Transcript

Hey listeners, this is the 80,000 Hours podcast, the show about the world’s most pressing problems and how you can use your career to solve them. I’m Rob Wiblin, Director of Research at 80,000 Hours.

It’s good to be back from our break. Working on 80,000 Hours’ annual review was fun and all but personally I prefer making podcasts.

Today’s conversation is about the ongoing efforts to get a research institute set up at the University of Oxford focussed on figuring out what are the most important priorities facing humanity. The hope is for this work to help guide money and talent to where they can have the biggest impact improving the world.

Sometimes we’re going to do specialist content that could go too deep in the weeds to be of interest to everyone.

So now that we’ve got quite a large number of regular subscribers I’m going to start giving advice at the start of some episodes about who will most find them worthwhile to listen to.

This episode is particularly valuable if you’re interested in learning about universities, academia, global priorities research, how to get new projects off the ground, or careers in research and research management more broadly. It’s also worthwhile if you’re a donor looking for projects to fund.

If none of those things make your eyes light up feel free to skip this one.

If you’re just keen to hear about the ideas bouncing around in the global priorities research community I’ve got conversations coming up with two researchers at the Global Priorities Institute – Will MacAskill and Hilary Greaves.

The other thing I should say is that the Global Priorities Institute we’re discussing is hiring right now for a Research Fellow and Senior Research Fellow in Philosophy. If you have a PhD in philosophy, and I know some listeners do, you should definitely check out the job advertisements which I’ll link to in the show notes.

Both posts are 4-year fixed-term research roles at the University of Oxford and will help to put effective altruism on the map in academia. Applications close strictly on the 4th of January 2018, so take a look soon.

Michelle is also going on maternity leave from March to September and so needs someone to cover for her for a while. So if you’d like to be GPI’s director of Operations you can apply right now. I also hear there will be two administrator roles opening up at GPI and the Future of Humanity Institute next door.

Without any further ado I bring you Michelle Hutchinson.

—

Robert Wiblin: Today I’m speaking with Michelle Hutchinson. Michelle is helping to set up the Global Priorities Institute at Oxford University. Before taking on that role, she was the Executive Director of Giving What We Can, and she has a PhD in Applied Ethics from Oxford on prioritization and global health. She’s also an old and dear friend of mine. So thanks for coming on the podcast, Michelle.

Michelle Hutchinson: Pleasure. Lovely to be here.

Robert Wiblin: Later on in the conversation, we’ll get to talking about how people can have a career in global priorities research themselves, but first off, tell me broadly what problem is the Global Priorities Institute, or GPI, set up to try to solve.

Michelle Hutchinson: GPI’s central research question is how with a given unit of resources we can really do the most good. We’re aiming to be at least the beginnings of the academic wing of the effective altruism movement. You might think that effective altruism and academia were already relatively well integrated, given that a lot of the founders of effective altruism were academics, but in fact, they’ve remained relatively separate, with even the academics who were also working on effective altruism working on it at different points and doing it with their non-academic time. In some ways, you might think that that actually isn’t such a problem, because a lot of excellent research in effective altruism typically needs to be done actually really quickly, because we need to be working out where we can give right now, rather than taking the long timelines that you typically do in academia.

But on the other hand, many of the world’s brightest people go into academia, and we’re really missing out on excellent research that’s being done and also on potentially talented people who might want to work outside of academia on other kinds of research, by ignoring academia. I think there are also some risks associated with not engaging well with academia as a movement. Some movements in the past seem to have had this kind of problem. If you look at cryonics and molecular nanotechnology as movements, although they came about in the ’70s and seemed to get quite a bit of press attention, they didn’t end up growing that well. And there are a lot of different reasons there might be for that, but one of them seems to be that the academic sectors similar to them, chemistry and cryobiology, tended to try to shut them out.

It seemed like there are a couple of potential reasons for that. One was actually that cryonicists and molecular technology workers were relatively, not hostile, but dismissive of the academics working in similar kinds of fields, which understandably put their backs up. In other words, they got quite a lot of media attention before a lot of the scientific work had been done, which made the academics feel that this superficial type of work was getting more attention, and therefore potentially more resources, than their real work. I think effective altruism could easily find itself in a kind of similar position, where it’s done somewhat shallow reasoning because it needed these answers really quickly and then has got a lot of attention and resources, and that could really irritate people who have spent decades working on similar kinds of problems, particularly if we’re a little bit dismissive of their work. So it’s really important to actually engage well with these things and really understand what research has been done so far and how much work went into producing the kinds of frameworks that we use all the time, like disability-adjusted life years.

Robert Wiblin: I think some people I’ve heard have been kind of dismissive of what academia has to say about these issues, because they think not much research has actually been done on this topic, and I guess sometimes I glibly say academics haven’t actually looked that much into which careers do the most good. I guess, as far as I know, there isn’t a huge amount of research out there, except perhaps some studies about teachers and police and those kind of careers and what impact they have. Do you think that there’s actually more stuff out there than we think, and we just haven’t been searching enough?

