Transcript

Rob’s intro [00:00:00]

Robert Wiblin: Hi listeners, this is the 80,000 Hours Podcast, where each week we have an unusually in-depth conversation about one of the world’s most pressing problems and how you can use your career to solve it. I’m Rob Wiblin, Director of Research at 80,000 Hours.

Sometimes we take a more indirect route, but today’s episode tackles that question, of the world’s most pressing problems and how you can use your career to solve them, in the most direct way possible.

Last year we published a single page summary of all our key ideas, which links to many of our other articles, and which we are aiming to keep updated as our opinions shift.

All of us added something to it, but the single biggest contributor was today’s guest, our CEO Ben Todd, who founded 80,000 Hours with Will MacAskill back in 2012.

This key ideas page is the most read on the site, and by itself can teach you a large fraction of the most important things we’ve discovered since we started investigating high impact careers.

But when I say single-page summary, it’s perhaps more accurate to say it’s a little book, as it weighs in at over 20,000 words.

Fortunately, it’s designed to be highly modular and many people work through it over multiple sessions, browsing through the articles it links to on each topic.

Perhaps though, you’d rather absorb most of our key ideas in the form of a conversation on this podcast. In which case you’ve hit play on the right MP3, because that’s exactly what this is.

One benefit of this podcast over the article is that we can more easily communicate uncertainty, and dive into the things we’re least sure about.

If you want to have a big impact with your career, and you say you’re only going to read one article from us, we say to read our key ideas page.

And likewise, if you’re only going to listen to one of our podcast episodes, it should be this one.

One note is that our advice is going to keep shifting, and we’re aiming to keep the key ideas page current as our thinking evolves over time.

Of course, that’s not going to be possible with this podcast, so this represents our views as of November 2019 when we launched the current version of the Key Ideas page, and actually recorded this episode.

OK, without further ado, here’s me and Ben Todd discussing our new guide to solving the world’s most pressing problems with your career.

The interview begins [00:02:18]

Robert Wiblin: Thanks for coming on the podcast, Ben.

Ben Todd: Hey, it’s great to be here.

Robert Wiblin: So this is going to be a pretty different episode from the usual ones because as you might guess listeners, I have a reasonable idea of what Ben is going to say in most cases as we’ve been working on this together for the last couple of years. So instead, we’re going to assume that listeners have either been subscribed to the show for awhile and have heard a few other episodes or at least have skimmed over the key ideas page. And then we’re going to basically introduce all the issues that we raised in the key ideas page and then try to talk about some of the things that we perhaps didn’t have space for, some of the more interesting and controversial issues that are in that topic area. But before we get to all of that, maybe we can start by Ben, could you explain how you ended up running 80,000 Hours and what the history is there?

Ben Todd: Yeah. So what got me involved was I saw Toby Ord give a talk at my college when I was an undergrad studying physics and philosophy and Toby was also at my college and he gave his ‘Taking Charity Seriously’ talk that you can see on YouTube and that basically convinced me pretty much right away… I’m not going to rehash the things that were in there, but it was basically the idea that some charities you could donate to are much more effective than others and the best ones are super effective. So it’s a really worthwhile thing to donate some of your money to those charities. And about a week later I signed the ‘Giving What We Can’ pledge and I think I was the first non-founding member of Giving What We Can. It was right after the launch and then I was just involved with this community in Oxford of people who were trying to figure out the most effective ways to have a positive impact.

Ben Todd: At that time, most of us were students. One of the big questions facing us was what should we do with our own careers? And I was wondering the same. And so I volunteered to give a talk on how to choose a career and how some of the ideas that Toby covered in Giving What We Can might also apply to career choice. Will MacAskill happened to be in the same room and was also coincidentally thinking of doing a talk on the exact same topic. So he suggested we team up and we worked on the talk together actually in my room in Balliol. And yeah, in early 2011 we gave it for the first time. We gave the first talk and that actually turned out to be the most successful talk we’ve ever given. So I think there was about low twenties people in the audience and in the end, over the years after that, about six of them I think made pretty big changes to their careers based on this.

Ben Todd: One of them, a year or two after the talk, joined 80,000 Hours and is still with us. Someone called Habiba decided to take the further pledge, which was a pledge to give all of her income above a threshold, I think probably around 25,000 pounds to charity, stayed involved and actually recently switched to working as a senior administrator at the Future of Humanity Institute. So yeah, these kinds of changes convinced us that we might be on to something. Some people in the audience such as Richard Batty actually approached Will and me and said maybe we should start an organization around this idea, in addition to Giving What We Can.

Robert Wiblin: Yeah, I wasn’t around there at the time. Did you guys spend very long debating whether you actually wanted to start an organization and kind of who should run it?

Ben Todd: Probably back then, we were a bit too cavalier with starting organizations. Probably early on we set up too many things and the Oxford community should have just focused on making one project go really, really well though obviously it turned out well in my case. So yeah, it didn’t seem like we debated it much but I mean of course we started it just as a volunteer side project while we were still students for a whole year and then we ended up getting pretty major press coverage on the BBC about the idea of earning to give and we grew our blog quite a bit and then we managed to raise a couple of hundred K of funding, which was enough to to hire full time staff. And so it was that over that year that all those, all that traction that we were getting was, at that point it seemed obvious that we were onto something.

Robert Wiblin: Yes. At that early stage, 80,000 Hours, I suppose it was called High Impact Careers first and then 70,000 Hours and then 80,000 Hours.

Ben Todd: Yeah. Not many people know that we were briefly called 70,000 Hours. Yeah. In the summer of 2011 it was High Impact Careers and we started with a beautiful red and black website. Then we rebranded to have our color be magenta, which most people called pink, and then we switched, well, we switched to the blue actually more than a year or two after that I think.

Robert Wiblin: Important design issues there. We should dig up the website actually from 2011 on archive.org I think that would be a pretty funny to look at. Anyway, early on 80,000 Hours was doing all kinds of things. It had a blog. It was running all sorts of events, but I guess over time it kind of gained a bit of focus and actually stopped using just interns and actually hired some people.

Ben Todd: Yeah, I think like most startup nonprofits we made the mistake of trying to do too many things at the same time and we were kind of thinking of ourselves partly as an advocacy campaign around certain big ideas in effective altruism. We were partly thinking of ourselves as giving one-on-one careers advice, partly doing lectures, partly doing research into what to actually say about these topics and all of these activities are pretty different. And for a small team it’s way too much. So the first few years we just really tried to focus the concept down a lot and eventually we focused on just what we’re about is giving people information and support to help them change their career. And now we just do two main things, which is online content and one on one support.

Robert Wiblin: You were at some point considering doing Earning to Give yourself then decided to found a nonprofit instead which I guess may have been kind of contradictory to our advice at the time. How did you end up making that decision?

Ben Todd: Yeah, I mean 80,000 Hours early on got very associated with Earning to Give and our very first lecture, the structure of that talk was the arguments for Earning to Give and why that could be better than most typical social impact jobs. But then the second section of that talk was about how you might be able to do even more good than Earning to Give, and we covered our research and advocacy careers, and government jobs and threw out a whole bunch of other ideas that seemed like they might turn out to be better. And so even at the very start we never thought that Earning to Give was typically the best option all considered. And yeah, that was very obvious in my choice where I was choosing between Earning to Give and working at 80,000 Hours. And so yeah, I had a job offer to work in investment, in finance, which is something I’d been interested in actually since I was a teenager.

Ben Todd: So I think that would have been pretty interesting. I would have probably enjoyed it. But yeah, we really got convinced by that time by what we call the multiplier argument, which is just this idea that if you could just change one person’s career and they’d go and do something really high impact, then that’s kind of having as much impact as you could have in the whole rest of your career. And so if every year you could help one person switch into a high impact career, then you’re having way more impact than you could yourself directly. And we actually thought we could do much more than help one person change career per year.

Ben Todd: So we thought it through and in my case, just carrying on with 80,000 Hours would be the highest impact thing to do. Actually back then we hadn’t really thought of the concept of career capital and so I just totally ignored that in the decision. But actually I think, just by luck, I actually turned out to get better career capital by doing 80,000 Hours than I would have in finance just because I’ve met lots of other people who want to focus on social impact and I learned a lot about how to run a social impact organization. And so actually maybe if I’d known about career capital back then it would’ve led me in the wrong direction. But fortunately I was doubly wrong and it canceled out.

Robert Wiblin: Career capital for those who don’t know is this term of art we use to describe everything that puts you in a better position to have a bigger impact in the long-term, including skills, connections and credentials.

Common misunderstandings of our advice [00:10:16]

Robert Wiblin: I guess part of the reason that we’ve written this key ideas article is to try and make our ideas super, super clear and very upfront because we found that sometimes people don’t quite understand what we’re saying, or they read some part of our advice and don’t understand the big picture and as a result, they end up doing something that we wouldn’t advise if we were talking to them one-on-one. What are examples of classic or common misunderstandings that people have had of our advice or cases where we’ve communicated our ideas in not such a good way and so that’s resulted in people making decisions that we wouldn’t necessarily have advised if we’d spoken to them one-on-one?

Ben Todd: One thing we’ve realized in recent years is that at least the core of our audience really do just want to get the big ideas directly and explained with a lot of nuance. I think partly this podcast has shown there’s an appetite for that. And so we decided to recently replace our career guide online with what we’ve called the key ideas series. The career guide is still up, but we now make the key ideas series the most prominent one and it just tries to lead with our most distinctive ideas quickly, such as longtermism, existential risk, things that make us the most different from other people who might give you information on what to do to have a social impact. Another part of why we want to do that is because we thought some of our most important concepts are actually getting a bit lost in the existing career guide.

Ben Todd: And so there’s been a couple of misunderstandings we think we’ve been coming across recently and the one that’s most on our mind at the minute is just how people think about career capital and in particular a lot of people getting the idea from our content that typically the best thing to do is to work in consulting early career or some other prestigious corporate job like that rather than trying to do something more directly relevant to a pressing problem. Another option is we’ve seen people who just have great direct impact options right away. They could do something like go and work at a top think tank on AI policy or they’re thinking, “Maybe I should go and work at PwC first to get better career capital”. And it just seems better to push on in policy because even just from a career capital point of view that’s giving you a lot of other options in addition to being closer to having an impact.

Ben Todd: One example we came across recently was a bit of an unusual one, but it was someone who had the realistic possibility of becoming a magician and maybe landing a major TV show in India and they were choosing between that or otherwise consulting. And it actually seemed to us like the magician case was more exciting because it would mean that, well firstly that’s kind of more of an impressive level of achievement. You might stand out more, but also, you’re learning all about media, building up an audience and you’re also doing this in the context of a really important country. So those kinds of things seem like you might well get better career capital from actually becoming a magician rather than consultant. Whereas I think most people’s reading our advice would be like, in general, a consultant is the thing that we’d want people to focus on.

Robert Wiblin: Yeah, it somehow kind of seems more serious and so they might think, “Oh yeah, that’s the thing that 80,000 Hours would recommend”. But I guess we’re very keen a lot of the time to have people explore interesting things that provide unique opportunities for them that other people don’t have. Of course it’s somewhat hard to communicate that as we’ve mentioned on the show before because of course we can’t have a priority path as ‘magician in India’ anywhere on the site because that’s such an idiosyncratic opportunity. But yeah, I guess that is a common misunderstanding that people think that we want to pigeonhole them into particular common career tracks rather than encouraging them to do things that are particularly promising opportunities for them.

Ben Todd: Well yeah, so you always have to weigh your personal fit and comparative advantage against how good the path is in general. And if you have sufficiently high personal fit in a path, it can can outweigh that area generally being less promising.

Robert Wiblin: We use this term comparative advantage which is a little bit jargony. I guess comparative advantage is about your relative goodness at something relative to the other people who are in your organisation. So you could end up doing something which you are in some sense absolutely less good at but you’re filling some gap relative to what other people can do who you’re trying to coordinate with.

Ben Todd: Yeah. And we have a whole article about comparative advantage to define it a bit more clearly.

Robert Wiblin: I guess another mistake that I sometimes see is people giving too little weight to choosing the right problem to work on. If you ask them, they might think working on one problem, problem A is like 10X more effective than working on problem B, kind of all else equal. But then, they’ll end up working on problem B just because they managed to get a job in that sooner, or they feel that they have a somewhat better personal fit for that or maybe they just haven’t spent that long thinking about it, and it feels that they’re potentially sacrificing a lot of gain there if they actually really researched which problem that they want to work on long term and made a conscious decision about that.

Ben Todd: Yeah, I mean it’s very hard to wrap our head around these massive differences in impact that might be possible and this is one of the big motivations for 80,000 Hours which is just it seems like some options might just have more than a hundred times more impact than another. And so that kind of means that, say, 10 years working on the more effective area could be equivalent to a thousand years working on the second. So really, really huge differences.

Ben Todd: Yeah. Sometimes it’s easy to maybe intellectually accept the case, but then when it comes down to specific options, kind of our intuition is more that things are roughly similar. We actually did a survey of this where you ask people how much they thought charities differed in effectiveness in terms of saving lives and I think the difference was only 1.5.

Robert Wiblin: Yeah, it was really small differences. I think people thought that the best charity versus the median charity work in the developing world was maybe twice as effective. Whereas I guess we would think of small like a hundred times or at least like tenfold.

Ben Todd: Exactly. Yeah. And this gets us to the whole idea of 80,000 Hours where if you can just spend one percent, if you can make your career just one percent better, it’s worth spending 800 hours doing that. If you can make your career a hundred times better, it’s worth almost spending your whole career saving up for that last six months.

Robert Wiblin: I’ve lost count of how many times I’ve said that over the years.

High level overview of what 80,000 Hours generally recommends [00:16:14]

Robert Wiblin: All right. So before we get into the body of the key ideas article, maybe let’s just start by mapping out the process that, at a very high level, we would recommend that people generally follow if they want to have a really high impact career because that will kind of map out the structure of the conversation to come.

Ben Todd: Here’s the very high level overview of the overall process we generally guide people through when we’re advising them on what their long-term plan should be and what the next step should be. And it’s also roughly the structure of the key ideas series itself, it moves through things in this order. So the first question we encourage people to think about is the question of which global problems are most pressing and then which ones you want to work on. Once you’ve got that, then it’s about coming up with career options that might help you especially effectively address those problems. So that’s like figuring out key bottlenecks facing those issues and how you might resolve them. And then from there you want to be trying to get a short list of promising long-term options together. So that’s the kind of idea generation phase. Then we have a strategy section, which is then more about how do you narrow those down?

Ben Todd: And so a really big question there is just which one’s the best personal fit for you? Another big question is whether it’s more effective to first gain career capital? Will that accelerate your career more or should you just enter directly into one of these long-term paths? And then finally there’s the section on decision making. How would you actually figure out, once you’ve got a short list, what your next step should be? And so the key thing is to try to identify your key uncertainties about these different options, investigate them, and then we recommend that generally people take what we call that upside option, which is the option that would be the highest impact if you performed towards the top end of your expectations.

Ben Todd: That said, doing that while having clear backup plans. And then finally we roughly recommend that people review their career about once a year or whenever they face a major decision point, and a little bit less often if you’re more confident and a little bit more often if you’re early career or you’re more uncertain about what to do. And then the rest of the time, just focus on doing as well as you can in the option that you’ve chosen.

Robert Wiblin: All right. So that’s a lot to digest, but we’ll gradually work through each of those phases and explain how we would imagine that someone could do a good job with them.

Key moral positions

Robert Wiblin: All right. So the key ideas article starts with a very foundational question of our moral views. So yeah, how would you summarize the key moral positions that 80,000 Hours takes that are relevant to people’s career decisions?

Ben Todd: So yeah, we start the key idea series with, we just state some of the key values behind our advice. We don’t really justify these things because that gets into a lot of complex debates in philosophy. But we just thought it’s important to try to clarify these things and be transparent about our values. And the kind of question you can have in mind here is just what does it actually mean to make a difference? What are we taking that project to mean and then how does that then carry through into the rest of our advice? So yeah, we highlight a couple of key principles on the key ideas page.

Impartial concern for welfare [00:18:17]

Ben Todd: So the first value we highlight is the idea of impartiality. And so we take 80,000 Hours to be the project of trying to do good from an impartial perspective.

Ben Todd: And yeah, exactly how to define impartial is, again, a very tricky issue and it’s debated exactly how it should be defined, but very roughly you could think of it as we want to treat everyone’s interests equally. So that means we don’t want to privilege people based on their particular nation or social group or ethnicity, where they are in the world, even potentially what time they live in. There’s a tricky issue then of how do nonhuman animals fit into that and yeah, we’re unsure about exactly how their interests should be weighed. But again, we would say their interests would be counted to at least some degree though we’re unsure about the exact weighting.

Robert Wiblin: I guess one approach would be that their interest should count to the degree that they’re capable of having interests. So you might think about different species potentially have different strengths of interests because they’re designed differently. That would be one way of reconciling impartiality without giving the exact same weight to a fly as a human.

Ben Todd: Yeah. Though, like I say, those are all controversial. So yeah, we’re trying to treat everyone’s interests as equal. Then there’s a question about what are people’s interests? Like in what way are we actually trying to help them from that perspective, and broadly we say that we should be roughly focusing on increasing welfare. By welfare we just mean a very broad thing about how good people’s lives are and how happy, long, fulfilling, and so on their lives are and how much they can avoid needless suffering. Again, there’s lots of debates about exactly what welfare should consist of. Ultimately, does it mean satisfying people’s preferences? Is it just a matter of people having more positive mental states than negative mental states? There’s other views as well. We’ve actually found that these questions don’t actually seem to make that much difference to our choice of problem areas, which is the kind of more decision relevant thing that we’ll come onto later.

Robert Wiblin: Of course people are going to potentially value other things other than just welfare, impartially considered, but I guess you’re saying our career advice is going to be focused on that because that’s potentially something that a very wide range of people care about to some extent and will have as one of their career goals.

Ben Todd: Exactly. Yeah. I think almost everyone cares about doing good in this impartial way to some degree. Very simple toy example where you might see this is if you could push a button and it would cause a stranger to suffer. Most people would think, well, I shouldn’t do that. Even though I’ve got no particular attachments to that person, I don’t know who they are. And that kind of shows that we do actually care a bit about the interests of strangers and just people in general. And I think many people also share an interest in welfare. Again, there could be many other things that matter, but all else equal, generally if people have better lives where they’re more fulfilled and more happy and they have less suffering, then many people can get behind that as one important goal to push towards.

