change much, unless you’re [referring to] the way we fundraise. I think there are more parties from whom we could draw support. It comes down to time constraints.

However, we just hired a director of operations, Abraham Rowe, who previously was the Executive Director of Wild Animal Initiative. He’s helping to ensure that I spend significantly less time on business operations. That’s like gold, as far as being able to focus more on fundraising strategy. One of our goals is to create more detailed plans to pitch to funders.

Bridget: Sorry — I wasn't very specific there. I was trying to get at whether you think you need to change your research priorities in order to attract more funding. Do you think you can do that without having to alter your vision?

Marcus: Naturally, I think everyone who's running a research organization would say they can. But I think it's true; we can both do very high-value work and get funding for it. I don't want to [take on] projects that I think are significantly less valuable just because someone will support them.

Obviously, there's a tradeoff. If I had to choose between the two — getting support or doing what’s important — then clearly, the number-one thing [would be doing what’s important.]

Bridget: Yes, okay.

A more specific question has come in: How do you work on comparing the suffering of humans and non-human animals? Is that something that you've been looking at, and what are your thoughts on it?

Marcus: That's a really easy question! [He laughs — that was sarcasm.]

The first thing I would say is if you're thinking about trading off between humans and non-humans, there are a number of dimensions in which they differ, and where they're the same.

We have an obvious advantage in that we can make long-term changes to the world that improve things for other animals. Basically, all other animals lack this capacity, so whatever your philosophical opinion about their status may be, it is clearly true that humans have an additional, huge advantage.

That difference alone is very difficult to understand. How many humans’ worth of value would it take, say, to decrease the suffering of 100 chickens? People give answers to those questions, and most organizations give implicit answers to them. But I think fundamentally, it's just extremely difficult to know what the correct answer is.

All that said, I think the kind of work [on moral weight] that Jason [Schukraft] is doing involves trying to get a better grasp of the underlying philosophical dilemma.

Resolving that dilemma also involves answering a huge number of scientific questions. I'm fairly “consequentialist,” so I think about what the variance in the welfare is across these different kinds of species; I'm thinking about how much someone can suffer; I'm thinking about how fast their temporal perception is.

You may think of us as more or less valuable than other animals. And this is true for a number of dimensions, such as the intensity of your experiences. Some animals experience a level of pain or pleasure more intensely, on average, than other animals. That might indicate something like phenomenal unity. This is more speculative, but perhaps there are creatures, up to and including humans, who have more than one state of experience happening at the same time.

Split-brain patients are perhaps the best-known case [suggesting that] something like this may be going on, where one hemisphere of the brain may be retrieving [perceptions] from the world separately from the other hemisphere. Of course, I don't know the proper way to determine if something like that is true. But it might influence how we think about this domain. Octopuses are a good example of an animal with a decentralized nervous system, where something like that might be happening.

These kinds of hard scientific questions are considerations that determine how I think about different animals’ capacity for welfare. But again, [there is a separate] sub-problem of the larger consideration about how humans could improve the world. [Another question is whether we should place more value on animals that can do] really shrewd things. And you have to figure out what you think about all of these tradeoffs. I don't have a clear answer to how I think about them overall, given all of these uncertainties.

Bridget: Yes, it's certainly a very big question. I think you've done a good job of outlining how complicated it is, and what the uncertainties are.

We are getting toward the end of the session, and I thought it would be interesting for the audience to hear about your experience in starting up an organization, and your thoughts on whether more people in effective altruism should consider doing that. If so, what advice would you give them?

Marcus: I'll share my own story. When I started to intern at the Charity Science Foundation, in the summer of 2015, I wasn't confident that this would be the direction of my career. I thought it was a great opportunity, and that their organization was impactful in that I could potentially help others.

In taking that internship, I was eventually offered a job to help start Charity Entrepreneurship. Again I thought, “This is a great opportunity. I'm not sure it’s going to be my ultimate career path, but it sounds potentially high-value.”

[That led to] starting Charity Science Health. It was there that I started to think more fundamentally about what kinds of organizations are valuable for EAs to start and how we should go about that. While I was there, I started to get more into research. Along with others at Charity Science at the time, I saw that there was a big list of potentially very important questions. A lot of them fall into the domain of some of the big EA considerations I mentioned earlier that are important to resolve if you’re trying to determine which cause to focus on, or how to try to make progress within those causes. [At that point], it became more appealing to start an organization that would try to answer some of those hard questions. [That’s why] I left Charity Science and started Rethink Priorities.