Michelle Hutchinson: I think it is pretty difficult to know precisely what’s out there, and the chances are we are missing quite a bit of stuff. I do also think that for the kinds of things that we’re trying to do, there isn’t as much research as you would expect, partly because of the types of incentives that academics typically have, because it’s so important to become a real specialist in one particular thing as an academic. There’s all this pressure to go into a really tiny niche rather than taking a big picture view, and the types of questions that effective altruists are wanting to answer actually require having knowledge of quite a few different areas. Simply in order to know what kind of impact a teacher’s having, you need to really understand economics and a whole bunch of different areas, and so it would be the kind of thing that it would be difficult to write, say, a stellar PhD thesis on. So I think there is actually some truth to this, but we should be mindful of why that’s the case, rather than assuming that academics don’t care or aren’t skilled enough to do this kind of work.

Robert Wiblin: Okay, so we want to reach out a bit into academia and at least make sure that we don’t make the professors mad at us. What’s the vision for GPI? How large do you think it might become? It’s based at Oxford, right?

Michelle Hutchinson: Yes. It’s part of Oxford University. Right now we’re working closely with the Future of Humanity Institute and hoping to spring off on our own in January. I think it’s difficult to know precisely how fast we’ll grow. Our aim is to make sure we’re growing sustainably as well as quickly, so we’re hoping to take on something like two or three new researchers per year over the next few years. I would guess that it makes sense, rather than growing really large as one institute, to then start working with other universities and try to grow other centers around the world.

Robert Wiblin: There’s J-PAL at MIT, which coordinates randomized controlled trials on development projects, and there’s only so many of them at MIT, but they have an enormous network of researchers across universities all over the world. Is that the kind of thing you’re thinking about, or is it more that you’d have a couple of different hubs with a dozen researchers at each one?

Michelle Hutchinson: I think it’s pretty difficult to know. I think that’s definitely one model that could work well, or the separate hubs model could also work well. I think we don’t yet have a full understanding of which things are going to work best. We wouldn’t be in quite the same position as J-PAL is, because it’s doing a whole bunch of empirical work, whereas we’re aiming to be more conceptual, and I would guess that growing centers in other universities is potentially going to be more useful if the thing you’re trying to attract is professors who are working on conceptual things and who are doing a lot of teaching and stuff. But as I say, this is early days.

Robert Wiblin: Yeah, it is early days. How many people do you have at the moment?

Michelle Hutchinson: Right now, Hilary Greaves is the Director, and William MacAskill is one of our researchers. We’re taking on another researcher, Andreas Mogensen in January, and other than that we have research associates who are not yet part of GPI but come to our seminars and provide advice on our research agenda and things.

Robert Wiblin: So they’re all philosophers, right?

Michelle Hutchinson: No, we have a couple of economists as well who are research associates who come to our workshops and things.

Robert Wiblin: Yeah. As you said, it’s notoriously difficult to have interdisciplinary research within academia. I guess the Institute would cut across philosophy and economics, and I guess possibly politics and some other fields as well. Are there any others?

Michelle Hutchinson: I think in the long run we certainly hope to have quite a lot of different fields involved. Psychology would be another obvious one. At the moment, though, we’re really trying to double down on philosophy and economics, because of this difficulty of being interdisciplinary, and I think there’s a big risk for GPI and the effective altruism movement as a whole to be seen as simply a philosophy thing, which is trying to do other disciplinary things but is not succeeding at it.

And certainly we’ve already found that it’s quite difficult if you’re based in the Philosophy Faculty as we are, because at Oxford you have to be based in a faculty, to hire economists, because they want to be based in the Economics Department. And so we’re trying to really double down on being a dual disciplinary thing in order that we get seen as properly interdisciplinary, rather than trying to do quite a lot of different things at the same time and ending up accidentally just being seen as philosophy plus.

Robert Wiblin: We’ve been a bit abstract about what the research questions are. Do you have any specific questions that people are looking at at the moment to give people a flavor of the kinds of topics you’ll be publishing on?

Michelle Hutchinson: As I had said, our overall aim is to work out how we can do the most good with a particular unit of resources. One of the most difficult questions there, and therefore one that’s been the least worked on and we have the least idea of, is how to compare amongst really different types of causes. Often people make comparisons like, say I save one life from malaria right now, how does that compare to averting a pandemic in hundreds of years’ time, given how many lives that saves. But that’s really an apples to oranges comparison, because saving a life from malaria now is going to have a lot of effect into the future. And those are the kind of effects that we might be able to have some handle on what they might be, although it will require pretty careful consideration.

So if we save a life now, is that going to lead to there being one extra person throughout history, or will the parent of the child who died actually end up having another child because that one died? Or on the other side, would that person actually go on to have more children? And so is the population of the world going to increase by that proportion throughout history? Those are the kinds of comparisons we need to be making, and those are pretty difficult questions to answer.

Similarly, in this kind of long-termism vein, there are a lot of questions like how will technological advancement improve or make the future worse, and so should we be trying to speed that up, and if so, what would that mean? Then there are also questions where the effective altruism movement seems to have one answer, but there’s quite an academic literature, perhaps pointing in a different direction, and it would be useful to know why that is. One example of this is the case of which kinds of consequences we should take into account. It’s pretty standard in effective altruism to assume that when I’m working out what action to take, I should consider all of the consequences of my action, not just the direct ones, but that’s not the case in particularly the philosophical literature.