Ben Todd: Yeah. So like we say, we didn’t think that’s the only thing that matters, but it’s an important goal. And it’s also one where we think just given how the world is right now, then just potentially, due to technology and all the amazing wealth that we find ourselves in and potentially the moment of history that we’re in, our actions today actually can have these really big effects on the long-term wellbeing of large numbers of people. And so it’s also a particularly important factor to focus on, just given how the world is right now.

Longtermism [00:22:32]

Robert Wiblin: Okay, so that’s impartiality. The second part of our moral philosophy is longtermism, which is this emerging set of moral ideas. We’ve had several episodes where we’ve discussed longtermism in some detail and even one episode that was over two hours, which was almost entirely focused on it. And even that didn’t manage to consider all of the different possible motivating reasons and possible objections and the back and forth about that. So longtermism is pretty hard to boil down but one kind of possible one sentence description would be that basically the most important moral consequences of our actions are the impacts that they have. Like more than a hundred years into the future, possibly more than a thousand years into the future. Now I’ll just try to lay out how you could potentially get to a conclusion like that.

Robert Wiblin: The first observation would be that, like many philosophical views imply, the welfare of people in the future is at least an important consideration. It’s not clear necessarily whether the welfare of people in the future is as important as the welfare of a person today, but at least it should be given some weight. The second thing is to note that the future could be extremely long, so the universe is going to be around for a very long time and also just the universe is extremely large. There’s a lot of space, a lot of energy, a lot of matter, and it could be converted into things that are valuable and that could support many people living for a very long time. And potentially another thing is that the lives of people in the future could be a whole lot better than they are now, inasmuch as science and technology over the last 200 years has probably improved the welfare of people and that we might expect that that will continue for at least some period of time.

Robert Wiblin: And then the third thing would be that there’s things that we could potentially do today that could improve the welfare of people who will be alive in the future and that we could have somewhat predictable effects and that they would be positive rather than just completely random. And one way of demonstrating that might be you can imagine that if an asteroid were heading to earth and it hit earth and everyone died, then there’ll be no people around in the future and their lives would just be completely ended. And if we could prevent that from happening, say that it’ll be clear that that would have quite a persistent impact and could result in humanity surviving for hundreds, thousands, maybe even longer than that, thousands of years more.

Ben Todd: Maybe one of the things to emphasize is that each of these three premises could easily be wrong. Maybe there won’t be that many people in the future. Maybe we can’t affect them. But the key thing is just that because there’s at least some possibility that there could be so many people in the future and there’s some possibility that our actions could affect them. Those two things mean that the argument kind of goes through and in particular just because the stakes are potentially so large, it could well be the dominating consideration.

Expected value and counterfactuals

Robert Wiblin: Yeah. So one kind of latent philosophical assumption that’s going on here is the idea of expected value. That we focus on maximizing expected value, which is the size of the reward times by the probability of getting it. So basically we always think that if something is good and you’ve got like a one percent chance of getting it, it’s twice as good if you have a two percent chance of getting it. Twice as likely is twice as good or twice as bad.

Ben Todd: Yeah. And then that’s an important component of how we end up with longtermism.

Robert Wiblin: Yeah. So here we’ve got a group that, in expectation, matters to some degree which is potentially extremely vast in number, like much more vast than the number of people who are alive today and we can potentially affect them through actions that we take now that will have kind of predictable consequences. And that could potentially get you to this conclusion of longtermism, that the most important effects of our actions are those that affect things in the world in more than a hundred years time. Of course there’s many objections that one could potentially field. So people might say, “Well actually, people in the long-term future don’t matter at all morally. They’re not moral patients as far as we’re concerned”. Which conceivably could be true. But I think we think that is fairly unlikely and not true in expectation.

Ben Todd: Probably the most important philosophical position that holds this is called person-affecting views. We’re not going to cover that here, but you can read more about some of the reasons for and against those views in our article on longtermism. And also you can listen to our podcast with Toby Ord.

Robert Wiblin: Maybe an even better source is actually the podcast with Hilary Greaves from the Global Priorities Institute. She’s written this really neat summary paper about population ethics, which goes into different person-affecting views, which I guess kind of march under the banner of, “We’re here about making people happy rather than making lots more happy people”. And she discusses some of the pros and cons of them. And I guess, also why she, I suppose like us, ultimately doesn’t find them to be the most plausible theory. Although I suppose they are, as you said, among the most popular philosophical views that people take which I guess would be a big challenge for longtermism potentially.

Robert Wiblin: You might also think, well, it’s just so unlikely that we could survive into the future so then in fact, there aren’t going to be many people in the future. But that just seems kind of empirically wrong. It seems there is a decent chance that humanity could survive for just a very long time. There could be a whole lot of people. Maybe a third, more complicated objection that some people will raise is just that yes, we have very big effects on the long-term future, but they’re just so unpredictable that it’s very hard for us to improve how well the future goes.

Robert Wiblin: Because it’s just so chaotic and random that one can’t really make it better through any deliberate actions that we take. And I think there’s something to be said for that. But again, there’s this expectation argument that it’s possible that the effects of actions are just completely random. Like whether they’re positive or negative or neutral. But I think there’s a reasonable chance that in fact there are things that we can do that would make the future better. Like, reduce the risk of a nuclear war–

Ben Todd: And we’re going to actually try and consider specific examples and I think the way to just assess that is just to consider specific proposals and whether they seem convincing or not.

Robert Wiblin: Yeah. And I guess, especially since we haven’t been looking for opportunities to do this for all that long, it already seems that we have some pretty plausible things that we could change about the world today that would potentially make the long-term future go a whole lot better. And then there’s even more philosophical and practical objections to all of this, which we don’t have time to go through today as I said, but that kind of gives a really quick synopsis of what longtermism is.

Ben Todd: Yeah. I mean one other big motivating idea in terms of why to prioritize taking this perspective is the neglectedness case. Our existing political systems and economic systems really seem to be optimizing a lot for the interests of people who are presently alive because we’re the people who have the votes, we’re the people who have the money, and so therefore are driving what happens in the world. Whereas the interests of future generations are often entirely neglected. But if we imagine that everyone in all of history had a vote, then they would all massively outvote presently alive people and have a big say over what happens. And because they’re being hugely neglected, that means we should actually expect there to be pretty effective ways of helping people who exist in the future.

Robert Wiblin: So yeah, I was making the argument, a big scale of the issue there because there’s so many and that it’s tractable that there might be things that we could do now that could change the long term. And you’re completely right. It’s very neglected by our economic system, which mostly makes consumer goods for people who are around today. It’s neglected by the government because people who don’t exist yet can’t vote, can’t influence them. And I guess it seems it’s even neglected kind of by the nonprofit or civil society or charitable foundations. I think, in part, just because many of them haven’t been–

Ben Todd: The issue is it’s a new idea.

Robert Wiblin: Yeah. It’s kind of a new idea. So it hasn’t been brought to people’s attention. So there’s potentially a lot of low hanging fruit that’s been left there by other people.

Ben Todd: Yeah. This position only really got staked out in the last couple of decades by philosophers. Derek Parfit was a big figure in that. And then Nick Bostrom and now several other philosophers at Oxford and other places. So overall you might think of longtermism as the idea that when we consider the question of ‘how to make a difference?’, we should just simplify that down to the question of ‘how can we best help the long-term future?’. And we’re not entirely sure that perspective is correct, but we’re kind of convinced enough by the arguments and also because that perspective is so neglected that this is one of the most promising areas to explore and potentially work on for someone who wants to maximize their impact from an impartial perspective.

Robert Wiblin: Yeah, it’s hard to find really high leverage opportunities to improve the world that someone isn’t already taking. To some extent, sometimes you have to make bets that people are making a mistake about something or that a not universally accepted view might be right. And I guess in terms of the bets that we could make that are on a view that isn’t universally accepted that seem especially promising and especially action guiding and that would give us concrete ideas for things that people could do with their career to have a really big impact, longtermism stood out as seeming unusually likely to be something that’s much more widely accepted in the future than it is today. And it could be a framework that would allow us to find many opportunities for our readers to have a big impact.

Ben Todd: One other common objection people give to working on longtermism is just the idea that it’s very hard to get motivated by it because it’s such an abstract idea. And I think that was true for many of us as well when we first got interested in this area. But I think over time, kind of thinking more about the specific problems that we can address, it’s become more motivating. One big thing for me was recently reading Daniel Ellsberg’s book about nuclear war. The level of insanity that is exhibited there where the present generation is just taking this massive gamble with the whole future where we’ve set up this system by which with almost no notice we could just all go up in flames due to this kind of crazy machine that we’ve built. And by doing that we would not only potentially cause this disaster in the present generation, but we might actually put all of history at risk.

Robert Wiblin: Yeah. All of these people who have no say who are just going to being completely screwed over by us because of our own incompetence or selfishness.

Ben Todd: Another thing was a recent paper about the Fermi paradox that Anders Sandberg explained on the podcast where he actually argues that there’s quite a good chance that our current civilization is the only sentient life in the universe. And again, just the idea that the universe might be entirely empty except for this one planet. And then again, there’s all these ways in which we’re just playing roulette with that entire future, I found really motivating.

Robert Wiblin: Yeah. I’m somewhat motivated by this image of a totally unnecessary nuclear war, all of the just totally unnecessary destruction that that would entail. But I think I’m also motivated on a day to day basis by a more positive vision of the future. So I sometimes imagine that if you could transport someone in time from 1500 to today. I think they’ll just be like, “This world is incredible”. There’s so many people and you’ve got laptops and medicine and television and all these drugs that improve your health and science. You understand so much about the universe that we didn’t understand.

Robert Wiblin: And then I just think that if we were able to transport ourselves 500 years into the future and civilization manages to stay on track and continue advancing in the same way that we have in the past. We would similarly just be astonished by how incredibly amazing humanity’s accomplishments are and probably how amazingly good their lives are. Because they just have so much more capability to make their lives so amazing and get rid of all of the problems that we have to worry about today. And it just seems so sad for us to, through our short-sightedness, snuff out this potentially amazing future when it’s completely unnecessary.

Ben Todd: Yeah. I mean one response to that is there are some ways in which the world has probably become worse in the last couple of hundred years, where factory farming is one of the examples we’ve talked about before. I just wonder what would you say to that?

Robert Wiblin: Yeah. I mean it is possible that on balance, the world actually has gotten worse in some ways. If you imagined it could be that the suffering on factory farms is so severe as to outweigh the benefits to people. I guess if you took that view then you might well just want to focus on making sure that those things don’t continue into the future. And I do think that work to eliminate factory farm or to make sure that factory farming definitely doesn’t persist forever is really valuable. On the factory farming case specifically, I think it’s very unlikely that it will persist forever because it just seems it’s a way of making food that we can already see that it might well become technologically obsolete within our lifetimes, let alone over 500 or a thousand years when we’ll be able to make food in just so many better ways.

Robert Wiblin: And just more generally it seems as if people have become more empowered to shape the world with more and more flexibility because you just have more technology, by and large they get rid of these really negative side effects when they can. Yeah. There’s ways that that might not happen, which is for example a reason to kind of spread good moral values so that people would just find it intolerable to have this suffering created for small benefits to them in the future. So that would be another possible longtermist approach that would be trying to shape those values so that the future goes better and try to eliminate the possibility of these really negative things existing for very long periods of time.

Ben Todd: Yeah. I think it’s just important to note that the case for longtermism doesn’t rely on the future necessarily being on this good trajectory where it’s getting better and better. The key thing is just ultimately, could be either much better futures or much worse futures and might there be something we could do about it? And in that case then you should be a longtermist. It might affect what longtermist projects you want to work on rather than whether to take a longtermist perspective in the first place.

Robert Wiblin: Yeah, it’s interesting with the factory farm case. So someone who is specifically not a longtermist who’s worried about factory farming might think we’d have to shrink factory farming right now. I think someone who is taking a longtermist perspective on that, might be like, “We have to make sure that we don’t get locked into a situation where factory farming persists forever, or doesn’t persist for thousands of years in the future”, which might lead to a slightly different edge on how you would tackle that problem. Although it could ultimately cash out on just trying to make beyond burgers really amazing.

Ben Todd: Just going back to motivation, I think one other quick point in terms of what actually motivates me day to day, there’s a lot of things like writing this specific article which is useful or advising this person was motivating because I helped them figure out their career decision. And I think a lot of jobs are like that. Even if your long-term aims are quite abstract, often there’s much more concrete things day to day that also provide a type of motivation even if you’ll never see the beneficiaries of your actions.

Moral uncertainty and moderation [00:35:39]

Robert Wiblin: All right. That was a bit of a diversion there into what actually motivates us to take an interest in longtermism on a practical level, but let’s move on to the third aspect of our ethical views, which is moral uncertainty and moderation. How would you describe that one?

Ben Todd: Yeah, I mean, this is actually just a bunch of different things pushed together, but the general idea is we don’t think you should just kind of be extremist about one moral position and one set of values. And there’s a couple of different reasons for that. One is just the kind of pragmatic reason that often people who are extremists about their values, it seems to actually just cause a bunch of harm. It doesn’t actually lead to the best consequences in the long term.

Ben Todd: If you look more generally at the track record of people who’ve just really tried to push one value system at the expense of others, it doesn’t look like a great track record. So there’s just this, “Well, this isn’t actually the best way to go about things” argument. There’s also a coordination trade argument because we want to be able to work with other people who maybe don’t exactly share our values, but have a bunch of overlap, or even people who just have different values. So again, it’s better to make some concessions to what is valued in general rather than just, again, trying to basically screw other people at every turn just to slightly get an edge on what you care about.

Ben Todd: Then there’s actual moral uncertainty itself, which was covered in the podcast with Will MacAskill. Just that we’re actually very uncertain about ultimately what matters. So we think that suggests we should consider a variety of different perspectives when it comes to what’s valuable and then try and pursue actions that seem reasonable on a balance of perspectives, at least ones that don’t seem terribly wrong on one perspective.

Robert Wiblin: This is an interesting one, because there’s so many different considerations pointing in this direction that almost each seem very strong and potentially quite decisive and then you throw them all together and it seems like a pretty robust general conclusion.

Ben Todd: Well yeah, I mean figuring out exactly what moral uncertainty implies is very challenging. But yeah, we kind of just boil it down to it seems like in general there’s an argument for moderation, and I think one big thing is just not doing things that seem clearly wrong from a common sense point of view. Even if your intellectual case for impact backs them up. Yeah. That “Don’t do crazy things” principle seems pretty reasonable.

Robert Wiblin: Yeah. So you’ve got this tendency probably that people don’t take enough account of philosophical uncertainty about these meta issues. Then we’ve also got that I think humans are not inclined to take full account of all the empirical uncertainty about all of the things that they might not know that might be relevant.

Robert Wiblin: And we’re also probably not inclined to take disagreement of other people in the world as seriously enough. We kind of tend to just go with our gut judgments, even if other people who seem as informed and smart as us disagree, we don’t go, “Well, someone else disagrees with me. Basically, I should just put equal weight on their view as on my own.”

Ben Todd: Well that’s almost a fourth argument it seems, which is something along the lines of people aren’t as epistemically humble as they should be, so you need to correct in that direction a bit more by being a bit more humble than intuitively you might think.

Robert Wiblin: Yeah. Then there’s this issue of if you’re just completely dogmatic about your own views and not interested in compromising with the rest of the world, then well, the rest of the world will hate you and potentially try to stop you, try to thwart you because they’re going to view you as kind of a hostile entity.

Robert Wiblin: Then there’s even another one, which is we want to create a culture internally among 80,000 Hours and the people who listen to our advice, and I guess the effective altruism community which we’re somewhat a part of, where people coordinate well and get along with one another and are friendly rather than backstabbing at every opportunity when they think that it’s advantageous to them.

Ben Todd: Yeah. Under pragmatic reasons, that’s one mechanism by which you get them.

Robert Wiblin: So yeah. Moral uncertainty, modesty and reasonableness seem pretty robust conclusions. Are there any philosophical schools that you think are at odds with these ideas that where it’s just clear that they wouldn’t be interested in our advice because they disagree too deeply?

Ben Todd: I think when people just debate morality in general you always focus on the interesting disagreements between positions and people don’t spend nearly as much time kind of talking about, “Well, we all agree that if you can save a life with a low cost to yourself, that’s a good thing to do. And if you can save two lives for similar costs, then that’s an even better thing to do.” But then that’s all you need to accept to think that you should take a lot of the kinds of things we’re going to cover super seriously. And yeah, it’s just because these big differences in impact aren’t intuitive, we tend to in practice not focus on them very much, but on an intellectual level, we can all agree they matter.

Robert Wiblin: I guess actually another topic that we haven’t taken a view on, and which doesn’t seem terribly decision relevant, is kind of demandingness, or whether it’s obligatory to help people or whether it’s just a good thing to do. And I guess there’s disagreement on the team and also it doesn’t seem to really affect what we should say or even really what people ought to do.

Ben Todd: No, I agree. But we’ll come later onto that question and actually, exactly how demanding you think morality is, turns out to not be quite as an important decision as at first it looks.

What are the most pressing problems to work on? [00:40:33]

Robert Wiblin: Okay. So that kind of tackles the really broad philosophy. Let’s narrow down to the next section, which is what are actually the most pressing problems to work on, which as we’ve said we think is a very important issue and potentially something that people don’t place as much weight on as they ought to. What do you talk about in that section?

Ben Todd: Yeah, so again we’re coming at this from a longtermist perspective and then the question is which issues are most crucial to address today in order to help the long-term future? And I mean this is really an emerging area of research, like I said, so we’re not clear what the answers are going to be, but kind of the general view that people are coalescing on is basically what you want to look for is these crucial moments in history which might be possible to change how the long-term future turns out.

Ben Todd: Phil Trammell has recently been calling these ‘hinge moments’. Though you should actually think of them more as ‘hinge periods’, because it’s not like a specific moment. Other people hate that nomenclature. Some people have suggested it should be called ‘pivotality’.

Robert Wiblin: I don’t like that. I think ‘critical junctures’ is one I’ve heard in history.

Ben Todd: Okay. Yeah. But the idea is most things probably just kind of wash out in the long term. We’re not really sure what persistent effects are going to have. Then it seems there are certain categories of things where there’s a kind of element of lock-in or irreversibility and there’s basically a point in history where something could really go one way or the other way and be like that for a very long time. That’s what you’re looking for in a hinge.