Whether other people should do that is such a hard question. When I decided to start Rethink Priorities, I had a big spreadsheet of plausible career paths, how much money I thought I'd make doing various jobs, how much impact I thought I'd have, my counterfactuals — i.e. what these other organization would do if they didn't hire me or roughly where I'd spend my money if I was going to earn to give.

[When I analyzed] my spreadsheet, it turned out to be wise for me to start an organization, given my particular background and skillset — or, in some cases, lack of skills. I don't have a background in, say, biology or biochemistry, whereas if someone else is in a similar position and thinking about starting something up, maybe the stronger career path option would be to go into one of those areas.

I think it has worked out. I hope others would agree that Rethink Priorities is worth doing. I think it's worthwhile. But as for whether any individual should be more focused on starting an organization, that’s a very difficult question to answer in the abstract.

There definitely are particular domains where it may make sense conceptually — where there’s a strong case for doing research or a particular intervention. But I don't have especially strong opinions, in the abstract, on whether EAs should start organizations.

Bridget: If you could [go back in time and] give yourself advice during your internship at Charity Science, what would you say? Is there anything you'd do differently from that point on?

Marcus: One advantage of being at Charity Science was that it’s a pretty small organization. I got the chance to do a lot of things. I tried cause-effect analysis. Looking back on that, I could have done better.

I probably would spend more time just learning how to manipulate R, which comes up a lot in data analysis. I spent a lot of time back then learning Python, but R is where the action is if you’re doing statistics.

[In terms of] more generally applicable advice, I’d probably say you could try to gain skills that are cross-applicable across multiple domains. I was kind of joking about using R, but a general skill like doing data analysis is really important if you want to do research projects of any kind.

Bridget: I feel like that's quite a useful piece of advice — that people should learn R instead of Python.

Marcus: Don't take that too seriously. It depends on your field. If you're working on “science science,” maybe it's more common to use Python. It varies widely. Python is actually one of the most popular programs in use. It's also great compared to R (that’s just a personal opinion after using the two languages).

Bridget: I feel like we've asked you quite a few difficult questions in this session. One of the top-voted questions, which I'll note was sent in by Peter Hurford, is “Why is Marcus so fabulously handsome?”

Marcus: Ah, hmm.

Bridget: We expect the same level of rigor that you gave the question comparing the suffering of humans and non-human animals.

Marcus: [Laughs.] I'm going to build a model and run my DNA through quantitative analysis to figure it out. I'll get back to you after six months and receiving $500,000 for this project.

Bridget: Well, I'm going to keep an eye out on the EA Forum for the results of that analysis.

We are coming toward the end of the session. To finish up, it would be good to hear your vision for Rethink Priorities. If we were to have another chat this time next year, what would you like to be able to say had happened in the intervening time? What are your goals for the next year, and also more broadly?

Marcus: I'd break that question up into where I’d like to see the organization and what I’d like to see the organization produce. One thing I've mentioned is we're trying to professionalize our operations and do better at setting ourselves up for the long term. There are a lot of things that exist in the boring bucket of running an organization that I'd like to see happen.

Then, there are more exclusively research-based questions I would like to see us investigate. I mentioned before that I’d like for us to [identify] something that no one is doing, and that is worth starting a charity [to focus on]. Uncovering an opportunity like that would, obviously, be high-value. One specific domain we’re interested in is animal welfare options in the EU. We're thinking about what might be high-value there. If we were to identify a particularly [promising] opportunity, we might make the case for additional grantmaking in that domain. Finding something worthwhile there is pretty pressing.

More structurally, I would like for our work to be more useful to the people we're interacting with, be it donors or organizations. I’d like for them to see clear value in the work we produce — not only to better serve them, but also potentially to help someone change their options. It’s hard to say what that might look like. So much depends on what the results of some of these projects are. But if we discover something worth changing, and are able to actually move people in that direction, [that would be exciting].

I’d also like to see us grow. As I mentioned, we're staff-constrained. I would like for us to no longer feel that way and have enough researchers on staff to simultaneously handle the projects that I think are all high-value.

Finally, I’m thinking about how the pandemic is still going on. There's a reason we're doing this session virtually, and I want the best for everyone who works for us and in the EA community. I’d like to make sure that we're all staying healthy through all of this.

Bridget: Yes, I think that’s a good note to finish on. Be mindful that you need to look after yourself. And it sounds like there's a lot of interesting work that's happening, and will be happening over the next five months or so, at Rethink Priorities. So keep an eye out for that.

We’re at the end of this session. Marcus, thanks very much for joining us and providing your insights here at EAGxVirtual. And thank you to everybody at home who is watching. Goodbye.