In the medical ethics literature, it’s basically assumed that when a doctor’s considering how to treat that patient, it’s imperative that they only take into account the effect on that actual patient. They shouldn’t take into account things like is this patient a corporate tax lawyer or are they a social worker, so should I assume that their value to society is really kind of negative or very positive? This would be grossly unfair. And so it seems pretty important to look more into whether there are good underpinnings for this idea that we shouldn’t take indirect effects into account, and if there are, maybe we should be thinking differently about how we weigh consequences.

Robert Wiblin: It sounds like you’re in the stream of thinking about the long-term future and also population ethics and the philosophy of that kind of thing. Is that right?

Michelle Hutchinson: Yeah, I think that’s definitely right, and also thinking about how we can get useful results from taking fairly theoretical literatures and applying them to more empirical problems. So in the case of population ethics, you might want to marry that with moral uncertainty and work out, given that we might think that we should be morally uncertain, how should we account for the fact that are a number of different population ethics that seem plausible? That’s the kind of thing that philosophers often think is kind of boring, because they worked out the interesting details of both moral uncertainty and population ethics, and simply applying one to another and thinking through particular situations isn’t going to be very interesting to them. But that’s the thing that actually gives you your results of how you might act.

Robert Wiblin: I know that academic publishing moves in kind of a glacial era kind of timescale, but have you guys managed to publish anything yet?

Michelle Hutchinson: As I said, we’re in pretty early stages so far. GPI doesn’t yet exist as a specific entity, so it hasn’t published. There is work that’s been done by Hilary Greaves and Will MacAskill that’s very much relevant to this, and I think the question of whether we specifically have published is one of those uncertain questions.

Robert Wiblin: It may not have an answer, in fact. All of those research topics on population ethics and the long-term future can all get very theoretical. Are there any kind of more concrete, mundane, even empirical questions that you might be interested in tackling in future?

Michelle Hutchinson: I don’t know about mundane, but certainly more economics-y. One question is how we might try to build some kind of framework that can compare more broadly than the quality-adjusted life year metric. Ideally, we’d have a metric which took into account things like education and economic advancement, and some work has been done by this by people like Toby Ord and Owen Cotton-Barratt and Cookson, but so far there isn’t anything like an acceptable metric in this space.

Another thing that we’re thinking of working on is how we might generalize from using simply randomized control trials. It is well-known within effective altruism that the gold standard in economics is using RCTs to work out which interventions are the most effective, but more recently they’ve come under some fire within economics, in part due to the question of how you can actually generalize from an RCT done in one particular context to what you might expect to happen in other contexts. This is the problem of external validity, and we’re actually planning to host a workshop on this, on causal inference and extrapolation, in March, in order to try and start a research stream on this.

Robert Wiblin: I know that one question that Hilary Greaves, who’s the Director of the Institute, has been looking into recently is this critique of effective altruism that, because we only think about how individuals ought to behave, we’re necessarily blinkered to larger changes that you could make in society, and as a result we’re going to miss some of the most effective things that people could do. What’s she been writing about that?

Michelle Hutchinson: Right. There’s this very plausible view that there are certain things that can only be done by collectives, and so as an individual you can make no difference. An example might be a large protest that as an individual going to a particular protest you’re making really no difference, but the whole group is making a big difference. And so the thinking might be that rather than treating the individual as the agent you’re thinking about, you should really be treating the collective as the relevant agent.

The example she actually looks at is one of vegetarianism, to see whether this holds. If you’re thinking of killing an animal and eating it yourself, it’s very clear what the harm you’re doing there is, but if you’re buying meat in a supermarket, it’s much less clear. It’s certainly not going to be the case that for every chicken you buy, another chicken is slaughtered. A much more plausible model of what’s going on there is that for every 50 chickens that are sold, they order in another 50 chickens. In that kind of a case, it might seem that as an individual it’s very likely that your not buying a chicken is going to make no difference, because you’re probably not the 50th person to buy a chicken. So you might think of this as one of these collectivist cases where we should not be thinking about what the individual should be doing, but only about what the group should be doing.

She then talks about the fact that actually the way we ought to be working out our actions in these cases is by using expected value. If I’m considering whether to buy a chicken, I actually don’t know whether I’m going to be buying the 50th or the 49th. So the way I actually ought to make my decision is using expected value. If there’s 49 chances that I’m not making a difference, but there’s a one in 50 chance that actually I am making the difference here, and so in that way, even simply looking at the individual, I can see how I’m making a difference. The idea is that actually that’s how we should see our participation in a protest or something as well. It’s very likely that we’re making no difference, but there is some chance that we’re making a really big difference.

One thing to note about this argument is that it doesn’t in any way point towards the idea that we should be purely taking unilateralist action. We absolutely should be coordinating with what other people are doing around us and thinking about what we should do on the margin. But it also shows that when we are taking into account what other people do, the relevant question is, as an individual, what should I do.

Robert Wiblin: We talked a little bit about this issue on the Toby Ord episode, and I think that the main thing is just to notice that almost whenever it’s the case that the world would be better if a whole group of people were coordinated to act in a particular way, it’s almost certainly going to be the case that for some individuals, it would be the most effective thing for them to do, to participate in that collective action. So I’m not sure, I think maybe you could contrive a weird case where they have a lot of difficulty coordinating, and so it does come apart whether you consider at the higher level or at the individual level, but in most typical cases, it seems like actually the answers are going to have to be more or less the same.