Ben Todd: So then from that there’s basically, we end up with two camps of longtermists. The first camp, I’m going to call them the ‘targeted existential risk’ camp. And they’re people who think, “Well, there are these things which look like they might be hinges.” Basically there are potential existential risks, so that’s something that could dramatically decrease the value of the future, indefinitely.

Ben Todd: A really clear example would be if an asteroid strike ended civilization, then obviously the value of the future is now zero from then on unless life evolves again on earth or something like that, but you know it’s at least a very long-term setback, probably. So yeah, if you think there might be some chance of an existential risk happening in our lifetime, that seems like it could easily be the crucial thing to work on from a longtermist perspective, so in how we prioritize causes and we think issues around AI safety, some of those might cause existential risks. Then secondly, global catastrophic biorisks. Thirdly, nuclear war. Forth, climate tail risks. Those are our kind of areas that we’ve selected within this category where it seems like maybe in the next couple of decades there’s important work to be done in reducing those risks.

Ben Todd: Then the second camp of people is quite a diverse camp, but basically in this camp, either you think, “Well, we don’t know what the hinge of history is yet. Maybe it’s those issues, but maybe there’s other issues that are even more important”. Or maybe you think you do know what the hinge of history is, but maybe it’s going to happen a very long time in the future. So you’ve got these two dimensions, you’re very uncertain, and the hinge might be a long way off. So there’s not really a name for this category yet. One name that’s being kind of used tongue in cheek is that that’s the boring longtermists and the boringness comes from the fact that you’re saying now is not an especially important moment in history. Now is just like any other moment, so we’ve no reason to think there’s a key hinge moment on our doorstep.

Ben Todd: Whereas the people who are really focused on AI, that’s more saying, “Well, the hinge might well be here. It’s AI.” So that’s where the name comes from. But you know, people also call it patient longtermism or maybe broad longtermism. Hopefully we’ll settle on a good name soon. So yeah, within that camp there’s then three main categories of problems that people work on, or issues that people work on. So if you’re just really not sure what the hinge of history is, then one option obviously is to just do more research and try and find that hinge. So that leads to global priorities research being a key area within that. And just in general, even another way to think about global priorities research being important is that we might also just be wrong about this whole longtermism thing. So for either reason you might think just doing more research into these big picture questions is really important. So, that’s the first one.

Ben Todd: The second one is you could do some kind of capacity building, so you could basically help grow the resources of people who care about the long-term future such that when this hinge moment does come up in the future we’re in a better position to deal with that. And that’s why we kind of see the effective altruism community, and building that, as really important. If we were convinced that it was, say, all about AI, or all about biorisk, then we would probably just focus on building the biorisk community or the AI safety community. But the point is that we might be wrong about what the key problem is, so by building this community of people who are interested in working on a wide range of issues and flexibly switching, we’re able to get there to be more people interested in working on whatever the most important problem turns out to be in the future.

Ben Todd: There’s also other ways you could build capacity. One option is just to try to save money and grow that as much as possible. If you think interest rates are kind of set by everyone else and everyone else is very impatient, they kind of want to get money to benefit themselves fairly soon, so that makes interest rates quite high. But if you’re just happy to wait and play the long game, you can just save your money and gradually compound it faster than the world economy grows and so thereby have more and more influence over time, and then in 1000 years your foundation will be able to have a big impact on the long-term future.

Robert Wiblin: Yeah, if it’s still around.

Ben Todd: Exactly. So it’s pretty unclear whether that’s a practical proposal then. There could be less extreme versions of that, that are sensible. You could also just kind of focus on building your own career capital or say building a network of people in policy who are interested in longtermism, so that maybe just in a 30 year time period or something like that, these ideas are much more represented among people who might be making the decisions that are relevant to these things.

Ben Todd: Yeah. Then the third category within boring longtermism is what gets called ‘broad interventions’. So these are just things to work on that will help society navigate lots of different future challenges, whether that’s future existential risks or other types of hinges. So the problem we’ve written about online in this category before is improving institutional decision making, where the idea is like many of these challenges will be navigated by big institutions in society and if you can help them be better at making complex decisions then we’ve got a better chance of overcoming these challenges, whatever they turn out to be.

Ben Todd: But yeah, this is also our greatest area of uncertainty about which problems we might recommend. In particular, a lot of people interested in boring longtermism think there’s going to be some other promising areas vaguely in the area of how to improve international coordination, how to improve international governance, how to reduce the risk of great power conflict. The basic idea there is just that a lot of these big global challenges really seem to boil down to coordination problems between major powers, and if you can just get those countries somehow working together better on issues like climate change but also like all the other issues we talk about, then you might make the world generally more robust to these challenges.

Ben Todd: So, yeah, I’ve been talking about the two camps: boring longtermism and targeted existential risk focused longtermism. I just want to clarify that this is actually a spectrum of options and everyone kind of agrees that both types of activities are important. The interesting question is more about exactly where the balance should be within the community between these two different things.

Robert Wiblin: And I suppose actually also what portfolio we want. Because presumably we don’t want to go all in on either one of these strategies. There are good reasons to have a bit of a range of different approaches both to hedge our bets, and also so we can learn about how well these different options are playing out.

Ben Todd: Well yeah, I mean I wouldn’t actually say it’s about hedging bets because that makes it sound like it’s about risk aversion. But each approach has some diminishing returns. So you want to do a little bit of each. And then as you said, it’s also about the exploration point where we’re uncertain about how good each approach is. So trying them out lets us learn that and then we can maybe reevaluate which ones to focus on later.

Robert Wiblin: You’ve done a really good job there mapping out the spectrum from people who think that we live in a special moment where you might be able to have an outsized influence, and we kind of know what the nature of the hinge is and how we might be able to influence it, to the other side, people who think the current time is unremarkable; there’ll be important moments in history when we’re going to affect the direction that things go, but they’ll be at some point in the future and we don’t know what they’ll look like so instead we have to kind of prepare for them in some general way or some much more flexible way.

Robert Wiblin: My impression is that you’ve become quite a bit more excited about the second category, the more boring longtermism lately. I’m a little bit more skeptical of that category and maybe think we don’t want to jump too quickly towards all those things. Because some of the challenges I think they face are, let’s say you’re trying to do a broad intervention where you improve decision making in the world as a whole. That’s difficult, even within one organization. There’s a million organizations in the world and it’s very hard to know that if you want to improve decision making, like who? Among where? You could end up having a lot of influence, but if it isn’t particularly the right people or the right places for what turns out to be the pivotal or the critical juncture moment in future, then it can all be wasted and that can potentially result in you dissipating your energy a lot.

Ben Todd: Yeah, I think I’d agree with that. I think within the second camp, the stuff I’m most excited about is more targeted community building work and global priorities research, so we actually recommend them on a slight tier above. And then things like improving institutional decision making in general, I’m a little bit less convinced that that’s the highest priority at the margin, though obviously, as always, if you had a particularly strong comparative advantage in that area, then I think it would be interesting.

Robert Wiblin: Yeah. I mean there are some exceptions to that where it does seem like there’s groups that we don’t know exactly how they’ll end up being pivotal, but there’s a good chance that they will be, like relations between the US and China. It’s pretty obvious that that going badly could really send humanity off track, so even if we don’t know exactly how they’re going to end up in conflict or how that could be avoided, building capacity to try to do that in the future, it seems like it’s a pretty decent bet.

Ben Todd: Yeah. Those broad interventions are our area of biggest uncertainty. We don’t really have strong, concrete suggestions right now, but it’s an area I’d most like to see more research done in to try to uncover maybe a more niche thing in there that turns out to be really good.

Robert Wiblin: It does seem like quite a few people are getting interested in that and starting to try to explore that space, so I’m hopeful that we’ll get a bit more clarity on it over the next couple of years. Especially, I guess people going into global priorities research might be able to bring some attention to it.

Ben Todd: I think partly we need more research into the area. I mean it’s also interesting to consider that it may be really helpful just to have people exploring these paths and trying to figure out good things to do within them. It’s quite a bold career move in a way to just try to explore a new problem area by yourself. But obviously if you succeed then you might uncover a thing that many other people in the community could go and do and have a big multiplier on your impact.

Robert Wiblin: A really difficult question here that I would love to know the answer to is how big are the differences in effectiveness of having an extra person or an extra million dollars going towards these different problem areas?

Ben Todd: Yeah. One problem rather than another.

Robert Wiblin: Yeah. As you were saying, I think people’s typical intuition is that working on the right problem might make you two or three times more effective than another one.

Ben Todd: Well, I would say the typical advice is just there’s no way to compare between different issues in the world. Just do what you’re most interested in, do what you’re most passionate about, and if you’re already passionate about recycling, work on that. If you’re really passionate about positively shaping the development of AI, then do that one, as many people are.

Robert Wiblin: Or a mix. Yeah. So maybe the typical advice is just don’t even consider it, but then I suppose there’s other people would think about it who might say, “Well, the differences are modest”. So I guess maybe they think these personal fit issues are more important.

Ben Todd: Exactly. And that’s why then following your interests, following your passion comes in. Because that’s, in theory, a proxy for personal fit.

Robert Wiblin: Yeah. Whereas I think we’re kind of trying to figure out if it is a 10X difference or a 100X difference, a 1000X difference, a 10,000X difference.

Ben Todd: I mean we’re pretty much totally the opposite to the common advice in a way, because we actually think probably your choice of issue is the most important question to get right.

Robert Wiblin: Yeah. And it’s a bit frustrating because it’s hard to prove this because we don’t really get the measurements of how much good did groups do, especially in this longtermist stuff where we actually won’t see whether that … well, we’ll probably never know even if we could see the fullness of history because we won’t know the counterfactual or we won’t know whether these things helped. But especially from where we stand today where they’re trying to have influence in the very long term, it’s impossible to really say with confidence the effect of any particular person or project right now. We kind of just have to have a model of the world and try to guess how much impact they’re having, or use broader theoretical considerations and try to go on that.

Robert Wiblin: But one reason that I think there probably are really massive differences in effectiveness between different problems is that we use this importance, tractability and neglectedness framework. We break down how effective it is to work on a problem into how big is the scale of the problem, how big would the benefit be if you could fix some fraction of it?

Ben Todd: Yeah. Though then for us, that actually boils down to just ‘how much does it help the long-term future’?

Robert Wiblin: Yeah, how much has it increased the expected value of a long-term future? Then, how many resources are going into working on this problem, where I think we have a pretty strong intuition that on average you get logarithmically declining returns to working on a problem. At least, once you’re talking about serious money or serious numbers that people going into working on them. So, that means each doubling of resources is probably equally useful. This is because you kind of run out of useful ways to solve a problem over time.

Ben Todd: And yeah, there’s quite a bit of empirical evidence that in many areas, you get these logarithmic returns. For instance I think with Moore’s law, they’ve actually roughly measured that the resources going into that type of research has kind of doubled consistently in order to maintain the constant growth rate.

Robert Wiblin: Yeah. Actually across scientific research in general, it seems like we’ve been just piling more and more researchers onto scientific questions and getting less and less return, probably because the questions are getting progressively more difficult.

Robert Wiblin: So then we’ve got the kind of third factor, tractability or solvability, where I guess we probably think that the differences there are somewhat smaller, maybe than other people think, or at least somewhat smaller than the other two factors. But it’s very hard to kind of measure that. So we’re perhaps a little bit unsure about it. But anyway, if you just naively plug in numbers for these things across lots of different problems, it seems like you get ludicrous differences. Like 10,000X differences between the problems that we recommend and the problems that many people work on.

Ben Todd: Yeah, maybe it’s just worth kind of working that through just very quickly and that taking this longtermist perspective with many issues, we’re just not really sure what effects they have on the long-term future. Whereas some of the things we’ve identified, such as if you can reduce an existential risk, that seems like it has a massive benefit for the long-term future. Although it’s hard to quantify, there seems to be some really big difference in scale there.

Ben Todd: And then when you look at neglectedness, again, it’s very hard to measure neglectedness precisely. One kind of guide is just looking at the amount of money flowing into different areas. So just to kind of look at that, US education has around $1 trillion a year spent on it, to an order of magnitude. Climate change has a couple of hundred billion dollars spent on it internationally. It seems that climate change is something like 10 times more neglected than US education as an area just measured with the total amount of resources going into it, but then issues like global catastrophic biorisks and AI safety research, they only have tens of millions of dollars per year spent on them. So I think that means then you’re looking at about a factor of 10,000 difference in the resources, at least the kind of direct resources going into them right now in the short term.

Robert Wiblin: And then I guess some people would want to say, “Well some of those things are so much harder to solve than others. It might be 10,000 times harder to solve”. But I suppose I don’t think that, and I think watching people try to go and pioneer these areas has vindicated that view that these things weren’t as hard to solve as people imagined because they do just seem to be getting traction as you would in kind of solving any other practical problem.

Ben Todd: Yeah, exactly. It doesn’t seem like it’s 10,000 times harder to make progress on AI safety than on climate change. In fact, actually it seems if anything, it’s easier because the field is so much smaller, which is exactly what you would expect from neglectedness.

Robert Wiblin: So if you just plug in these numbers naively, you get these very large seeming multiples, but there’s a whole bunch of factors which I guess kind of attenuate that conclusion or should make us doubt it. One is just that it’s a very strong and surprising conclusion on its face. So it kind of that extraordinary claims require extraordinary evidence, and here we’ve kind of offered theoretical considerations, these really broad measurements, so it would make sense to be reluctant to draw such a strong conclusion from evidence that doesn’t seem super compelling.

Ben Todd: There is a debate about whether we should even find it surprising, which is the issue of how strong should our priors be about how big these differences should be. And in one school of thought, it’s kind of just if you think there’s a factor of a hundred thousand difference, that means you think, say, one person working on the most effective area has the impact of 100,000 people working on more mainstream problem areas. And you may just think that seems implausible, that one person can easily have such an outsized impact on the world, so we should be skeptical of it.

Ben Todd: But then there’s another side of things where it’s just, well, very little effort is going into helping the long-term future because of our political system. General people’s interests are all focused on very short timescales. So we might just think that society’s not doing a very good job at helping improve the long-term future, and that would just mean that there might be amazing opportunities lying around that just aren’t being taken. So for someone who then really does care about helping the long-term future, they’re then able to have way more impact than normal from a longtermist point of view.

Ben Todd: This is called the issue of market efficiency for doing good; is it an efficient market, in which case it’s hard to have outsized impact, or is it just wildly inefficient in which case we shouldn’t be surprised that we find really big differences.

Robert Wiblin: Yeah, I guess if you just look around and see that nobody is trying to improve the long-term future in the way that we’re conceiving of it and then it does seem a lot less surprising that you get a 10,000 fold difference in the effectiveness of things that people are doing, because there aren’t people crowding into the areas where they could have a large impact because no one’s even thinking about it. But I guess there’s other reasons to be a little bit more skeptical. One is that you might get flow over benefits. So someone who’s working on a project that’s not the most important one in our view it might have kind of side effects that benefit the things that we think are most important.

Robert Wiblin: So for example, if a hundred people go and work on global poverty, they might kind of solve that problem to some extent, and this will cause someone else to decide not to work on that problem because it seems like it’s more handled and then go to work on a different issue.

Ben Todd: Once you get to this point where there are people who might work on either issue and they’re kind of unsure about which ones to do, then it seems like it starts to get quite hard to get massive differences if they’re influenced by everyone else. Maybe this happens on quite a large scale. Like Bill Gates has said he thinks AI safety is an important issue but he’s not working on it himself because he’s already chosen global health as his focus and he’s going to stick with that. But maybe if no one was doing anything on AI safety at all, Gates would switch into it.

Robert Wiblin: Yeah. Or alternatively you can imagine, because Gates is doing the global poverty stuff, that’s going to prompt some other billionaire to go and work on global catastrophic risks instead.

Ben Todd: Exactly. Yeah. And even if that just happens to a tiny, tiny degree, like 1 part in 1000, of the resources funge like that, then it kind of caps the difference in effectiveness between the two things at a factor of a thousand.

Robert Wiblin: Yeah. So funging is this slightly strange term of art that people use in effective altruism to refer to spillover, flow over effects when an action that you take changes the behavior of other people. A classic case would be if I donate money to a charity and then that uses up their room for funding, then another donor who would have supported them instead doesn’t give to that organization but gives to a different one. And in a sense, what my donation has actually caused is the funding of this second organization that this other donor then gave to as a result of my donation. So that can sometimes be a little bit hard to control: the effect of your actions because of how it changes other people’s actions, and you get the same kind of phenomenon potentially where if you take a job then it’s possible that someone else would have taken that job if you hadn’t taken it. So in fact what you’re causing is the spillover of the job that they go and take instead.

Ben Todd: Yeah. And we would say that that second person has funged you and the job that you took. So you can imagine the donation case. There’s two donors supporting a charity. The charity has a budget of $100. Each donor gives $50. If you were to donate even more, say you would donate $60 instead, but then have the other donor just reduce their donation. So the overall charity was still at $100, then that donor has funged your donation because they’ve kind of made it that you have less impact on the charity.

Robert Wiblin: Yeah, it comes from the term of fungibility, which describes objects that are all completely replaceable with one another. Just as a $1 bill is the same as any other dollar bill. But again, anyway, we’ll move on from that.

Ben Todd: Yeah, I guess there’s an Open Phil post where they talk about some of the challenges that GiveWell have faced with this.

Robert Wiblin: Another thing is even than resources funging, you can also have side effects in a project that have a positive effect on the long-term future even if it wasn’t designed for that. So you might think, say, working on US education in general is not terribly effective because lots of people are trying to do it and it seems relatively difficult to fix the problems that are remaining. But even if it’s relatively ineffective, improving, in general, wisdom among students making people smarter and more informed probably has some diffuse positive effects on the long-term future, which would kind of set, again, if you think it’s 1/100th as effective on the long-term future by that measure, then that sets an upper bound on the ratio of the effectiveness of working on that versus something that’s more targeted.

Ben Todd: Yeah, I mean maybe this is a bit of a digression, but lots of things potentially help improve the long-term future in a way it actually makes the scope of everyday actions much larger. So just having a kid, maybe that just increases the population, in expectation, by a slight fraction for a really long time. So you’re actually having this big positive impact. Though, that one would be more a speed up in character rather than a trajectory change, so probably is dominated by quite a large degree by reduction in existential risk. Whereas something like education that maybe does actually also reduce existential risk a bit.