Michelle Hutchinson: Right. There might also be psychological factors at play, so it could be very difficult to make yourself act on some very uncertain chance of making a difference, or it might just be kind of difficult to understand what’s going on with these small chancy things, and so it might be easier in general when you’re working out what to do to think in terms of what should the collective do: okay, I’m one of the collective so that’s what I should do. But we should recognize that what’s going on there is the psychological thing rather than the underlying thing.

Robert Wiblin: You mentioned one reason to do work within academia is that it might easier to hire really smart people, and I know that some organizations like CEA and Open Philanthropy sometimes feel that they’re somewhat limited by their ability to attract really top research talent. Have you found that it’s easier to attract good researchers so far, because they can also maintain their academic career options open?

Michelle Hutchinson: I think “easy” would be stretching it, but we definitely have found that there are very talented researchers out there who are only willing to work for us because it doesn’t shut down their options of going into other academic places later. Academia tends to be very snobbish in this kind of way, where it would be very difficult for someone to get a top professorship in future if they’ve basically ever left academia. And so it’s extremely important that we’re part of Oxford University for doing this. As I said, in fact it can even be a problem that you’re attached to the wrong department within a university, even if you’re in not only a university but an excellent university.

Robert Wiblin: Interesting. So academics are very picky, and I guess they have to be, because it’s such a competitive career path that going to the wrong place can sabotage you in the long term?

Michelle Hutchinson: Exactly, because there are so many people wanting to go into academia and the standard is so high, it ends up being the case that even a small black mark can really count against you for years to come.

Robert Wiblin: On that note, do they find the topics that effective altruism is interested in looking into a little bit strange and non-traditional, and is that a potential career risk for people who might work there?

Michelle Hutchinson: Yes, I think it definitely is. A lot of the kinds of topics that we would want to work on could potentially be rather difficult to publish on in really top journals, which is very important for the calibre of candidates that we’re trying to attract. So for that reason, we’re aiming to make our jobs the kind where the expectation is that you spend 50% of your time on our central research agenda, which would be collaboratively created so that people weren’t having to work on things they didn’t think were useful, but topics that everyone agreed were important.

And then 50% of their time could be spent on other research, which could potentially be entirely irrelevant, where we expect that one key aim people will have is in their other 50% of time to work on topics that are particularly well thought-of in academia that could definitely be publishable in top journals, in order to make sure that their long-term career capital isn’t penalized.

Robert Wiblin: I’ve heard that one of the big biases in academic publishing is that the top journals tend to be very theoretical, at least in economics. Is that one of the issues that makes it difficult to publish in top journals, or is it another concern?

Michelle Hutchinson: I think that’s a concern to some degree, because it’s certainly also the case in philosophy that the top journals are very theoretical. I don’t anticipate that being a big problem for us, because we’re trying to do pretty conceptual work, and a lot of the work that’s most questionable within the more applied space is that you tend to get people working on it who don’t have such good formal foundations. And so I expect that quite a lot of the work we do will be theoretical, deliberately.

Robert Wiblin: Are there any other downsides to working through academia rather than, say, starting a nonprofit to do this research?

Michelle Hutchinson: Well, I’m doing the operations side of this, and I can conclusively say there are. I mean, part of it is simply what you might expect from working within a really large, very old organization, compared to a start-up, which can do whatever it wants.

Robert Wiblin: I guess Oxford is literally 200 times older than the Centre for Effective Altruism. No, much more than that. I guess it would be like 400 times older. Anyway, it’s many multiples, so I guess they’ve got a lot more rules that they’ve built up over that last 800 or 900 years.

Michelle Hutchinson: They do, and not only do you have the problem of rules, you also have the problem of no one knowing what any of the rules are, because there are a lot of people, and the rules were come up with gradually. And as a big organization where there are a lot of people at stake, they simply have a lot more to lose than the Centre for Effective Altruism, and so there are a lot more constraints. For example, in order to hire a person, you have to have the money to pay for their entire contract. So if you’re going to hire someone on a five-year contract, then you have to know where all of that money’s coming from. If you want to hire someone for an indefinite period, then you have to have an endowment, the interest of which pays that person’s salary indefinitely. So it costs in the region of £5 million to hire someone indefinitely.

Robert Wiblin: So this is very different from a typical business or start-up, where you might only … well, at a limit, you might only have enough to make payroll that month, plausibly, and you can still hire people if you’re willing to take the risk.

Michelle Hutchinson: Yes.

Robert Wiblin: I noticed that originally people were referring to the plan for this organization as the Institute for Effective Altruism, but at some point it switched to the Global Priorities Institute. Was that to keep the university happy, or was there some other reason for the name switch?

Michelle Hutchinson: The university was quite happy with the name switch, but we had a number of different reasons for it. One was we wanted to make really clear that the aims of the Institute are broad and to be thinking about not just what individuals should do with their resources, but what larger bodies should be doing with their resources, whether government or inter-governmental organizations. Although the effective altruism movement also obviously cares about that, the name itself doesn’t necessarily lend itself to that interpretation. People hearing about it for the first time can easily assume that it’s only really individualistic. We also wanted to make clear that our main focus was the most important priorities, rather than smaller considerations, which again, to individuals can be pretty significant, like should I be buying fair trade or should I instead be buying coffee from countries that are poorer. That kind of thing. That wasn’t going to be our main focus.