Ben Todd: Though again, it wouldn’t be too surprising to me if it was a factor of 10,000 less levered because it’s just so undirected at the particular hinges that we think are most important. Yeah, most things probably do help the long-term future but you might still have very large differences between them, I think, on that argument.

Robert Wiblin: I mean, that’s a somewhat open question: “Do these actions generally help?” So you can imagine some people think that having more children is bad for the long-term future because it creates more problems in the short term before we can fix them. So even if it’s positive, because there’s more people in the short run, if it has a very, very slight effect on the probability of extinction or catastrophe, then that would end up swamping that.

Ben Todd: Yeah, exactly. I mean, I think you only need one kind of funging argument like the one we mentioned to already think that there’s a cap in the difference in potential effectiveness.

Robert Wiblin: We can also throw in empirical and moral uncertainty about these things, and it’s very hard to know exactly how to quantify that.

Ben Todd: Exactly, and in theory that should be taken into account in your assessment of scale. But then it’s very hard to do all these things in practice, so you probably want to correct for them later.

Robert Wiblin: Another thing is that it’s exceedingly hard to know how actually neglected problems are. So sometimes we say, “Oh, well only $10 million is being spent on this. So yeah, only $10 million has been spent on it so far.” But that can be very misleading for two reasons. One is that in the past, lots of money has been spent solving problems in general and very often some of that subtly indirectly works on this other problem.

Robert Wiblin: So, for example, not very much money specifically works on preventing aging. There’s relatively few labs that would call themselves anti-aging labs. But then everyone who’s been working on biomedical research in general, they’re going to make improvements in chemistry. Improving our understanding of biology has, to some extent, indirectly been helping us, bringing us closer to the point where we can tackle that problem directly, and it’s very hard to know how to quantify all that.

Robert Wiblin: So that’s in the past. Then in the future it also seems issues that are very pressing, really, really burning issues where the scale is really big and no one’s working on it, there’s a tendency, at least at the moment, or it seems there’s some tendency for resources to be directed towards those. So if it’s particularly neglected now, it may not be so neglected in the future. Then if you consider, on net, across all of time, is it neglected?

Ben Todd: Well, yeah. Well I guess what you actually care about is how many resources get spent on a hinge before that hinge happens.

Robert Wiblin: Right yeah, before that hinge, and so we don’t know when the hinge exactly will be.

Ben Todd: So then you’re trying to predict what resources are going to be spent on that thing in the future as well. So yeah, this all means that just those naive numbers that I said earlier, that’s probably not the correct way of thinking about it. Probably the differences are less severe than the kind of very direct numbers that I mentioned. But then they were such large differences in neglectedness on the naive measure, say a factor of 10,000, and still expect that to be large differences even once we take into account these attenuating factors. I think there’s also just a very strong argument that we should expect helping future generations to be neglected, as I mentioned earlier.

Robert Wiblin: Because I guess the market doesn’t really reward it very well. Voters don’t really reward it for the government, and we can just directly observe that not that many people are thinking about it.

Ben Todd: There’s an old Paul Christiano blog post on our website where he talks about, if most people only care about a thing 1% as much as you care about it, you should expect to have roughly a hundred times as much impact by focusing on it.

Robert Wiblin: Yeah, I guess another thing is that we’re often focused on kind of tail risks and unlikely events, and for those, well it seems like humans are very bad at thinking about those because they weren’t so important in the environment in which we evolved, and it’s just also just very difficult for any mind to really wrap their heads around.

Ben Todd: Yeah. Though I’ve never been quite convinced by the heuristics and biases arguments for why X-risks should be neglected, because we know in other cases people like–

Robert Wiblin:: Like terrorism.

Ben Todd: … way overreact to small probabilities of bad things happening. So I think you have to make a more detailed argument about the nature of the risks.

Robert Wiblin:: Yeah, I think that’s right. I think it’s just the case that with unlikely things, we’re just all over the place. It’s not obvious that we spend too much or too little, it’s just that we–

Ben Todd: Sometimes we spend way too much, sometimes way too little.

Robert Wiblin:: Exactly. Yeah.

Ben Todd: So just an argument against efficiency.

Robert Wiblin:: Yeah. And so it’s an argument that if you thought about it a lot, and really tried to estimate the likelihood of these things concretely, then maybe you could get a pretty big edge there.

Ben Todd: Yeah, that makes sense. So then the question is, “What is our overall thoughts on how big the differences in scale are”? We tried to survey different leaders of organizations in the effective altruism community last year, and we asked them just at the current margin, if you had $1 to donate, how would you trade off between donations to the EA Long Term Future Fund and the EA Global Health Fund? And the median response to that was people thought the long term fund was about a factor of 20X more effective, though with huge differences in opinions around them. An order of magnitude at each side of that. Yeah. And so then that suggests that they thought on average, was about a factor of 20 difference. I should say that those groups of people were probably selected for being more interested in longtermism. So that maybe has a bias in favor of the longtermist direction.

Ben Todd: But then yeah, my guess is that many… We didn’t actually ask this question, but I would guess that, say, if you just picked a random median US focused charity, compared it to the EA Global Health Fund, people would think that the Global Health Fund was about a factor of a hundred times more effective, which is kind of based on the beneficiaries being a hundred times poorer. Also, I think GiveWell now says that they think their top charities tend to have a five or 10x multiplier compared to just giving the cash, and so that maybe gets you over to a factor of a hundred difference. And so then if we put those two super speculative figures together, you’d roughly get to a factor of 2,000 difference between a median US charity and the EA long term fund, and that would be a rough estimate of the total spread.

Ben Todd: Though, as I say, that’s very controversial. And that’s maybe one of the biggest uncertainties within our ranking, is just how does say, nuclear security compare to global catastrophic biorisks? Is it that actually there’s only a slight difference between them, or maybe is biorisk actually 10x or even 100x more effective?

Ben Todd: Either way. Yeah. Even if there was only a factor of 100 difference in effectiveness between issues, that would be an absolutely crucial consideration for choosing a career, because again, it would be like in one year, you could achieve as much as a hundred years of work in this other area, at least if you kind of treat all those years as equally effective. And so it’s really, really important to take that seriously.

Ben Todd: As we’ll come onto, your personal fit is also really important and can vary a lot between different areas, and so sometimes personal fit can outweigh the differences in different areas, and it could be better to kind of do an area that you think is a bit less important in general, at the margin, but where you have a really unusual personal fit with it. But I think we also see people who kind of intellectually agree that maybe there are these large differences, but then when it comes to actually making decisions, they kind of treat all these areas as roughly similar, which is how our intuitions work. Our intuitions do not tackle these massive differences in effectiveness very well at all. And so you really have to try to resist that tendency, and it can help to try and put specific numbers on things, and try and think it through.

Robert Wiblin:: All right. Is there anything that we haven’t covered already, before you move on from this issue of what are the world’s most pressing problems?

Ben Todd: Yeah, so we’ve covered our view of which problems are most pressing. And we’ve been really explaining how that’s driven by longtermism, and the other philosophical positions we mentioned earlier. But I would just say that I think actually many people with slightly different focus areas could get behind many of these issues, even if they’re not, say, as focused on longtermism as us.

Ben Todd: I think a wide variety of moral positions can agree that nuclear war would be very bad. And also, actually, nuclear security seems pretty neglected still compared to many more conventional issues. And so we talk about climate change way more than we talk about the risk of nuclear war, even though we could literally all go up in flames at any moment. It’s actually a very big problem. Yeah. So you don’t need to be into this very full on longtermism to think maybe that could be one of the most pressing problems to work on.

Ben Todd: Likewise, the whole idea of global priorities research and building the effective altruism community is that we are very unsure about which problem is most pressing. And so there’s room for people with all kinds of different values and backgrounds to work together on that project of figuring out which problems are most pressing, building a community that just wants to work on whichever issues turn out to be most important, based on the arguments.

Robert Wiblin:: Yeah. Every so often we kind of do a back of the envelope calculation of how much would it cost to save a life by trying to prevent nuclear war, or doing other things to reduce catastrophic risks, that is of the present, of people alive now. And it seems it’s kind of competitive with other opportunities to save lives through really cheap healthcare and things like that. But it’s very, very speculative, obviously.

Ben Todd: Yeah. Probably depends a lot on the specific existential risk as well.

Robert Wiblin:: Yeah. Like whether you can get a lot of leverage there.

Which careers effectively contribute to solving these problems? [01:11:10]

Robert Wiblin:: All right, let’s push on. So we’ve talked about what problems we might suggest that people work on solving. The next step might be to think about what are the methods that you can use to get a lot of leverage in the world, to have a lot of influence, in order to solve a large fraction of these problems. Yeah. What do we talk about in that section?

Ben Todd: Yeah, so the next section is now getting into concretely which careers could you actually take to help with these issues? Supposing you roughly share our view of global priorities. You at least, you’d want to consider working on some of these issues. Then what might you actually do to to help with them?

Ben Todd: And I should just kind of say as a way of caveat, people really focus a lot on our particular ranking of things, but these are really just ways to help you brainstorm things that might be high impact. Then you need to then run it through our full decision process. Consider your individual circumstances, your personal fit, other options that might be really promising but aren’t on our list, and to actually come to your individual decision. We’re not providing a kind of, all considered, ‘these are the best careers for everyone’ list. That’s unfortunately not possible.

Robert Wiblin:: Yeah. You’re all too different. If only you were all identical widgets, then we could do that, but sadly.

Ben Todd: So yeah, these are just ways to brainstorm ideas and so we start with what we call our five categories. These are kind of broad, very broad categories that can help you generate options. I’m not going to go through them all in detail, but kind of add a few, I think, more mistakes people make when they’re thinking about these areas.

Ben Todd: So the first one is just trying to find a really good nonprofit that’s addressing one of these problems, and working at that. Ideally, a nonprofit that is a bit talent constrained as well, rather than just only very heavily funding constrained. And yeah, we list a lot of… We’re trying to list more and more organizations on our job board.

Ben Todd: I think one thing here is that it’s very easy to focus on what gets called the ‘effective altruism organizations’, which are the ones that have an official EA stamp on them, but actually there’s many, many nonprofits addressing some of these issues, and they can all be potentially good to work at as well, so we’d encourage people to think a bit broader about options within that category. And we’re trying to make the list of jobs on the jobs board broader, and we’re now up to almost 300 listed positions, and we’re pushing to get that wider still in the coming months.

Ben Todd: The second one is trying to do some kind of relevant graduate study, and then go into some work on a research question that’s relevant to one of the problems we’ve covered. Going down the grad school route is good if you might have good fit for it, though you’ll want to be quite careful about which area you choose to go into. And so some of the areas that we think are the most generally exciting are machine learning and economics, because they’re both… Economics is very relevant to global priorities research, as well as other areas. Machine learning is obviously very relevant to AI safety.

Ben Todd: But then another nice thing about graduate studies in those areas is that they also give you a lot of general backup options. You’re not just punting your career on this one path. But then yeah, lots of other areas that are worth considering: bioengineering, synthetic biology, all the kind of stuff in that area, because of its focus with biorisks. And then there’s all kinds of policy relevant issues that can be very useful for policy careers. There’s security studies, even law, and then go into policy, international relations, again, economics, but more with that focus. Yeah, war studies it gets called at King’s in London. Yeah, master’s of public policy. And so you can do all this research that’s relevant to either the broad longtermist interventions, or to take a policy approach to AI safety, or to biorisk. So yeah, those are some of the areas that we’ve been most excited about recently within research.

Ben Todd: The third one then maybe is the one I want to highlight the most, which is government and policy careers in roles that are relevant to these different problems we’ve covered. And I think this is the area that maybe is still, even though we’ve been banging on about it for awhile, the most neglected by our readers, and one which I think really a lot more people could seriously consider maybe making applications to than currently do. It also seems like one of the biggest needs in the community of people addressing these issues, and where there’s a lot of roles that might be very high impact.

Ben Todd: I think partly one problem that stops more people going into these areas is that it seems very vague on what to do in this area. But one thing is just, in order to get started, there’s actually a couple of very concrete ways to get started. One is think tank internships and research positions, and we’ve got a profile about those.

Ben Todd: In the US, you’ve got joining a political campaign, and you can often just do that pretty much straight from undergraduate. There’s trying to work as a congressional staffer, and then there’s doing graduate study in a relevant area. In the US as well, it’s pretty common to do a master’s of public policy and then go into policy from that. Also do a law degree. And so I call these the kind of entry policy positions, and they are just things that kind of generally establish you, get you a network, and then you can go from them into many different positions. Oh, and the other one is you can just try to get directly into the executive branch, or the White House, and just get positions in those straight away. Sometimes you can kind of leap ahead more quickly through these accelerator programs. So those are kind of how you’d establish yourself.

Ben Todd: Another big misconception we come across is people think you have to be a real people person in order to work in policy. But actually, there’s just a really wide range of roles in this area, and there’s definitely a lot of roles for more research oriented people, who aren’t just people with really good social skills.

Robert Wiblin:: It’s not all kissing babies. Some of them are just shuffling paper.

Ben Todd: There are obviously lots of roles where it is really important to have good social skills, but you don’t necessarily have to.

Ben Todd: Another big thing is people don’t realize that you can just enter a lot of these roles directly from undergraduate. You can just try to make a round of applications initially, and then if that doesn’t work, then you can go to graduate school, or gain some other type of career capital and try again later. But then yeah, you don’t need to have a policy degree. We sometimes find people who have already got a PhD, but then they think they need to do a master’s of public policy in order to work in policy. But actually, you’ve got a PhD, you can just go straight into policy positions. In fact you don’t even need the PhD.

Ben Todd: And so yeah, I think maybe one thing we were speculating which is going on here is just after you’ve just done six or seven years doing this very narrow PhD thing, it’s very hard to imagine that someone would just let you loose and set policy, and figuring that out, but that is actually how the world goes sometimes.

Robert Wiblin:: It is odd, isn’t it?

Ben Todd: One other caveat about policy is it is probably easier than average to accidentally make things worse in policy. And we’re going to come onto doing accidental harm later in the episode.

Ben Todd: Okay, so fourth path?

Robert Wiblin:: Yeah. So then we’ve got direct work, research, policy and government.

Ben Todd: Yeah, yeah.

Robert Wiblin:: Fourth one.

Ben Todd: The fourth one is just if you already have a strong skill, then obviously consider how you might be able to apply that to one of the most pressing problems. And that’s just a bit of a catchall. Our categories are most relevant to people who are early in their career, and they have a lot of flexibility over what they end up doing. If you already have some very well developed skill, it’s harder for us to give general advice, but there may well be a great way to apply that to one of these issues. And so if you’re in that category, unfortunately you’re going to have to do a bit more work to go and meet people in the problem area that you’re focused on, and try and ask them what you could do with the skills that you have.

Robert Wiblin:: I think during the Ebola crisis it turned out that anthropologists were really important because they understood burial practices and how you might be able to shift burial practices to reduce the transmission of Ebola. It’s just pretty random, and you wouldn’t necessarily think anthropology is going to be the top thing to study.

Ben Todd:: To help with biorisk, yeah.

Robert Wiblin:: Yeah. To help with biorisk. But as it turned out, sometimes just having an expert in it. Maybe you want to have a few around in all kinds of different areas, because you don’t know what the future holds.

Ben Todd: Yeah. That’s a nice example of how a wide variety of skills can be applied to these pressing problems. And it’s also an illustration of how people interested in longtermism, and just effective altruism more broadly, we really need there to be people who share that kind of approach, but also really understand all the different academic disciplines that ultimately might be applicable. I mean, maybe medieval poetry or something, it’s a bit harder to see how that’s going to turn out to be the crucial thing, but —

Robert Wiblin:: Yeah, the future’s unpredictable, but not quite that unpredictable.

Ben Todd: Yeah, but then there’s still a very wide range of things that are needed. And one thing that has, I think, been on some people’s mind at the Global Priorities Institute is that there’s very few people who have a strong history background who are interested in effective altruism. But actually when we try and study longtermism, often you really want to be looking at historical analogs. There’s questions like, “Could the Industrial Revolution have gone differently”? There’s questions about, “Has welfare gone up or down over time”? And these are basically questions in history that are really interesting and useful, but we haven’t found anyone to study them.

Ben Todd: And so that’s just one example of an area that’s not normally considered a classic thing that we want lots of people to do. But it’d be really great if there were some people doing that path.

Robert Wiblin:: Yeah. So what about earning to give?

Ben Todd: Yes. So then we have that as the fifth category. And so I like to think of earning to give as kind of deliberately earning more than you would have, and then using that to donate more. So you don’t necessarily have to be in a super high earning career, anyone can earn to give to some degree by choosing a higher earning option, and donating it. And yeah, the key thing here is just money is always useful to help the most pressing problems.

Ben Todd: Even if you’re not sure how to use money right now, you can always save money and grow it and have more later. And then when opportunities do arise, you can use it. But actually, I think there are just great places for donors to give right now. If you’re not sure about doing your own research, you can donate to the EA Longterm Fund, or one of the other funds. And so yeah, I mean, I think earning to give is still an interesting option, because it’s a way for anyone to convert their skills into resources that are working on the world’s most pressing problems. Yeah. If you’re going to follow that option, then I’d say also try to do it in a path that gets you good career capital, so that you do also have the option of switching later.

Ben Todd: Yeah. One of the reasons why we highlight it a bit less than we did in the past is that over the last say five years, there has been a lot more money going into some of these longtermist focus areas, especially from Open Philanthropy. And that has created a bit of a kind of overhang, where it seems like a really key bottleneck is basically people able to do research, entrepreneurship, policy, within these pressing areas, and slightly less so, money. Though again, having said that, there are still uses for additional money. For instance, Open Philanthropy often only wants to fund about half of the budget of an organization, because they don’t want that org to have to only depend on them. And so that means that you at least need people to match Open Philanthropy one-to-one.

Robert Wiblin:: I guess other comments with earning to give is that if this idea of saving and investing money for a very long time pans out as being a really good method, at least at some point that we should switch to doing that, then earning to give is going to look a lot better.

Ben Todd: Yeah exactly. It’s quite hard to compound human capital. You can kind of do it by growing a community, and then hoping that the community will kind of carry forward the values after you die. But generally, most people’s career capital drops to zero at the end, whereas money can be saved and used after your death. And so if you think the key hinge moment is going to be after your death, then that could be another argument for earning to give. Though we’re getting to pretty speculative territory at this point.