Then we wanted to make sure that we attracted really talented researchers, and so we wanted to have a name that didn’t already have too many connotations, so effective altruism was pretty fine for that, although there is some connotations to effective altruism within academia. That tended to rule out some other names, which already had quite a bit of work done in academia, like “well-being” and “welfare” that had particular understandings, which weren’t necessarily the understandings that we would take. And so we were trying to take a name that was a bit more of a blank slate.

And then finally, we wanted to have a name that was definitely future-proof. You can imagine that a young social movement might want to change its name in a few years’ time if some particular thing happens, whereas setting up an institute at a university that’s a thousand years old is not going to be able to change its name in that kind of way, and so it needs to be the name you’re going to have for a hundred years.

Robert Wiblin: You won’t be able to get approval for a name change within kind of a human time scale. So how do you envisage your research being different from what’s already been done in academia before?

Michelle Hutchinson: So far, a lot of the work that’s been done in academia on effective altruism has been relatively piecemeal and ad hoc, understandably. It’s individual academics taking the initiative and putting on a particular workshop or teaching a course to their students on effective altruism, without necessarily coordinating on what the best curriculum might be and things like that. I think in general that’s a great thing for people to do, but there’s a big risk for effective altruism that it could fall into a pattern where, in general in academia, certainly in philosophy and also in economics, the more applied something is, the less well thought-of it is. So typically, metaphysics is thought to attract better philosophers than ethics is, and applied ethics is thought to be worse again.

Effective altruism, by its very nature, is very applied, and so could easily end up not being thought well of. I think that’s the kind of risk that we can head off, and there are other fields which have managed to head off this kind of risk in the past. One example I like to use is that of neoliberalism, which generally had a pretty bad reputation in say the ’40s and ’50s compared to more socialist kind of economics, but managed, at least partly through concerted effort, to become the dominant economics line of thought by the ’70s. I think in order to do that kind of thing, though, we have to be pretty considered in how we do this, and so we should make sure to have a more systematic approach than what’s been taken before.

Another thing that I think is different about it is that so far, a lot of the work on effective altruism within academia has been very much on the question of should we be effective altruists, and we would like to get away from asking should we be trying to do the most good and towards if we are going to do the most good, how precisely are we going to do that.

Robert Wiblin: A staff member at the Centre for Effective Altruism wrote a blog post about this neoliberalism movement, and to be clear, not saying that the neoliberalism movement is great, but rather what lessons can we learn from their success, and one of the things they did was build a base of support in academia. Is that right?

Michelle Hutchinson: Yes, that’s my impression, that they worked quite hard on starting by building up these ideas within academia, part of the reason being that it’s a good place for really improving and testing the ideas and working out what the exact correct formulations are and getting it battled from all sides, in order that you make sure that you really have a rigorous understanding of why these are the right ideas and these ones aren’t. And another reason being that academia is extremely influential. People don’t necessarily think of it this way, but academics are the experts who are consulted by the media, by policy makers. They’re also the people who train the next leaders of the country, and so it’s important that academics have the right kinds of frameworks to be teaching their students.

Robert Wiblin: In starting up a new organization, it’s often very challenging to kind of match up all of the things that you initially need in order to get it off the ground. You’re going to need probably at least one researcher to get something written, in order to show people what you’re about. You’re going to need one operations person to file the paperwork and deal with the university bureaucracy, and you also need probably some funding, some seed funding just to make even the plans, to figure out what you’re going to do. Has that kind of been a difficult element of getting this off the ground?

Michelle Hutchinson: We’re in the pretty fortunate position that we have some researchers already working on this stuff, so Hilary and Will I mentioned. Then we managed to get fairly early on a grant for an operations person, which is me. And because of the broader effective altruism movement, getting seed funding for this kind of project also wasn’t too tricky. I think we’re going to have more problems in the future, because as I mentioned, it’s actually very expensive to do things within the university system compared to typical charities, and so it’s going to be a bit more difficult to make the case for we need £5 million to hire a person in the long run than it is we need some money to hire a person for two years to get this off the ground.

Robert Wiblin: You mean it will be difficult to persuade effective altruist donors that this is the most cost-effective use of their money relative to hiring similar researchers in a nonprofit?

Michelle Hutchinson: I think in particular, effective altruist donors. I think donors in general, though. Academia also doesn’t have at all a good … well, at least philosophy doesn’t have a good track record for attracting individual donations. So I think it will be tricky regardless of where we try to fundraise from.

Robert Wiblin: I guess the argument would rest on the issues raised earlier, that it might be easier to hire and that it’s really important to have a foothold in academia and to engage with some of the really smart people that are there, and also, I guess, to get the credibility that comes with having an academic institute.

Michelle Hutchinson: Yes, exactly. And often even just to pivot people who are already within academia and doing good research but not the most important research. It could actually be very significant to get them working in a centre like this rather than in isolation, but if they’re already an established academic, you really need to give them a proper academic post.

Robert Wiblin: Do you know how much more it costs to hire someone within the university system rather than just, say, through a nonprofit to pay them to do the research?