Robert Wiblin:: I guess also, we just don’t know what the future will hold. It seems at the moment these problems are … Well, they’re a lot less funding constrained than there were five years ago. And it’s possible that in five years’ time, money will look more important than it does right now. We could be in kind of a temporary period where our money is less the bottleneck.

Ben Todd: I mean, yeah. Just the other point is, with all of these things, you have to consider your personal fit, and so if you’re an unusually good fit with earning to give compared to your other options, then that could well be the top thing for you.

Robert Wiblin:: Another one that kind of didn’t make this list, and it’s a bit conspicuous, is advocacy. What were the reasons for that?

Ben Todd: Yeah. Well, we should maybe say what we mean by advocacy. In a career context, I mean trying to get into jobs where you have a public platform of some kind, such as being a journalist, or working at a campaigning organization, and then using that to promote important ideas. I think that can be very high impact, but it’s a bit harder to make a general recommendation like that. Partly just because it’s very difficult to actually do that. And then another reason is just, I think advocacy is one of the areas where it’s easiest to accidentally cause harm, and that’s because once an idea is out there, it’s very hard to unwind it. And first impressions are really important. And so it’s very easy to do a bunch of advocacy, but it’s a bit worse than it could have been done otherwise. And then you’ve actually set the cause back. So we didn’t want to just make that a very general recommendation to people.

Robert Wiblin:: Yeah, I think other reasons are that it seems especially hard to figure out what the call to action is in some of these problems that we care a lot about. If we’re worried about risks from new technologies, it’s like, what is the mass campaign that you’re running to try and get people to take action? We don’t yet know what we want a large number of people to do. So it seems if you’re going to be doing advocacy, you have to be fairly targeted.

Ben Todd: Yeah, and that’s a big change in our advice. Like if you’re focused on global health, say, it seems much more obvious that you want people to be in favor of international aid. Probably if lots of people, millions of people donate 1% of their income, that’s a really major way of helping with that issue. And so, having people earn-to-give and do advocacy and mass outreach seems much more useful, whereas it’s much harder to see exactly how mass outreach is the key bottleneck for something like AI safety right now. It’s more a question of just figuring out what the hell to do at all.

Robert Wiblin:: Yeah. I guess it’s a bit ironic for us to say that given what we’re doing exactly right now, which is something that’s reasonably like advocacy.

Ben Todd: Well, I don’t know if this is mass advocacy: five hour long podcasts about philosophy and…

Robert Wiblin:: Yeah, maybe a little bit more niche.

Ben Todd: Yeah, if you listen to hundreds of hours of podcast content, you might be able to work out what to do with your career differently. It’s not exactly that we just want everyone to donate a bit to the Against Malaria Foundation. It’s much more complicated to work out.

Robert Wiblin:: Yeah. So you raised the issue of the conclusion about what method you want to use varies a lot, depending on what problem you might be focused on. So if you’re focused on poverty, then earning to give maybe looks much better, or mass advocacy does. Or earning to give, because there’s obvious things you can do that seem really effective that can just absorb tons of funding. Even if it’s just cash transfers to people in dire poverty.

Ben Todd: The top GiveWell charities also have… Not sure exactly what the funding gap is now, but it’s around 100 million, maybe more. And they think that they could find a bunch more if they did more research, which is what they’ve just started doing.

Robert Wiblin:: And then if you’re trying to eliminate factory farming, I guess probably I actually think that the best thing to do is probably going into these clean meat, or plant-based meat businesses, or become a scientific researcher focused on how to make those products really good. But there’s also obviously, opportunities in kind of mass moral advocacy to try and get people to change their values, and maybe worry about that issue more, and that the message there seems a lot less fragile.

Ben Todd: Yeah. Torturing pigs is bad. Don’t do it.

Robert Wiblin:: Okay. So that’s the discussion of broader methods. Let’s get more concrete and talk about specific paths that people can take to use these methods to do good.

Ben Todd: Well, yeah, I mean I would frame it slightly differently, where the five categories we’ve just considered, the idea is that those are very broad, and hopefully everyone can generate some options that are in those five categories. Then we have secondly, what we call the priority paths, and these are some of the areas that we’re most excited about. But then they also tend to be the most difficult ones to take. Basically the message there is, if you might be a good fit for one of these priority paths, then it’s often one of the top things to consider. Though, again, even then you still should compare it to other options.

Ben Todd: So yeah, we have this list of about 10 paths. I’m not going to recap them all now. Just maybe some comments about how our views have changed on these over the last few years.

Ben Todd: Right now, we actually list at the top… They’re not really ranked, but we do have at the very top AI policy, and that’s an area that I’d say we’ve doubled down on in the last year or so. Recently, we released a big guide to US AI policy careers. I think there’s still a lot of room for more people to go into a pretty wide range of roles. And that’s all laid out in that guide.

Ben Todd: An area that we rank a little bit lower down the list than one or two years ago is operations management. In that I think there’s actually been a big success there, where we released a bunch of content about operations management last year, and people actually really responded to that, and a bunch of people went and took jobs in these positions, such as Habiba, who I mentioned at the very start.

Ben Todd: We actually also noticed that it was easier for people to come from outside the community and fill these positions than we had thought. So we have a great new director of operations at CEA called Josh Axford, who was working at a tech startup in an operations role and went in. Open Philanthropy actually hired the head of operations, the COO of the Clinton campaign. So, pretty high powered hire there.

Ben Todd: So there’s less open positions than there was a year ago, so it seems a little bit less pressing than then. Having said that, we still have it on the list, and there’s definitely going to be a long-term need for people in this path. If you think in general, for every kind of five or 10 staff you have, you need at least one operations management person. So that just has to scale up with the rest of the community, otherwise it’s going to be a big bottleneck. And yeah, we still often see people kind of a bit prejudiced against this path, where sometimes we come across people who think that they’re just kind of generally smart, so they must be able to go into a management position within operations management right away. But actually, it’s a very challenging skill in and of itself, and you generally need to have a bit of experience, and learn that particular skill. And yeah, we really need, in the long term, people to specialize in this kind of general function.

Robert Wiblin:: Yeah. That raises this general problem, that kind of category of roles that’s bleeding for people, it’s desperate for people right now – might not be that way in a couple of years time, or in 10 or 20 years time. And that’s just extremely difficult for people, including us, to anticipate exactly how that’s going to go.

Ben Todd: Yeah. We do try to estimate the room for talent in different paths. But yeah, I mean it’s very hard to communicate. We think it’s a bit less pressing than last year, but still very, very pressing.

Robert Wiblin:: Another one is, oh, it’s really pressing now. We think it won’t be as much in the future, but only because lots of you will read this and take the advice. So it becomes a little bit circular.

Ben Todd: Well there’s that whole problem as well, yeah.

Ben Todd: But yeah, just wanted to flag that some of the things in that old article we wrote about operations management aren’t quite as current, but we still list it in the priority paths.

Robert Wiblin:: Yeah. Any other priority paths you want to comment on specifically?

Ben Todd: Well, I think the other most interesting category to discuss would be other things we’ve considered adding, but haven’t written up yet. So there’s a number of things here. I think one interesting one is something like nonprofit entrepreneurship. And I definitely think that’s a big bottleneck facing the community, and that there is funding around to build more organizations, but we lack the entrepreneurs able to actually develop the concept, and make it into something shovel ready, and then raise funding for it.

Ben Todd: One reason why we haven’t listed it is just because it seems it’s very difficult, and there’s not many people who have both entrepreneurial experience and a lot of background in the specific problem area. I think it’s very hard to just go into one of these complex areas like AI safety, and just start a new organization, that’s often a bad idea. You need to instead build up your experience within that particular problem area for several years first, and then found the organization later.

Ben Todd: But if you’re listening to that, and you might be one of those people, you’ve worked in one of these problem areas for a while and you’re interested in starting something, then we’d very likely be interested to talk to you. And so I’d encourage you to apply to advising. It’s very difficult to give general advice, because it just depends so much on the specific idea, whether it’s worth doing or not. But we’d be keen to try and help you make introductions, or help you find funding for that kind of option. So yeah, that’s entrepreneurship.

Ben Todd: Another area of priority paths might just be if we do end up having international coordination, or reducing great power conflict as a problem area, then there might be some priority paths associated with that, that will then just turn out on whatever seems to be the most promising ways to work on those issues. Maybe that’s just going to work in research around international governance, or maybe working at international organizations, or working in the foreign service, trying to improve relations between different countries. So yeah, like I say, that’s very uncertain at the minute, what might be there, but we’re definitely thinking about it.

Robert Wiblin:: We’ve also considered adding a catchall extra option to this, which is just some other way of dealing with our priority problems that we aren’t going to be able to think of, because it’s very specific to you, but does seem really, really valuable.

Ben Todd: Yeah, totally. I mean, like I was saying, people tend to treat the list as comprehensive, but it’s really just a list of ideas to consider. And so maybe we need to make it really, really salient that, yeah, there’s actually a couple of different categories of things here. One category of things is a thing that might be a priority path, but we just haven’t had time to investigate yet. And if you think you found something like that, you could consider just kind of trying to pioneer that area, then telling us about it. And if we’re convinced, we’ll be able to have a podcast about it, or write about it, and maybe get more people doing it. So it’s kind of a high risk thing to do, but maybe potentially much more high impact.

Ben Todd: Another category is just… There’s certain things that maybe it’d be really great if just a couple of people did them. So, with priority paths, we tend to only want to recommend things that quite a lot of people could do, because otherwise our advice would maybe immediately become obsolete. But something like we talked about earlier was being a global priorities researcher, but who’s a historian, seems like it’d be great to have a couple of those, but maybe it’s not a kind of general enough recommendation that we’d want to make it a priority path. Yeah.

Ben Todd: Another category is just maybe you have a very strong comparative advantage in something which normally wouldn’t be a good idea to do. That’s more like the magician example, where we’re never going to make a magician a priority path, but if you are unusually talented in that area, or successful in that area, then it could be worth doing.

Robert Wiblin:: Yeah. Maybe you could find a way to spin that into something really useful later on. Just through your general successfulness.

Ben Todd: Exactly. That gets into career capital, really.

Ben Todd: Yeah. I mean, then the other area is other ways to tackle AI. And so recently, there was a really interesting post on the EA forum by Claire Zabel and Luke Muehlhauser about information security as a key need. Yeah, we should put up a link to that. We were considering making that a priority path.

Ben Todd: Another thing has been working in AI hardware. Some people think that AI hardware might turn out to be the biggest bottleneck, and just having people who really, really understand that deeply, and work in the relevant companies, might be super important.

Ben Todd: And then also, there was a talk recently at EAG by Vishal at DeepMind, arguing that there’s a real need for people to do kind of nontechnical roles at the top AI organizations, such as operations, but also things like communications and assistant roles. And although you’re not working directly on AI technical safety research, there might be many, many roles for people where it’s still great if they really care about longtermism, and may be able to use their roles to help.

Robert Wiblin:: So what do we have to say in terms of priority paths, to people who are kind of already more mid career, someone who’s listened to the show, but is more like 40, and might find it hard to envisage switching into these paths wholesale?

Ben Todd: Yeah. So we do have very rough article that we link to on the key idea series, about if you already have well developed skills in an area, then what might be the highest impact thing for you. It’s very hard to give general advice. But it would come back to: try and meet people in the problem area, and ask them about how you might be able to apply your skills to help the area. I think that’s basically what that comes down to.

Ben Todd: There’s lots of areas where it would be great to have more experienced people in the community, I think. Management skills has been a thing that’s been very highlighted in our most recent survey of skill needs. Another thing is anyone with experience of government and policy could be really, really useful in particular.

Ben Todd: The other thing I wanted to end on is if none of these options really seem like a good fit for you, then what do you do? How do you come up with even more options? And we do try to link to a kind of general brainstorming process that you can work through to help come up with more options yourself. Also, lots of brainstorming questions to kind of act as a checklist to think through.

Ben Todd: And then the kind of other category of things is you can always, if you’re not sure how to contribute right now, just focus on building generally useful career capital, and then in a few years, maybe an opportunity will come up. You can also contribute by donating. And maybe you can also contribute by doing a little bit of outreach to your friends. Maybe you don’t have lots of knowledge of AI, but maybe you know someone who does work in that industry, and you could actually have an impact by doing that. That was slightly covered in the Cass Sunstein podcast that we had recently, where he pointed out that by helping to subtly shift social norms, some people can just have a really big impact. And so maybe even just by kind of mentioning that you think we should care more about future generations, does actually maybe have more impact than you might think, because it could help move us towards a tipping point where that’s just a very mainstream idea in the future.

Robert Wiblin:: Yeah. Do you want to comment at all on how people can try to figure out what is the key bottleneck in the problem that they’re focused on? It typically seems that people in these different problem areas often have pretty strong intuitions about what’s really the rate-limiting step for them.

Ben Todd: I mean, I think it is difficult to have general rules of thumb. It often just comes down to trying to understand that area, and think about what things are most delaying progress at the time. I think we do try to sketch some general rules of thumb in the high impact careers article. And so one thing is just if a problem is a very new and niche area, then probably one of the top bottlenecks is just getting some kind of proof of concept, that there’s actually useful stuff to do in this area. And so recently technical AI safety has gotten to this point where there have been actual papers trying to just solve technical challenges in AI safety, and that’s kind of shown that that’s a thing, and now more researchers are getting mobilized behind it.

Ben Todd: Whereas, AI strategy, which is kind of all the other messy political and social issues around AI, that’s at a more early stage where there’s kind of not very clearly defined issues at all, or what to work on. And so it seems there the key bottleneck is something more like it’s… It’s been called disentanglement research in one post, and that’s just defining the key questions in the area in the first place.

Ben Todd: Once an area is more established, then often the bottleneck just becomes kind of a more mass thing. You can kind of think with ending malaria, there’s a logistics bottleneck of actually getting these solutions rolled out on the ground, and then there’s a funding bottleneck, and it’s more just a question of getting more and more resources going into those two big bottlenecks.

Career strategy

Robert Wiblin: Alright, let’s move on from talking about the priority path to career strategy, which I guess under the heading we have a whole lot of different considerations like career capital, how much to explore personal fit and how to weigh that against other things as well as avoiding harm and I guess weighing your career against personal things like how much do you enjoy it and the sacrifices that you’re willing to make and not willing to make.

Career capital [01:37:47]

Robert Wiblin: Maybe the first and the most significant one is career capital. What do we mean by career capital and what’s our overall take on it?

Benjamin Todd: Yeah. So in this section we cover all these strategic considerations that are cross-cutting and determine which options are best at different times in your career. And one of those is career capital. By that we mean the skills, connections, credentials that you get from a job which puts you in a better position to make an impact in the future. One common misunderstanding here is by career capital we mean just the credentials that you get. So having a degree or a formal qualification. But actually, we mean anything that puts you in a better position. So that’s the people you know, it’s the skills you’ve learned, personal developments, your productivity habits. All of those things can go under that banner.

Robert Wiblin: Yeah, and I guess how important do we think it is as against having an impact right away, or other factors that might cause you to choose a job?

Benjamin Todd: Yeah, so kind of in abstract, it seems like it could be really important. One way of seeing that is just for most people in their first couple of jobs, especially the first couple of years of work, they probably just don’t have a particularly big impact. We sometimes see people who want to just have a huge impact right away, but normally your impact is going to increase significantly in your first couple of jobs. And kind of taking it from another angle, we can see that if you look over a very long term, it seems people get much more productive through their careers. And there’s a bunch of different strands of evidence for this. One is studies of expert performers. So top researchers typically find that they, depending on the fields, hit their peak productivity often at age 40 to 50.

Benjamin Todd: And so that suggests that they’ve been increasing their output for several decades and presumably a top researcher who’s at the top of their game is having much more output than that same person was as a graduate student. Another strand of evidence is people’s income typically more than doubles over the course of their career. Or kind of more turning to the effective altruism community, a third strand of evidence is in our surveys when we asked people at organizations in the community to place a dollar value on different hires they’d made, for senior hires compared to junior hires, they often gave figures that were about three times higher. But often there’s not that many years of experience separating those two positions, which kind of suggests that people have managed to triple their impact in that early part of their career.

Robert Wiblin: Yeah. I guess another way of seeing it is just that there’s some really valuable positions, say being a senior academic or you know, being an elected politician, which you just can’t get when you’re in your early twenties, almost never. You need to do things like get a PhD to even have the possibility of getting those kinds of high impact roles.

Benjamin Todd: Yeah, so that’s another way to see that career capital can be important is that sometimes doing a certain thing just opens up an option that just wasn’t an option before such as an academic career just isn’t really an option if you don’t have a PhD and so if you think that’s really good, then you can kind of clearly increase your impact significantly if you thought that option was significantly better than what you could’ve done otherwise.

Robert Wiblin: Yeah. I guess another thing is that even if you’re only, in some fundamental sense, a couple of times more productive… If you’re a couple of times more productive, you might get access to 10 or a hundred or a thousand times as many resources or be able to manage many, many more people than you would otherwise cause your position in terms of your ability relative to others is going to be much higher and so that will allow you to amplify the gains in your actual kind of skill level to have even more impact.

Benjamin Todd: Yeah. That could be right. Another point is that career capital is more important the earlier you are in your career. That’s for at least two reasons. One is that there’s probably some diminishing returns from career capital. So as you gain more it becomes less useful, though actually I guess Rob’s argument just there was almost an argument for increasing returns from career capital, so that’s a little bit unclear. I think the point that’s more clear is just the earlier you get career capital, the longer you have to use it. And so obviously learning a new skill, age 64 and then you only have one more year to use that skill is much less useful than using learning at the start of your career and then having 40 years to use it. So for that reason, it means that younger people should focus significantly more on career capital than late career people.

Robert Wiblin: Yeah. We said earlier that there are times when people do have their greatest impact when they’re in their twenties or relatively early in their career if they’re in the right place at the right time and just have an opportunity to do something that other people haven’t seen, for example, I guess famous startups are examples of that. So we just don’t want people to ignore that possibility and always focus on career capital. But it’s most likely the default for most graduates is that they still have to build their skills and build everything they need to have a bigger impact later on.

Benjamin Todd: Yeah, there’s definitely important exceptions. I think actually within the effective altruism community there’s, in particular, partly just because it’s a young community, there’s been lots of people who have had a big contribution early on.