Michelle Hutchinson: It would depend on a lot of things. So, it would depend on the salary, and it would depend on whether you were trying to hire them long-term or not. I’ve mentioned this problem of in a charity, you can give someone an ongoing contract and say we will fundraise in a way that you can’t in the university system. Then there are a lot of separate costs for hiring a researcher within the university, where it’s not totally clear which of them particular people have to pay. There’s kind of a flat infrastructure charge for things like libraries and administrative staff and things, which is charged to everyone, but then there are quite a lot of indirect costs, which are at least half the person’s main salary, typically gets up to basically all of their salary, which are charged in some cases but not in others, so it’s a pretty difficult question to answer in the abstract.

Robert Wiblin: Would you say at the moment you’re more limited by more access to more funding or by the ability to hire the right people?

Michelle Hutchinson: It’s a little bit tough to tell. I think probably the biggest bottleneck right now would be finding the best people to hire, although we’re at a fairly early stage in fundraising, and so it seems like it’s going relatively well, but we’re not fundraising from huge numbers of donors, so it could easily end up that our grant application doesn’t go well, and then suddenly we’re very constrained on money. Then another bottleneck that doesn’t typically occur for other EA organizations is the administrative one, where it’s a little tricky to hire for people when the university doesn’t yet acknowledge GPI as its own thing, as distinct from Future of Humanity Institute, which officially employs people like Hilary. So even when we have good people in mind who we would like to hire and the funds to do that, it still is going to take us a while to actually hire them.

Robert Wiblin: Just quickly, roughly how long does it take you to go from advertising a job to actually being able to hire someone within the university system?

Michelle Hutchinson: It’s actually not the advertising that’s necessarily the problem. It’s getting an advert out there, because there’s quite a lot of different levels to go through to work out how much the person should be paid, whether the job advert is the right one, that kind of thing, because there are very strong incentives within the university that there’s just not enough money going around, and that there are permanent posts which aren’t fully funded, and so the fundraising should be prioritizing funding posts that already exist. And so justifying we’re going to open a new position and hire for that is quite tricky.

Robert Wiblin: Sounds like you’re doing the Lord’s work dealing with all of this. Let’s come back to the funding for a minute. It might be a little bit challenging to persuade EA donors to donate to this, but because you’re part of the university, you have access to a whole lot of grants that I guess are specific for universities. Is that where the fundraising is going well at the moment?

Michelle Hutchinson: No. We did some grant writing for academic grants early on, and we did have some luck with small grants, but large grants we didn’t end up having much luck with, I would guess in part because we were trying to have a fairly fast turnaround, so rather than putting a lot of time and effort into it, we were doing a fairly quick job to figure out whether we could get a large amount of funding relatively easily. And the answer was that we couldn’t.

I think one hesitation we have with using that kind of route for fundraising is that you have to put a lot of work in, and it’s work that’s not necessarily particularly relevant typically, because often the funder will have quite a strong view of what they would like, and it’s not necessarily aligned with us, whereas if we’re trying to fundraise from effective altruist donors, in those cases our incentives are very much aligned, so the kind of work they would want us to do is the kind of work that would be very useful for us anyway. You know, think more about your model. Come up with more of a plan. We think that this thing’s problematic. Have you considered … And so often, they’ll be providing extremely useful feedback. So even if the process is protracted, which it typically isn’t nearly as much as the more university-focused ones, it will be a very useful process.

Robert Wiblin: Yeah, writing grants sounds like a nightmare for a whole lot of reasons. You don’t get a whole lot of feedback. It takes quite a long time to prepare them. You’ve got relatively low odds of getting the actual money. I guess it’s quite risky, because you put in a couple of grant applications, and even if they’re all good, it’s possible that you could fail to get all of them just because of a bit of bad luck. Have you found it to be a frustrating process yourself?

Michelle Hutchinson: Yeah, I think I have found it pretty frustrating, in the sense of, as you say, often grant agencies will give you no feedback at all, so you won’t even know why you’ve been turned down, and then there’s also a very long lead time. Typically, you wouldn’t expect to get the money for about a year from when you put in the grant application, and it makes it very difficult to plan and to hire people if you’re saying, “Well, we might have money in a year’s time.”

Robert Wiblin: Yeah. Is it the case that academic institutes tend to grow pretty slowly for this reason, that the time scales of hiring and attracting funding are over many years?

Michelle Hutchinson: Yeah, I think that’s about right. There are also not very many research institutes in philosophy typically, and often the kinds of institutes that are, aren’t doing research themselves. They’re more collecting together people who already work for the university and getting them together for workshops or to do that kind of thing, rather than saying as a research institute, we’re going to expect you to do different work from what you were doing, and in exchange we’re going to offer you a contract that involves no teaching and things like that.

Robert Wiblin: So what you’re doing is a little bit unusual, perhaps, within the university system?

Michelle Hutchinson: Yes.

Robert Wiblin: So if the grants don’t come through because perhaps there just aren’t that many grants that are designed to support exactly this kind of research, you’ll be looking more at donors who are interested in the answers to these questions and particularly want to fund it because they value this research?

Michelle Hutchinson: Yes, that’s right. We’ve so far had most luck with individual donors who are fairly effective altruist-aligned and really want answers to these. Also some donors who think that maybe the more promising route to doing research that’s action-relevant in effective altruism is the less academic ones or through open philanthropy, because they can get answers faster, but who think that it would be really beneficial to have academia in general engaging more with these kinds of questions, and also who are very interested in seeing ideas that may be already accepted by effective altruists, like that there should be a zero percent discount rate or something, written up extremely cleanly and well in order to be more persuasive.