Robert Wiblin: Yeah. Alright. So it’s, if not the biggest consideration early in your career, then it’s probably at least one of the top considerations. So the million dollar question I guess is which are actually the best options for career capital, which is something we’ve been trying to answer over the last couple of years and I guess we’re not completely sure about, but we have a bunch of pretty promising ideas.

Benjamin Todd: Yeah. I think this is where the question gets a lot more complicated because, to some degree, as we’re gonna kind of try and sketch out, just focusing on what has a big impact in the long run also gets you career capital indirectly. So it can be a bit unclear. Even though career capital could be really important in an abstract sense, how much you actually need to focus on it as a separate factor.

Robert Wiblin: So in terms of the best options, I guess an option that a lot of people have in mind, especially I guess because they’re focusing on the credentialing aspect of career capital, is to go into management consulting or some other prestigious corporate job which is going to be widely recognized as having skilled you up to some extent.

Benjamin Todd: Yeah. Also investment banking, professional services, law, things like that.

Robert Wiblin: Yeah, that’s definitely something that we’ve talked about in terms of career capital in the past and still do today. I guess where do we come down on how that trades off against other options and you know, for how many people might it be the best option?

Benjamin Todd: Yeah. So I think it probably got emphasized a little bit too much in our advice and because it’s a memorable thing that people are already interested in as an option, people probably end up focusing on it too much after reading our advice and yeah we released an article that goes into this in a lot more depth recently that is linked to from the key ideas series. But overall we would say it’s a good option and it’s worth many people considering.

Benjamin Todd: But there’s often an even better option for getting into the problem areas and the priority paths that we think are highest impact right now. And the main way we’ve arrived at that conclusion is just looking at these areas and then trying to think through what are the most effective routes into them by speaking to people in the areas and then it rarely seems like management consulting is the most direct route into any of them.

Robert Wiblin: Yeah. I guess a bunch of people have gone into consulting who were involved with the effective altruism community over the years and it seems like they’re doing well. Their careers are progressing well, but they don’t seem to have come out well ahead of people who just went into direct roles straight out. And I suppose that’s made us a little bit more cautious about suggesting that people kind of do a detour into prestigious corporate jobs.

Benjamin Todd: Yeah. It seems like when they come back they kind of end up in a position that’s similarly senior to what they would have if they’d been working within these nonprofits the whole time. Obviously it’s hard to tell, but at least it seems like they haven’t obviously got greatly better or on the other hand, significantly worse career capital than they would’ve got.

Robert Wiblin: Yeah, I mean we should point out that there isn’t room in these organizations for everyone to go and get a job immediately, so for people who can’t get into those things, going and getting career capital in some kind of prestigious corporate job might be extremely sensible. But I suppose if you’re choosing between the two, it’s not obvious that the prestigious corporate job gets you far ahead.

Robert Wiblin: Setting that aside, which options then do seem best for progressing your career as quickly as possible in the priority paths?

Benjamin Todd: Yeah, so this depends a lot on your skill set. If you’re a bit more on the research and analytical end and you can afford it, then graduate school is often an attractive option, especially a masters. And that’s just because having a masters in a relevant area just often does seem to be useful in many of the paths that we most recommend, especially because it’s very helpful in lots of policy careers and also sometimes just having that specific expertise is useful as well.

Robert Wiblin: Yeah. So I guess in terms of going to graduate school, some options that stand out are security studies, international relations, public policy.

Benjamin Todd: Yeah, that cluster is the policy careers cluster.

Robert Wiblin: Yeah. And then there’s some subfields of biology which seem pretty useful.

Benjamin Todd: Which is relevant to biorisk.

Robert Wiblin: Yeah. And then there’s economics and history I guess within the social sciences, if you count history as a social science.

Benjamin Todd: Well, yeah. So the ideal subjects for graduate school are ones that also give you lots of backup options outside of academia. And I think economics and machine learning and some other kind of more applied mathematical subjects really stand out on that front. Whereas something like history… It would be great if there were more people who studied history within our community of readers though, you know, doing a history PhD doesn’t give you that many great backup options outside of academia. So that would be more one for a smaller number of people who have really good personal fit with that area to consider.

Robert Wiblin: Yeah, I guess if you’re an analytical person and considering going to grad school, another option might be going and working in a think tank or some other kind of direct, perhaps more junior policy role?

Benjamin Todd: Exactly. Yeah. And I mean if we compare that to management consulting, well you’re getting a kind of general purpose training in these research skills, many think tanks are relatively prestigious. So you’re getting a bit of a credential and you’re also getting to learn about a potentially important area of policy and build a network of people who are interested in public policy. So you can see how it potentially can be very relevant for a lot of the areas we want to work on.

Robert Wiblin: Yeah. What about for people who don’t want to go back to school? Perhaps they have more of a doer personality. They want to go out and take a job and actually be creating things.

Benjamin Todd: Yeah, so I think again, there’s a bunch of positions in policy that could be a good fit for someone like that. I mean in general, there’s just so many options in policy. There’s things for all kinds of different skill sets. But yeah, outside of think tanks, you’ve got all kinds of jobs working for a politician. So in the US you could become a congressional staffer. Also in the US, you could work for an election campaign. That’s often one of the ways that people go in.

Robert Wiblin: There’s also a bunch of interesting executive branch positions.

Benjamin Todd: Exactly, and then that’s the other big class of options, is just trying to get into a relevant part of the executive branch or some other parts of the US government right away.

Robert Wiblin: Yeah. What about business? I suppose tech startups is something that a lot of people are interested in doing. How do they look for career capital?

Benjamin Todd: Yeah, so it seems for tech startups there’s obviously huge amounts of variation in how good they are to work at. But we have seen a decent proof of concept in that we’ve seen people go to tech startups and then often learn a useful skill set like management or operations and then come back and it’s seemed to be clearly useful in social impact organizations. So I think that’s often a pretty attractive option, especially if you are in the US or London or somewhere where there’s a bigger tech sector that you can join, though obviously you want to be selective about which startup you work at and try to find one where you can work with great people or it has a lot of buzz around it or you think it might be on an unusually good trajectory, which is a bit easier to do if you have a little bit of a network in that area.

Robert Wiblin: Yeah. I guess a benefit if it’s growing really quickly is that you can potentially end up taking on more responsibility at a young age than you might get if you’re working at a think tank, which is probably a bit slower growing and a bit more established.

Benjamin Todd: Yeah. I think even if the startup doesn’t do that well, you can often end up with quite a bit of responsibility early on just because it’s a small team. So everyone has to do everything and that’s one good advantage of it.

Robert Wiblin: Potentially give you pretty generalist training or lots of transferable skills.

Benjamin Todd: Yeah, exactly. And it’s skills that are relevant to more small and scrappy organizations, which is similar to many of the nonprofits and other types of organizations that we’re interested in people working at longer term.

Robert Wiblin: Yeah. So we have this pretty big interest in artificial intelligence. So I suppose also just going to work in any artificial intelligence related organization. I guess DeepMind, OpenAI are promising ones at the moment.

Benjamin Todd: Yeah. But I think it can also be worth considering a lot of the other AI companies as well, even though those two would be ideal to work at. And then again, that can give you similar training to a tech startup or the larger organizations. Maybe it’s a bit more similar to management consulting, but you’re also learning about and getting a big network in AI which is this really important area.

Robert Wiblin: Yeah. That’s one way that roles that don’t even have anything specifically to do with AI alignment. Just working in an AI related organization full stop can potentially introduce you to the right people and get you the opportunity to get a job that you think is really useful later on.

Benjamin Todd: Yeah, exactly. Maybe it’s also worth mentioning some kind of catchall categories as well because there’s loads of ways that you can potentially get really good career capital and we can’t list them all. And so some categories to consider is partly just if you can do something impressive in almost any area, that can open up lots of opportunities in the future because you’ll meet other interesting people. It’ll be a type of credential. I think just related to that, almost a subset of that is anything that can have a big impact in the next five years. And then thirdly, just anytime you get to work with a really good team or with a really good mentor, in almost any area you can learn so much from the people you work with, and so that can also be really worth considering even if it doesn’t obviously fit into any of the categories that we’ve listed.

Robert Wiblin: Yeah. You’re right. I was falling into the trap of just listing off a whole bunch of stuff which doesn’t make for great podcasting, but I guess the people who want more options can find that on the website on the key ideas page and other pages that we link off to from there.

Benjamin Todd: Yeah. And we do hope to write more about it in the future as well.

Robert Wiblin: I suppose a kind of a high level trade off that people often face is between getting very specialized career capital that really advances them in in this one track that they’re probably already on versus developing very generic skills like just being able to write well, which is probably useful in almost any job. Yeah. Where do we come down on this if someone has a choice between a job that’s very specialized versus one that develops generic skills.

Benjamin Todd: Yeah. I mean maybe just to step back a little bit. I think if you look at the options we’ve just been listing that seemed to be really good for career capital, just pursuing those options is not actually that different from just trying to get into directly what you think might be really good in five or 10 years. And so that’s kind of going back to the point I made earlier that sometimes you don’t need to kind of directly optimize for career capital as much, because if you just optimize for having a big impact over five or 10 years and doing something with good personal fit, then you’re already going to be getting good career capital as a side effect. So the trade off is a little bit less than I think it first seems. But then one place where there’s a trade off that can bite a bit more, is there does seem to be this trade off between specialization and transferability. So some next steps you could go into, they give you general skills like management that you could use in many different problem areas and many different organizations in the future. Whereas some things are much more narrow. So if you just became an expert in a really specific area of policy, then that would be less transferable outside of that. Though, I should say I think people think of policy careers as being a bit more specialized than they actually are because people do often move around within policy careers quite a bit.

How much risk should you take? [01:53:03]

Robert Wiblin: Yeah, I guess the risk if you go into one of those priority paths just straight out and then develop lots of specialized career capital is what if you change your mind about what you want to work on or the situation changes such that it doesn’t seem like it’s so valuable to work on that problem anymore. Then you could get caught with kind of all this stranded career capital that you’ve built up and then it’s a bit hard to move out of it.

Benjamin Todd: Yeah. Well so I just want to clarify that it’s not so much that the priority path is narrow. The crucial thing is what you’re doing over the next couple of years. Many of the steps into priority paths do actually give you lots of options such as doing an ML masters. That’s often the best thing to do if you want to work on technical AI safety or AI policy. But then doing that, even if you decide to switch out of those priority paths, is still going to be a useful masters to have studied for lots of other things. So in that case, you don’t need to be as worried about specialization too much, but maybe the specialization worry would come in a bit later. Having said that though, some next steps are giving you much less transferable career capital than others. And yeah, the obviously the downside of that is that if you decide to switch path, then if you’ve gained a bunch of career capital that’s only relevant to one thing, then you’ll be kind of throwing that away.

Robert Wiblin: Yeah. It seems that the community maybe has a slightly different trade off or a group has a slightly different trade off than individuals do, where individuals often don’t want to close off options. There’s this risk aversion that makes people reluctant to do that. But then as a group of people, you need some people to go out and explore and do specific things that necessarily kind of close off some options for them.

Benjamin Todd: Yeah. I think if you just think of the whole composition of the community, it’d be pretty bad to have just a thousand generalists but no one knowing anything particular about anything.

Robert Wiblin: Hard to get anything done.

Benjamin Todd: Yeah. It seems like it’d be much better to have a large number of people be specialists and then they can all work together to do the particular projects that need to be done. Whereas yeah, I mean if you think about it more from a personal satisfaction point of view, then there’s a stronger argument for making sure you can stay flexible.

Robert Wiblin: Yeah. So maybe a reasonable way to conclude the section might be perhaps to focus on generic skills very early on, but then you do at some point want to actually start specializing and not be afraid to do that.

Benjamin Todd: Yeah. So I think it is a difficult trade off to make because the other point is just that some degree of specialization does eventually seem to be necessary for almost any of the options that seem highest impact. At the very least, you have to learn a lot about the specific problem area that you want to work on at some point while gaining career capital. And so it seems like it is often worth specializing more and paying the trade off of less flexibility. Though if there were two steps you needed to take towards your long-term option and one gave you more transferable career capital, then clearly do the more transferable one first and then delay the specialization if you can. But it does seem like you need to do it at some point.

Robert Wiblin: Yeah.

Benjamin Todd: To sum up career capital overall for someone right at the start of their career. The rough way I would approach it is: try to think about some long-term options that you might want to aim for that seem like they could be really high impact and work back from there. Think about which things might accelerate you the most in those paths and then you could probably choose between one of those based on personal fit and that’s roughly how I would weigh it. If you have a really good opportunity to accelerate yourself in a priority path, then probably take that. But if you have an exceptional opportunity to have an impact, then probably take that.

Robert Wiblin: Alright. Trivial. I’m sure everyone has figured out now what they’re going to do. So suppose to try to wrap up on career capital. The situations can get very complicated, or the trade offs can get pretty difficult to weigh up, once you start thinking about concrete examples, at least if you have multiple options that seem kind of sensible. A classic one might be, let’s say you’re already working as an AI engineer and you’re thinking, “Should I go and get a PhD and try to become a research scientist in AI instead?”. On the one hand, you could be doing something that’s more useful right away and also building relevant career capital by actually working on a project. On the other hand, you can go off to school for a bunch of years, but then try to jump up to a higher level. It depends a lot on the specifics of the person’s situation. So is there any quick way of wrapping up how people should weigh career capital against other stuff?

Benjamin Todd: Yeah, so we’ve been talking about the easier case when you’re in your first couple of jobs and you’re not really having any impact, so it’s basically all about career capital ultimately. Whereas later on in your career, you have these more fine grain trade offs where you could have a bunch of impact on one path or you could have less impact for a while and then more impact later and weighing that gets a lot more complicated. One big question to just bear in mind is which path will enable you to contribute the most quality adjusted labor to that problem area over the course of your career? And you can just try to think through the two trajectories and how much more productive you would be made. But then there’s a few other considerations. One is there’s a risk that your efforts become moot in the meantime. So if you’re trying to prevent existential risk, then obviously that risk happens then —

Robert Wiblin: You arrived too late to the party.

Benjamin Todd: Yeah. And there’s other things that could cause your efforts to become moot. So that can create a reason for urgency. And then the other thing is that there could just be timing considerations. In general it seems like as you get closer to one of these hinge moments in history that we talked about earlier, it becomes easier to see what to do about it and there’s an argument for then you just want to spend all your resources and try and make that hinge moment go well. And then if you’re not near a hinge moment then you want to invest as much as possible. And so then you’ve got these four factors that we just mentioned and you’re kind of trying to weigh them all.

Benjamin Todd: Where that seems to shake out in practice is that in general, you get one group of people who are a bit more concerned about the urgent existential risks and they’re a bit more in favor of having an impact sooner and not focusing quite as much on career capital though usually still somewhat on career capital and then the people who are more focused on the boring longtermist perspective that we mentioned earlier and they would be relatively more keen on gaining career capital. We have already recorded a podcast that is all about this now versus later trade off.

Robert Wiblin: Yeah. Fingers crossed that will be out in the next couple of months if it’s not already by the time this one gets out.

Personal fit [01:59:00]

Robert Wiblin: Alright, let’s move on and talk about another kind of key career strategy issue which is personal fit. I suppose personal fit is kind of most of what you would hear about if you went for classic careers advice perhaps if you’re about to graduate university. I suppose for us it’s somewhat de-emphasized, but we still think it’s pretty important. Yeah, how big a deal do you think personal fit is in the scheme of things?

Benjamin Todd: Yeah. I mean we have a bit of a challenge in that we think that most careers advice mainly focuses on things around personal fit and it ignores all the other stuff we’ve just been talking about. So for instance, how some problem areas might be much more effective than others. And so we see a big part of 80,000 Hours as just pointing out that these other things matter as well and they could maybe be even more important for determining how much impact you have. And that then kind of means that people who’ve read our advice a lot sometimes then end up underemphasizing personal fit because that’s not the central idea that we focus on, but it is still really, really important. And so one way to think about that is you can kind of imagine that in some career paths you probably just wouldn’t stick with them at all and so you’d basically end up having zero expected impact in them. Probably most career paths. Whereas in others maybe you’d do really well and then in others you might have a chance to be really exceptional in that path. And so the difference in your impact between that top scenario and the bottom scenario is potentially orders of magnitude of difference maybe.

Robert Wiblin: Yeah, I guess once you realize that most of your impact is probably more than a decade in the future. And also just recognize that… I guess I don’t know anyone who kind of stuck with a career for more than 10 years that they hated, at least if they had other options that they could seriously get into. Then you realize that taking a job in which you’re not flourishing, which you resent having to go to work is probably not a great idea.

Benjamin Todd: Yeah. One way to see that personal fit might actually be even more important than common sense advice would say is if you are in one of these areas where the people who perform best in the field have significantly more impact than the typical person in the field, then it’s extra crucial to increase your chances of getting into that top end of the field. Not every career has that pattern, but some of the ones that we talk about, such as policy careers, research, entrepreneurship, they seem a bit more to have that character.

Benjamin Todd: And so then that actually means that personal fit matters even more than people normally think. I think another big reason why personal fit matters a lot is that, as I was alluding to just now, the more personal fit you have in your career, then the more successful you are in it and then that also gets you better career capital. So it’s also just this really useful instrumental thing. Doing really well at almost anything gives you an interesting network and it shows that you can achieve things and be a top performer and that opens up other opportunities and if the thing is somewhat relevant to the cause areas that you want to work on as well then that’s even better.

Robert Wiblin: Yeah, I guess over the last 10 or 15 years I’ve just seen a bunch of cases where people are doing well at some job and then they switch the task that they’re doing to something that’s a bit different or work at a different organization and suddenly they start really flourishing and just super taking off in their career. Whereas before, they were just doing well. Just kind of keeping up with everyone else. I guess this somewhat informs my view that it’s worth doing a bit of exploration or moving around early on to find something in which you’re not merely good, but you’re excellent and that perhaps relatively small changes in what it seems like you’re doing or who you’re working with can make a big difference to fit.

Benjamin Todd: Yeah, I think that’s right. There was another point in the recent anonymous advice series where one of the people we interviewed pointed out that being really good at your job has all these other benefits that we were talking about, but also being really good at your job is harder than most people think in that we’re used to being at university where you just work for a few months and then you get an A and you’re doing well. Whereas becoming really successful in any of the areas we talk about could easily take five years, 10 years of full time work to get there. And so that just means it’s just essential that you stick with it for the long term.

Robert Wiblin: Yeah. I guess how often do you think people should take a job that seems worse in other ways or work on a problem that maybe we don’t think as, in general, quite as high a priority because they’re a better personal fit for it.

Benjamin Todd: Yeah, I mean you could think of your long-term impact, just to super simplify, as the multiple of how good the path is in general and your degree of personal fit with it. And that would actually mean both factors are very important and it could well be worth taking something that maybe you’d see it as second tier in terms of how good it is in general, but if you have significantly better personal fit for that path, that could well win out.

Robert Wiblin: One complication with personal fit is that it seems like people are not super good at predicting what things they’re going to really flourish at and what things they’re gonna really love. Sometimes you have people who think, “I’m going to be great at being a doctor” and then they try it out for a couple of years and just find that they’re not flourishing as much as they expected to. Not by a long way.

Benjamin Todd: Yeah, and there’s all that literature about people being bad at predicting what makes them happy and satisfied in the long term. And yeah, I mean also the literature where people tried to predict job performance, there just aren’t any predictors of job performance or expertise that are super reliable.

Robert Wiblin: Yeah. It’s kind of an interesting phenomenon. Amongst a bunch of not fantastic ways of trying to figure out what things you have great personal fit for, which ones look least bad?

Benjamin Todd: Yeah, so we cite a meta analysis of ways to predict job performance on the website. And there’s a recent update to that meta analysis and it finds that structured interviews are not too bad a way of doing it. And I think you could proxy that by speaking to people in the field and asking them what are the ingredients of success needed for someone in this area and then how do I stack up on these and just trying to get an honest appraisal from someone who knows the area. I think that can often be useful. Another thing that can be useful and is a bit more time consuming, is to actually try out the work in some way and work sample tests also can be a pretty reasonable predictor of job performance. And yeah, if you can go down the path for a little bit or do a couple of week projects in the area and try and get a sense of how you’re comparing, that can also be very helpful.

Robert Wiblin: I guess people sometimes have this problem that people don’t want to discourage them in some career paths so they get kind of over-inflated feedback on how good they are at something. Back in the episode with Tetlock, the second interview I did with him, we talked about how very often PhD students get encouragement from supervisors even when probably their supervisors know in their heart that they’re pretty unlikely to get an academic job. I guess Tetlock suggested that you should go to them and ask for a brutally honest opinion, even if it’s going to be very unpleasant because the alternative can sometimes be going many years down a path that then just leads you into a dead end which is much more costly.

Benjamin Todd: Yeah, I mean when it comes to predicting personal fit more broadly, it is one of these difficult prediction challenges and a lot of Tetlock’s advice on just how to make good predictions in general. You can just directly apply to that question. So there’s lots more interesting thoughts in there.

Robert Wiblin: Alright, let’s wrap up on personal fit. Was there any last things you wanted to say about that?

Benjamin Todd: I’d just say yeah, it’s one of the most important considerations in choosing between options and often the way it seems to go when we do one on one advising is someone will generate a bunch of promising next steps or promising long-term options, and then often it seems figuring out personal fit is the main thing that is then used to tie break between the options and a lot of the focus of the effort in the later part of a career decision. So yeah, once you narrow down a bit, it’s one of the most important considerations.

Robert Wiblin: Yeah, I mean all of these things can be quite hard to predict, but maybe what you’re actually motivated to do at any given point in time is among the more measurable things that you can know. So inasmuch as you’re super psyched about a job, maybe even if being psyched isn’t the number one consideration, it might be among the more measurable things that you can figure out.

Benjamin Todd: Yeah, I mean I would definitely want to consider that and I think about that a lot in figuring out what I want to work on personally. I guess I wonder a little bit about how exactly it squares with the literature about how people claim a thing will make them really happy and then actually when they’re there and they survey them on how happy they are, it didn’t make any difference.

Robert Wiblin: Yeah, maybe it works better in the short term if you’re like, “What do I feel psyched about doing today?”, then you have a better measure of that than, “What will I still enjoy doing in a year’s time?”.

Benjamin Todd: Yeah, that seems right.

Exploration [02:07:07]

Robert Wiblin: Okay. So the third career strategy issue that we wanted to talk about was exploration, which I guess is highly related to the previous two, to figure out the job where you’re gonna have the best personal fit, you’re gonna have to potentially try out a bunch of different things because you can’t just predict it ahead of time.

Benjamin Todd: All of the strategy considerations overlap and interrelate a lot.

Robert Wiblin: Yeah. So what are the key messages that we have for people about exploration?

Benjamin Todd: Well I think one big point is just that it’s just very, very uncertain which career will be best for you. We’ve just been saying it’s really hard to predict personal fit and that just means that also, one of the most important priorities early career should just be to try to learn more about which things you’re going to be best at in the long term. And in particular to try out some of the options that you think might be most promising longer term. Yeah, so you could think of that as… Technically it gets called value of information, I’m just going to call it information value. And so early career you actually could kind of think of yourself as both optimizing for impact and career capital and information value and trying to do the next steps that give you the best balance of all of these things.

Robert Wiblin: Yeah. So I guess that has the perverse outcome that you could potentially go into a job and completely bomb at it, be completely terrible and be like, “Wow, that was fantastic because I gained so much information. Now that’s helped me to rule out tons of options and so now I can go out and try something that I’ll actually potentially be good at”.

Benjamin Todd: Yeah, I think it is really helpful to think of, especially early on in your career, think of these jobs you’re taking as tests and in some ways it doesn’t matter how well it goes. What matters is whether you learned about what’s going to be a good fit for you longer term.

Robert Wiblin: I guess not all jobs that people can try out are equally informative. We have this idea that in some jobs there’s a greater dispersion or a greater variance in how well people perform and perhaps you can get the most information by trying options where it’s like shooting for the stars and if you’re a superstar, then that’ll be massively good. And so you get more information inasmuch as there’s more uncertainty about how good you’re going to be in the first place.

Benjamin Todd: Yeah. So one thing is if it seems like you’re in one of these fat tailed areas where the people who do the best in the area have way more impact than others, then your only concern would be to figure out if you’re going to be in that tail or not, just to really simplify. And so then if you try that area and then if you get evidence that you’re probably not, if the probability is sufficiently low that you would be in that tail, then you’d probably want to switch into something else and keep trying and keep trying and keep trying until you eventually find the thing where you’re more likely to be in the tail.

Robert Wiblin: Yeah. I suppose a clear example of that might be, let’s say you want to become a famous singer. There’s only room for a relatively small number of famous singers in the world and so you could perhaps relatively early find out, you know, I’m actually just not that great a singer and so the odds of me becoming a superstar are actually now pretty close to zero.

Robert Wiblin: So that’s a good sign that maybe you should go off and try something else.

Benjamin Todd: Though I’m sure there are all kinds of stories about people who tried for decades and then suddenly broke through. Yeah, it can be complicated.

Robert Wiblin: Okay. So in order to avoid wasting time, how do people gradually increase their investment into different paths? What’s the escalation that people can do as they’re exploring their fit for something.

Benjamin Todd: Okay. So we’ve said that because there’s so much uncertainty, it’s really important to early on try and test out things and learn about where you have the best personal fit. So you could imagine an early career plan could be to identify the option that you think is most promising. Try that first. If that doesn’t work, go to the next one, go to the next one and kind of keep trying like that. Though we actually think you can do a little bit better than that. And so you can consider the question, “Which option, would you learn the most from by trying it out?”. And that’s the option that has the highest value of information. And that seems to be basically an option that might be really, really, really good, but where you’re really, really uncertain about it. So you can think of that as an option where the probability distribution of the outcomes is really wide and there’s this tail of outcomes that might be really good, but you’re very uncertain whether you might get it. It can sometimes be worth taking something like that over an option that might actually have higher expected value than it if you can test it early and figure out whether you’re going to be in the really good outcome or the median outcome.

Robert Wiblin: So I guess there’s a sense in which it’s kind of a win-win because if you’re particularly good at that, then you stay with it and if you’re not good at it, then you just quit and go do the next thing and it hasn’t been that costly to experiment with it. Is that the logic?

Benjamin Todd: Exactly. So another way of seeing it is as a kind of asymmetry where if you go into this thing which has a really high upside but might easily not work out, then if you can try it for a year or two and then it actually does turn out to be in the really good scenario, then you just carry on with that maybe for decades. So you’ve got this huge prize from doing that. Or, if it turns out to indeed actually not be as effective, which is maybe the default outcome, then hopefully you’ve only spent a year or two doing it and then as long as you’ve kind of planned ahead, try something else.

Robert Wiblin: Yeah. So that’s one approach, but there has to be a limit to how much you can do that. You can imagine someone who just keeps year after year trying something different every time. That’s kind of a high risk thing that they mostly expect won’t work out. But there’s only so long you can do that before… Maybe you’d just be exhausted with failing and the other considerations in life start creeping in.

Benjamin Todd: Exactly. So information value, just career capital, you should focus on it more early career and so one and then again for a very similar reason, the earlier on in your career you learn about something really good, the longer you can do it and so the more valuable it is to have that information. So that just naturally means you should do a bit less exploration as your career goes on and kind of become more conservative.

Robert Wiblin: So yeah, I guess become more risk averse. I guess the bar of the probability of you succeeding at a job should kind of go up and up over time. I guess when you’re 22 you can really shoot for the stars, whereas when you’re 30 you kind of want to have at least a reasonable chance of success.

Benjamin Todd: Yeah. So technically I would be focusing on the options which have these really big upside scenarios early and then just focusing more and more on expected value over time. It just converges to expected value.

Robert Wiblin: Yeah. So I guess there’s kind of that declining benefit thing but then there’s also just, I suppose people often want to settle down at some point. So I guess it’s easy to recover from getting fired from your job when you’re 22. But maybe when you’re 35 it becomes a bunch more costly because you’ve got other commitments in life that mean you just can’t take unlimited career risks.

Benjamin Todd: Yeah. So I would actually try to factor that in. Maybe it’s just better to think of that as a different consideration, which is well, you have your impact goals and you have your personal goals and when it comes to your personal goals, you want to be quite risk averse probably because, well, yeah, lots of reasons. But just like losing all your money is much, much worse than gaining the same amount of money is good. So you really want to not have the worst scenarios happen. Whereas from an impact point of view, you can be like, it’s fine to take a high risk, high return option because if that has a high expected value. So you then just need to trade those two things against each other depending on how much personal risk you’re willing to put on. And then I suppose you’re right that in general people are less willing to take personal risk later on in their careers and that’s because they have dependents and so that’s kind of increasing the downsides.

Robert Wiblin: I suppose there’s this whole other thing about life circumstances that can change how much you’re willing to take risks. It’s like, do you have a whole lot of savings? I mean this is a reason to have that aspect of career capital, which is having a bunch of money in the bank because it just allows you to take risks that have very positive expected value but might expose you to some significant downside.

Benjamin Todd: Yeah, but that’s a good example of why you’d actually be able to take more risks later career because you tend to be more financially setup later career.

Robert Wiblin: Yeah. That’s interesting.

Benjamin Todd: And you do see people who’ve done that. They’ve retired from their corporate job and now they just kind of do whatever they think is the highest impact even if it could easily not work out.

Robert Wiblin: Yeah. That’s a great point. I suppose just generally you’ve got this question of, “How good is my backup option if this doesn’t work out?”. I suppose if you’ve already retired from your corporate job and have enough money to live forever, then your backup option looks pretty great.

Benjamin Todd: Yeah, I think there’s probably other reasons why it’s easier to explore early. One thing is just it seems society is more tolerant to people switching jobs a lot early because it’s kind of expected and you also have these opportunities like internships which can be good for testing out a thing but then older people aren’t allowed to take internships. So you have these unusually good opportunities for exploration that you get early but not later. Probably there’s also just diminishing returns as well where once you’ve tried out a bunch of jobs, you’ve learned a bunch about your personal fit and you’ll learn a bit less from trying out future things.

Accidental harm [02:15:25]

Robert Wiblin: Yeah. One consideration we haven’t talked about is that sometimes it’s not only a question of having a big upside versus no upside. It could also be a question of big upside versus massive downside. And if you’re in an area where you potentially do a lot of harm, then just taking massive risks is not necessarily such a winner.

Benjamin Todd: Yeah, exactly. So I think with this kind of approach of focusing a bit more on these high risk, high return things, there are two important caveats to that. One is you need to also manage your personal risk while doing it, which we just mentioned. And then having a good backup option is important for that. And then also you want to avoid any risks of having a significant negative impact in the field that you’re focusing on. So one way you can think about it is just filter out anything like that first and then from among the remaining things, focus more on the things with the potentially really big upsides.

Robert Wiblin: Yeah, we won’t talk much about ways that people can accidentally do harm here, but we have a whole article, I think it’s “Nine ways you can accidentally do harm in your career”, which we link to from the key ideas page. I suppose this issue of exploration raises another weakness that you can potentially get by going into prestigious corporate jobs which is that almost no one expects that working in strategy consulting just for businesses in general is going to be their very best long-term option. And as a result, they’re not really narrowing down this question of what is their very best option that they really want to end up in the long term. It kind of slightly fails in exploration in that sense.

Benjamin Todd: Yeah. So we quite often come across people who say, “Well I’m not really sure what to do long term so I’m going to do one of these corporate job things to basically buy time”, which could sometimes be reasonable but, if you’re really uncertain what to do long term, then the most important priority is to try to figure it out. And the sooner you can figure it out the better. And so it’s really important to think with exploration, what you actually want to explore and test out, is the things that might be best for you long term and you want to try and test those as soon as possible and falsify them or gain evidence for them being your best thing. And then, yeah, if something like consulting is not something that you’re planning to focus on long term, then you’re not getting that benefit from doing management consulting. Though I’d say management consulting is better than many other prestigious corporate jobs for exploration because you do at least get to try out working in several different industries. So it does have that benefit. Whereas if you think of one of the other categories like law, that doesn’t have that.

Robert Wiblin: Yeah, and I guess even if you’re not learning a lot about specific paths, you’ll learn potentially a bunch about yourself, like what kinds of tasks you enjoy doing, what your strengths are.

Benjamin Todd: Exactly. Though you get that in almost any job. So if we’re just considering the delta between different things then…

Robert Wiblin: Yeah. Another challenge that we’ve seen some people face is that they go into jobs in order to explore their options and then they realize that they want to leave that option, but they find that they’re just so busy doing their actual job all the time that they barely have any time to actually learn about ways of ways of getting out, which I suppose is a trap that people should potentially be aware that they might fall into and try to avoid.

Benjamin Todd: Yeah, exactly. Sometimes people have done this job for a couple of years because they were so uncertain, but then they come out and they find out that they’re just as uncertain as they were going in.

Robert Wiblin: Yeah. Alright. Is there anything else to say about exploration? I suppose there’s a whole lot more on this topic in our podcast that both of us did actually with Brian Christian who wrote “Algorithms to Live by” where we talk about a bunch of mathematical models of this explore-exploit trade off. But is there anything else that we want to say here?

Benjamin Todd: So yeah, this whole discussion has been all about trying to figure out what long-term path is best. But if you’re facing a choice between your next step, so something you might just do for the next couple of years. Then again, all this uncertainty means that testing out and doing more exploration can be really important as well. And you can kind of think, the bigger the stakes of the decision, the more exploring testing research you should do before making the decision. So if you’re choosing where to spend several years, then it could easily be worth spending 10% of that. So that would be like a month or two, trying to figure out which option is best. We do often come across people in our advising who’re facing a decision and then they’re just only trying to analyze it and trying to figure it all out from a desk and just think it through. When often it can be really useful to try to think, “Well concretely what am I actually uncertain about and then how could I just go out into the world and learn about that uncertainty?”.

Robert Wiblin: Alright. Okay. So that was career capital, personal fit, exploration and I suppose a bit about risk management as well. Is there any way that we can kind of tie together all these different points into a cohesive vision of career strategy? There’s a lot of different considerations here that people are trying to weigh up and I suppose one answer is just that there is no one dominant consideration here. There’s just a lot of things that you have to put down on the page and then try to weigh them up against one another.

Benjamin Todd: Yeah. So I think the best way to think it through is to consider all these different considerations and try and balance them against each other and that’s what we lead people through in our career decision process where we recommend for your next step, you could score them on impact, career capital, personal fit, maybe degree of personal risk. What other factors do you think are most important and try and take the option that has the best balance of those things. But I think there are obviously some patterns in what we’ve been saying. And so you can see that in general, early on in your career, say especially in your first couple of jobs, that’s really about trying to test out promising long-term options and get career capital that could accelerate you in those options. Whereas later career it becomes more about just trying to do the thing that you think has the highest expected value. Yeah and then personal fit is probably just important throughout your career, later career because doing something with better personal fit is higher impact and early career because it builds your career capital faster. Also in general, you’ll always want to have a backup option for the thing you’re focusing on the most as a type of risk mitigation. So yeah, we kind of sum up some of these points on the key ideas series.

Robert Wiblin: Alright, nice. It’s a bit of a shame that there’s so many things to take into account when you’re making career decisions. But no one said that careers were gonna be easy.

Benjamin Todd: Once you consider these different strategic factors there do tend to be a couple of formats or schema like career plan schemas that often make sense. One of those is just try and figure out your most promising couple of options and then aim to try out all of them and then choose the one that seems best after trying them in a couple of years time. We emphasized that quite a bit in the old career guide. Another schema is just to choose the thing that has the highest expected value and then do that and if it doesn’t work out, try the next best thing and so on. But as we covered, that plan is actually neglecting information value. You can then do a slight modification of that plan which is then to take the best upside option first.

Benjamin Todd: So that’s the thing where if it goes unusually well, if you end up at the top end of your expectations, then the thing that’s best in that scenario then do that one and then if that doesn’t work out, try your next best upside option, and then as we were saying, get more conservative over time as your career goes on. And so that’s the tryout upside options, career plan schema. The fourth career plan schema is just to… This is the one to do if you’re super, super uncertain is just build transferable career capital and then decide later. And I think yeah, in general, you want to be doing the career capital one or trying out lots of options at the most degree of uncertainty. So maybe in your first one or two jobs and then I think the kind of upside option one is the one that we most commonly would recommend and then the “Just do expected value one” is more the late career option. I’m sure that clarifies everything.

Robert Wiblin: One slightly convenient simplifying thing is that very often, the thing that you have the best fit for both allows you to have the biggest direct impact and also to build the most career capital and also working on the problem that seems the most pressing that you want to do long term also potentially offers you the best career capital. So it’s not as always as if these factors are completely independent and then there’s always big trade offs. Sometimes they just all come together.