Robert Wiblin: So if anyone is listening to this and might be interested in donating to GPI, or at least hearing the case for doing so, then they can email you at [email protected], is that right?

Michelle Hutchinson: Yes, that’s right.

Robert Wiblin: Okay, great. If someone did donate to GPI, what kind of outcomes might they hope for in the next couple of years? What might be your key performance indicator? What kind of metrics would you be looking at each month to tell whether the project’s going well?

Michelle Hutchinson: The thing that in the short- to medium-term we’re really going to be looking at is how many articles on global priorities topics we’ve had published in top academic journals. A nice thing about this measure is that it’s relatively objective. People tend to agree on what the best philosophy and economics journals are. On the other hand, it’s a metric which is going to take a little while to come to fruition, and also one that’s not going to produce result on a month-by-month basis. So there are a few other kinds of things that we’re using.

An internal one is how many informal write-ups we have that we think could be turned into good journal articles. Another is whether we’re getting good academic visitors. We, as well as hiring researchers, would like to have top academics at all levels of career coming to visit in order to get them engaged with the project, as well as putting on more general workshops and things, and in the longer run, getting good academics around the world writing on these kinds of issues. Shorter run, we hope also to increase the number of universities at which courses are run on global priorities research and effective altruism in general. So those are the kinds of things we’re going to be thinking about in measuring our impact.

Robert Wiblin: Let’s move on from talking about funding to talking more about hiring. What kinds of people would you be open to hiring in the coming years?

Michelle Hutchinson: In terms of researchers, we’re really looking at people at a number of different stages of their career, from coming straight out of PhD’s to fairly established within academia. We actually would be interested to talk to people even if they’re relatively early on in their PhD’s, because we want to be making sure that we build relationships with good researchers from early on. At the moment, we’re primarily looking at philosophers and economists, and we’re concerned to hire not too many people but really excellent researchers. It obviously helps to have gone to a top university, but the thing that we’re going to primarily look at is the quality of people’s research.

We also care a lot about having people who are really driven to work on the most important problems and how we can help people the most, rather than people who happen to be in somewhat related areas, because it’s relatively difficult as a research institute to make sure that people work together on a research agenda and remain focused. So we want to have people who are relatively value-aligned from the start.

Robert Wiblin: I’ve heard that managing researchers is a little bit like herding cats, so you want people who are kind of intrinsically motivated to work on topics that you as management think are particularly valuable, is that right?

Michelle Hutchinson: That’s the hope.

Robert Wiblin: Okay. Is it important that they’ve done a PhD in the relevant area, or if they’re very smart and they’ve got some good publications, or they’ve done good research in the past, are you open to people switching to work on the topics that you care about?

Michelle Hutchinson: We’re definitely open to people switching. In some cases it might be difficult to hire people if they don’t have a relevant PhD, and it also depends how far they’re switching. For example, it could be that someone’s PhD is in logic or something, but actually they want to move to working on ethics, and that’s the kind of switch that we could accommodate if the person feels that they would be happy to do that. If someone’s PhD is in, say, maths and they want to switch to working on ethics, that would be a bit trickier, but we would assess it candidate by candidate.

Robert Wiblin: What things do you think listeners could do if they’re undergraduates or perhaps post-graduates to put themselves in a good position to potentially apply to GPI in the future?

Michelle Hutchinson: One thing that’s obviously important is really to double down on their academic work and produce work which really showcases their talent. Another thing is thinking carefully about what topic to work on. As we touched on earlier, the incentives in academia don’t necessarily push towards working on the highest priority issues, and it can often be rather difficult to work out how to make an academic career in those kinds of things. I would absolutely recommend that people engage with other people, particularly in the effective altruism community, to work how that might be possible.

There’s a Facebook group, EAs in Academia, which might be a good first starting point for discussing this kind of thing with other people, because I know when I was choosing what to do my PhD thesis on, it was far from clear how you go about that at all.

Robert Wiblin: Yeah, if I can be a bit blunt, maybe blunter than you want to be, going into academia is extremely difficult, as we mentioned, and there’s a lot of different filters that you have to get through, and many people who are trying to do it. So it’s very competitive, even for people who are very academically bright. And it’s also the case that global priorities research is very selective as well. It requires people who can be quite creative, but also have really good foundations for thinking abstractly and working through these difficult questions where sometimes it’s hard even to tell what the right question to ask is, let alone to figure out what the answer is.

So there’s a lot of people who are interested in working in global priorities research, and I think the reality is that most of them are not going to be able to pursue a career in that, and probably even a smaller fraction of people would be able to pursue in it within academia, because that presents additional barriers. Would you say I’m broadly on the right track there?

Michelle Hutchinson: Yes, I think that’s about right, and I think the one thing that people don’t consider enough is research-adjacent types of roles, so roles as research assistants or as project managers, that kind of thing, because actually that also requires really talented people, and it’s the kind of thing that because within academia is not thought well of, often doesn’t attract very good candidates at all. And I think that’s really where people can find niches where they can have a huge amount of impact.

Robert Wiblin: How hard is it to hire good operations staff and research assistants?

Michelle Hutchinson: It’s pretty difficult. Operations types of things often isn’t what people studied in their undergraduate degrees, and so it’s not something they’re really aware of as a good option. I got into doing operations really entirely by chance when we were setting up the Centre for Effective Altruism, where I’d been doing research. We needed someone to figure out how we set up a charity and how we hire staff, and then it turned out that I really enjoyed doing that and was very pleased to go into this role doing a similar kind of thing, but working out how you do that in a university setting. But many people simply don’t have that kind of opportunity.

Then you also get fewer applicants than you could, because it’s often not terribly well thought of, compared to research. It’s thought that you have to be really smart and generally excellent to be a researcher, but anyone could be a good operations person, which is not at all the case. And then even if you’re getting good applicants, it’s not necessarily easy to figure out who the best applicants are, because whereas when you’re looking at a researcher, it’s fairly obvious how you tell. It’s both from their academic pedigree and also by reading their work. In the case of operations, the work that people have done is often very dispersed, and it’s not as clear what things were down to them versus the whole organization. So there are a number of barriers to hiring good operations staff.

Robert Wiblin: Yeah, this is something that 80,000 Hours has been pushing for a little while. People really don’t often appreciate how much finding good just very competent, reliable operations staff is often a limiting factor on an organization’s ability to grow. We’ve just heard this again and again. Being a research assistant for someone who’s really good, who’s doing good research, you can potentially increase their output 50%, possibly even 100%, if you’re extremely good at doing all of the work, of freeing up their time, basically, so they can always be writing, and then you help with everything else that they have to do. So I think it’s something that many more people should really be considering, and we have some more research on those topics coming out in coming weeks. I’ll announce those on the podcast once we’ve actually published them.

Michelle Hutchinson: I’m really glad to hear that. I think having good judgment also is something that isn’t discussed very often, probably because it’s very difficult to pin down, but the difference between having a research assistant and operations person who has really good judgment and understands the work that’s being done and what needs to happen, and having someone who’s really just trying to work out precisely what you said and just do that, can be really massive.

Robert Wiblin: Yeah, because in one case you can let them do a whole lot of things quite independently with fairly little oversight, and you trust them, whereas in the other case you have to look at everything all the time, so it may not be even freeing up any of your time at all because of the management overhead.

Michelle Hutchinson: Right.

Robert Wiblin: I imagine a question quite a lot of people might have is what are the odds of this research really having an effect on the real world? A lot of people have this impression that academic papers are often only read by the person who writes them and maybe the two people who peer review it, and maybe their parents. Is this a bit of a risky thing to do, inasmuch as you might publish these papers but then they don’t really get listened to by the relevant folks?

Michelle Hutchinson: I think it is, and it’s an important question. In the cases of papers that we’re writing where we’re really trying to answer questions that the effective altruism movement doesn’t yet have the answers to, I think one thing that we importantly have going for us is a close connection to people who are really trying to make a difference in the world. We actually share an office with the Centre for Effective Altruism, and we also have links with other organizations. We have an Open Philanthropy staff member on our board in order to try and make sure that we can ask them for what kinds of work we should be doing that would be useful to them, and also whether we’re producing in a form that’s useful.

On the other hand, then there’s the research where we’re trying to influence academics rather than people trying to work out how to do the most good in the world, and in those cases we’re simply going to have to keep a fairly careful track on to what extent our research is getting read. One of the metrics we’re likely to use that I didn’t mention is to what extent we’re getting response papers written to the work that we’re writing, because a lot of academic threads of thought end up coming about by you have one paper that’s got a pretty novel concept, and then that sparks a number of other people to write on similar kinds of topics. So you can really tell whether you’re actually getting read and thought about or not. If we find that actually we’re not, we’ll really have to try and rethink our strategy.

Robert Wiblin: I guess that would also allow you to get leverage on your funding in a way, because you’ll take in a particular amount of money and write some papers, and then you’ll be able to goad other people into doing some of your research for you by perhaps disagreeing with you or wanting to build on the literature.

Michelle Hutchinson: That would certainly be the hope.

Robert Wiblin: Well, it’s been good to get all of this on the record in the very early stages of GPI, and I guess we can revisit it in a couple years’ time and see whether things went to plan or whether they didn’t, and whether your guesses were right about what was going to be the bottlenecks and so on.

Michelle Hutchinson: Just trying to catch me out, Rob?

Robert Wiblin: I guess if there’s any professors of economics or philosophy out there,
they should get in touch with you and potentially apply for a job?

Michelle Hutchinson: That would be wonderful.

Robert Wiblin: At 80,000 Hours, we’re always interested in coaching people who are interested in working in global priorities research, so we’ll have a link to the coaching form in the notes on this episode and on the blog post about it as well. And obviously, we’ll put up links to most of the things that we talked about in the conversation today. So it’s been great to have you on. My guest today has been Michelle Hutchinson.

Michelle Hutchinson: Thanks. It’s been a lot of fun. Hopefully, things go to plan.

Robert Wiblin: Fingers crossed.

—

Outtro: Just a reminder about the two research roles that are currently open at GPI. If you have a philosophy PhD please check them out ASAP.

We’re also keen to coach people who either want to do global priorities research themselves, or work similar to Michelle. There’s a link to the coaching application form in the show notes.

Keiran Harris helped produce today’s episode.

Thanks for joining, talk to you next week.