Benjamin Todd: Yeah, that’s a really big important clarification. So the idea of the upside option is by pursuing that long-term upside path, firstly you’re getting information value because the path might be super good, but you’re unsure about it. Secondly, while you’re on this good long-term path, you’re getting some relatively good career capital. Thirdly, you’ve probably got good personal fit for that thing because otherwise it wouldn’t be one of your top upside options. And so it seems often all the factors can converge on something.

Robert Wiblin: I guess the one that sometimes doesn’t is exploration. Like the thing that you know is good versus the thing that might be even better. That’s one that you can potentially face at any point.

Benjamin Todd: Yeah. I think also the specialised versus transferable career capital thing can be a tricky one.

Personal wellbeing [02:24:02]

Robert Wiblin: Okay. Let’s move on to another career strategy issue, which is how do you make your career fit into your broader life and the other concerns that you have? I suppose this is something that people in the effective altruism community have debated a lot over the years and it has sometimes been a bit of a controversial one. What do you think personally and I guess what does 80,000 Hours recommend?

Ben Todd: Okay. Yeah. So the issue is what happens when you have a conflict with what would make you personally happy versus what seems to do good in general, which is our focus at 80,000 Hours. So there’s a couple of points to make about this. One is that I think if you pursue the vision of a career that we’re laying out here, you are trying to do something meaningful. You’re trying to find something that you’re good at, you’re trying to work with other people to contribute to a pressing problem. These are all things that are very satisfying to do as well. And there’s many types of satisfying career, but I think that is one of them. And so by kind of following the 80,000 Hours advice, often you can end up quite personally satisfied as well from doing that. So there’s a bit less trade off than it first looks.

Ben Todd: That said though, I think there can definitely be trade offs that you face. It’d be very, very lucky if the morally best thing perfectly lined up with the thing that was best for your own wellbeing or just your other goals. And then exactly how to deal with those is just a very tricky issue that’s very debated in moral philosophy and you have to really weigh up your specific circumstances. One line of thought on this is the kind of Peter Singer argument that if you can help other people with little cost to yourself, then you ought to do that as much as you can until it becomes a major sacrifice to you. In Singer’s paper,’ Famine, Affluence and Morality’, he argues that we should never have any luxury spending, we should give away most of our money to help those in global poverty. That would really argue for the main focus of your career and your life should just be helping others. One point about Peter Singer’s argument is, people often attack it by saying, “Oh actually you can’t help in international development very much. There’s nothing effective to do there”. But Singer’s argument is actually much more general than that. As long as there’s any way to help others without a huge cost to ourselves, then you should spend almost all of your resources doing that.

Robert Wiblin: It’d be pretty astonishing if there was no way anyone could find to do a lot of good to help other people.

Ben Todd: Yeah, and the way we approach this is, if there’s some chance that now is an unusually important time in history, we’ve gained all this technology and wealth that’s pretty unprecedented in history and our actions today might affect the entire long-term future, then maybe just the stakes of our actions are really, really high, and again, that should be the main thing we focus on. That’s the arguments on those side. But then there’s lots of philosophers who argued that the demands of morality are not so pressing as that. It’s permissible to also have your personal goals and work on those as well.

Ben Todd: I’d say me personally, I just see making a difference partially as one of my important goals, but not my only goal in life. The way I see it is it’s the main aim of my career and I give 10% of my income or forgo 10% of my income to charity, but then with the rest of the 90% of my income and my free time, I’m free to do whatever I want with that. I definitely do lots of things that are very ineffective from a ‘having an impact’ point of view. Buy overpriced clothes or …

Robert Wiblin: Eat fancy food.

Ben Todd: Get too into fancy food, yeah. That’s kind of how I approach it, and I think many of our readers do as well. Then finally, there’s another point of convergence that just everyone agrees, even if you fully accepted the Peter Singer argument that there’s no point making yourself miserable and working yourself ragged and then burning out. That’s definitely not the best thing to do even from an impartial, doing good point of view. I think actually probably if you want to maximize your impact you should probably be doing something you enjoy and can do sustainably. You need to treat yourself with some compassion, realize that we’re not perfectly moral and we’re not going to be able to always do the abstract best thing, both for your own being able to keep up with it, but also, if you can get other people interested in doing good as well, that’s another way of having an impact. If everyone trying to do good is super miserable all the time, it’s probably not going to be the most effective thing.

Robert Wiblin: Yeah, I think it’s interesting. I feel I’m a lot more willing than you I think to take the more extreme, dogmatic, very demanding philosophical position or to say that that just may well be correct, that morality wasn’t made for people and people don’t want morality to be really demanding, but it just happens that it actually is. But then I think I might go even further than you in saying that there’s not that many practical tradeoffs in real cases, because we can go through a whole lot of different considerations. One is, you seem to have most of your impact after you spent a long time in a career and you’ve done a lot of expertise, and then if you’re not enjoying it then most people find it extremely hard to stick with the career for many years or decades. Yeah, if you hate something, how likely are you still to be working on that when you’re 60? Things will intervene. Then there’s also just, yeah, people who are unhappy, I find that it tends to cloud their judgment or cause them to behave strangely and to get stressed out and not get along with their colleagues. It’s all of these negative effects, that being negative has on your life and your–

Ben Todd: Being really motivated by guilt doesn’t necessarily seem like it leads you to do the most effective things.

Robert Wiblin: Yeah, it seems really crushing. People, they seem to develop odd complexes when they’re motivated by guilt. I’m not quite sure I can put my finger on it here, but it doesn’t seem like it leads to good decision making relative to being really happy. I think this issue of, if you’re unhappy then you’re going to repel other people from wanting to work on the same things that you do and from wanting to collaborate with you and from agreeing with you. That can be a really enormous factor. If you imagine that over the course of your career you might recruit one, two, three, four other people to join in the similar work to what you’re doing because you’ll just be potentially a charismatic and appealing person to be associated with. But let’s say that you won’t be able to do that if you’re constantly miserable because you’re working every hour of the day and get no time off. That is an enormous factor, I think. You have to really think you’re getting a lot more work done.

Ben Todd: There is the counterpoint to that that people who do extreme sacrifice can be very inspiring as well, because they act as these saint-like figures that motivate others.

Robert Wiblin: Yeah. That’s interesting. Yeah, I agree with that, although how many people do they attract to then try to emulate them and go and get another job? Just most people aren’t interested in self sacrifice in the same way, whereas they would be if they could live a perfectly good life while doing a lot of good.

Ben Todd: That’s true. Maybe you get a lot of attention but not many people following you.

Robert Wiblin: There’s also the fact that many of the jobs that we recommend pay reasonably well and they’re very intellectually stimulating and also they’re relatively high status in society. People tend to like you for doing the things that we suggest.

Ben Todd: Well yeah, that’s if you are able to land one of these really good things. I think probably where a lot of the trade-off comes in is if you focus on this very impact point of view, then you’re going to want to be pursuing these slightly higher risk things that might be really good but there’s a good chance they don’t work out. So the trade-off comes in taking greater personal risk where maybe you spend several years trying a thing and it doesn’t go anywhere. At least that’s only several years of cost, but it might set back the rest of your career quite a bit compared to if you’d just done a more standard option that was going.

Robert Wiblin: No, I think one area where I might think that it is good for people to make a sacrifice is to take big risks in their career and to accept the possibility that their career could end up being not as interesting or not as successful in the long term because they take a bunch of risks and none of them pan out. But I guess maybe this is a personal perspective, but I feel like even people who do that, it’s not as if they end up with unpleasant lives on a day to day basis, or if their lives are unfulfilling, all things considered. Especially the people that we tend to be advising tend to have many interesting career options, and even their fallbacks are pretty stimulating and don’t end up with them in penury.

Ben Todd: Yeah, I would always say try to have … Well, we talk about having a plan Z no matter what you do, something that would still be acceptable to fall back into, and if you can’t see what the plan Z is from this path, then I think it’s also pretty reasonable to not do that, and do something where you can see a plan Z or you don’t need a plan Z.

Robert Wiblin: Yeah. I should say I’m partly saying these things in reaction to people I know who are really dedicated to doing good who have decided to make this the number one thing that they think about all the time. For them, I often want to say, chill out. You don’t have to rush, you don’t have to stress everyday about whether you’re having lots of impact. Just get the big picture right, be heading in the right direction. That’s a pretty good thing to do. Just in expectation, things will work themselves out, because most people are not really thinking that much about social impact in their career in general. For them I would recommend, maybe they should think about that a little bit more often and be willing to make some personal sacrifice or take some risks.

Coordination [02:32:46]

Robert Wiblin: Another important aspect of career strategy is figuring out how you can have a larger impact by coordinating with other people in such a way that you can share fixed costs, get economies of scale, use your comparative advantage, so each of you is doing the best thing or the thing that you’re most capable of doing.

Robert Wiblin: How does the career strategy picture look different when you’re working as a group of people rather than just as an individual?

Ben Todd: Yeah, in general it seems that people can have a lot more impact if they work together rather than individually. A way of seeing that is, if you’re going to start a company, one way of doing it would be everyone’s be a generalist, but it seems like it’s generally more effective for one person to be the product person, one person to be the sales person, one person to be the marketing person. Then by doing that, the collective impact will be sufficiently big that they’ll all end up with a bigger share than if you tried to do everything yourself. Those are all the mechanisms that Rob mentioned behind that, and so yeah, we want to do the same when doing good. I think it seems like taking that perspective might actually change which options seem most promising quite a bit.

Ben Todd: Though this is still quite an active area of uncertainty for us, and our article that we have on this is still not fully finished, I’m just going to talk through some of our initial ideas here. If you more ignore coordination, then the challenge is more just trying to identify the single best opportunity at the margin, and then taking that and holding everyone else fixed. I call that the single player approach. As an example of how that can go wrong, if you’re thinking, should I work at this nonprofit? Then on the single player approach, you might think, well if I don’t work at this nonprofit someone else will take that job otherwise, and so that doesn’t really seem like a very effective opportunity. But we can see that if that other person who would take the job shares your values and also wants to work on these problems, then by you taking the job, you free them up to go and do something else that’s potentially high impact. The question then doesn’t so much become what’s the job that no one else is going to take at the margin, it’s more, of all the job seekers currently interested in these problem areas, what’s the ideal allocation of these people between all these opportunities that are currently open?

Ben Todd: That’s how the picture changes a bit with, say, job opportunities. Here’s another example of where the top options might change, and that’s the question of which problem areas people should work on. From a single player perspective, you’d probably just be trying to identify the single problem that you think is most pressing and that you have good personal fit with, and then you just go and work on that, and you wouldn’t really think about what other people were doing too much. But if you think there’s this whole community of people who are all potentially working together and shifting between opportunities, then as more people go into one area, that area will have diminishing returns.

Ben Todd: That, and other reasons, will mean that from a large scale community point of view, the community should invest in several different areas, similarly to how a large foundation would want to donate to several different causes. How would you then work out which area you should personally work on? The question then becomes, ideally, if you could choose exactly how the community was allocated across problem areas, what percentage of resources would you put into all the different areas? That’s the ideal allocation. That’s the first question. Then you want to think, okay, in what ways are we out of whack with that allocation compared to where we’d ideally be? Then you probably want to focus on the areas which are then most under the optimal allocation and trying to bring them up to the balance.

Ben Todd: Then the third question would be also what’s your comparative advantage compared to other community members? If you’re unusually suited for one problem area, then that makes sense for you to work on that freeing up other people to work on areas that they’re better suited to. A more concrete example, within AI safety, one of the really big questions is, is powerful AI likely to be developed soon or maybe in 50 or a hundred years in the future, which we’ll call a longer timeline? From a community point of view, you’d probably want to have people allocated in working on both short timeline scenarios and long timeline scenarios, because there’s going to be diminishing returns to each type of opportunity. Then you should think, is the community currently under investing in either the short term or the long-term ones, and then working on the one that’s most under invested in? And then again also considering your comparative advantage. That could actually mean that maybe you think, in general, your guess is that AI timelines are short, say, but you actually think the community is really underinvesting in the long-term timeline scenarios. That might actually mean it’s better for you to work on the long-term ones rather than your best guess option.

Robert Wiblin: Yeah, I guess another scenario might be, let’s say that you think that you’re much better placed to work on the short term problems, but in fact you think long AI timelines is more likely, but you think that if you start working on the short term scenarios then that will cause someone else who would have done that to go and work on the long-term scenarios who will have a comparative advantage working on those. You can funge someone in a positive way out into something that’s more their comparative advantage relative to you.

Ben Todd: Exactly, and yeah, we have a whole article about comparative advantage. It can actually be a bit different from personal fit. We think of your personal fit as what you’re best at compared to, say, people in that area in general or that career path in general. But sometimes it’s actually better to do something that you’re not that good at, if you can then free up someone else who can then go and do something even better than what you would have done otherwise.

Robert Wiblin: One case where that could potentially bite is if you could imagine two academics or two people who are very good at research in the scheme of things trying to set up some new research institute. Maybe they can’t easily hire other people for whatever practical reason, but they have to raise money to fund this project it to get it off the ground to begin with, but both of them are, their strength is in research or their personal fit would seem to be in research. But nonetheless, one of them has to do fundraising. Just because of the complementarities between them, someone whose personal fit is not fundraising could end up having to go into that.

Robert Wiblin: And even further than that, it could be the case that the person who is relatively worse at fundraising out of those two should go and become the fundraiser because it could be that that would then free up the other person who’s actually better than them at research. You’re going with someone who’s not doing their personal fit and also not doing what is their absolute advantage relative to their colleague, working on this project. Which is slightly perverse, but I think actually might be a relatively common scenario when you’re coordinating with large groups.

Ben Todd: Yeah. Especially it often seems to actually come up within small teams where you’ve got a fixed number of employees in the short term.

Robert Wiblin: And a fixed number of roles that have to be done or fixed number of tasks that have to be done.

Ben Todd: But yeah, like I say, this gets pretty complicated. We have a full article that goes through all these cases and explains the circumstances under which it applies and doesn’t apply. But yeah, going back to the overall approach about what’s the ideal allocation, how can you push it in the right direction and what’s your comparative advantage? Owen Cotton-Barratt called that the portfolio perspective, and that potentially applies to how we should think about many issues faced by the community. We’ve already mentioned ‘how to allocate across AI timelines?’, another thing is ‘how to allocate across cause areas?’ But then also it could be more broad things like, should we prioritize broad longtermism or targeted longtermism as we talked about earlier?

Ben Todd: It could even apply to things like career capital, because we may want some people investing in that career capital now so they’re in a really good position in 20 years, but then other people should just be taking these short run opportunities that are really good. Yeah, you could worry, is the community underinvesting in the medium term investment things, or is it underinvesting in the immediate impact opportunities? Again, you should be thinking about, do you have unusually good opportunities in either of those two categories, and then if so, maybe you should do that and let other people take the other opportunities. Those are some examples of how your view of what’s highest impact might shift if you start to take into account more of these coordination considerations. Yeah, we also have a much longer article that goes into a bunch of other potential implications of these.

Robert Wiblin: Another consideration that shows up here is that when you’re working with a group and trying to harness all of these benefits of coordinating between people, it can pay to be extremely nice to one another, so that it’s possible for you to avoid having fights that prevent you from harnessing all of these gains. I guess especially if you want to maintain a very high level of trust. A lot of this stuff breaks down if you think that people are going to mislead one another whenever they get the opportunity to do that to benefit their project. Because then people won’t believe what one another was saying. So it makes it much harder to coordinate across different projects or for funders to believe what fundraisers are saying. Yeah, inasmuch as you think that there’s big gains from coordination, that gives a reason to have an unusually high level of integrity and honesty and probity, and also I guess to demonstrate valuing coordination and cooperation so that we can all accomplish more together.

Ben Todd: Another side of that is, if you’re working with people who really share your values, then normally it’s actually worth being very helpful to people because you might both be able to mutually benefit each other. More formally, you can make trades for mutual gain, but if other people show your values, then you don’t even care about them benefiting you back. Just helping them also furthers your own values. Because if someone else is trying to reduce existential risk and then you help them find a job, then in a sense that’s just more existential risk reduction from your point of view and you don’t really care who takes the job itself. There’s lots of strong reasons to be very helpful in daily life, but actually when you’re now part of a community that’s all got the same goal, then there’s actually even stronger reasons to be even more super helpful. This could cash out as things like doing small favors for people, giving them advice, helping them figure out their plans in an area, all these kinds of things should be maybe done to an unusually large degree. Having a good culture is also really important for getting people wanting to work on these areas and involved for the long term. It’s also just really important for actually growing the community, working well with other communities. That’s a whole other area of reasons.

Robert Wiblin: There’s a lot more that we could say about. There are many considerations that come up with coordination. But as I was saying, it’s quite technical and quite thick with considerations and modeling that comes out of that.

Ben Todd: If you ever needed technical reasons to be a nice person, we’ve got you covered.

Tips on making career plans [02:42:58]

Robert Wiblin: Yeah. Okay, so we’ve been talking about a lot of different considerations that people could bring to bear when planning out their career. Is it possible to bring this into any cohesive decision making process, a step by step process that people can use to do all these things in a way that’s not overwhelming.

Ben Todd: It’s funny you say that. We have just what you’re looking for.

Robert Wiblin: We didn’t plan this ahead of time. It’s all spontaneous.

Ben Todd: The final section of the key idea series, we cover a couple of checklists that you can work through to make your decision, and so one we call the process summary, which is just what are the steps to go through to actually make a long-term career plan and figure out your next step. Then we also have a separate process which is called the decision making process, which is when you have a couple of career options, A, B and C, and you want to choose between them specifically. In our advice in general, it’s really easy to focus on the concrete career paths that we talk about because they’re really salient, but with any career decision it’s going to be a really individual thing where you’re going to have to factor in a lot of personal circumstances, your own strengths, all those kinds of things.

Ben Todd: Unfortunately we can’t just come up with a list of the top 10 careers that everyone should do. You always have to think through your own decision, and so we try and help with that by providing these processes as well. I’m not going to go over them now as it’s better just to read through them when you’re facing a specific decision, but for now could just highlight a couple of points to particularly emphasize. I think one big thing is, it’s easy to focus a bit too much on analysis rather than concretely gaining information. Sometimes you come across people who’ve been racking their brains analyzing these options, whereas they could’ve just gone and found out something that would have immediately made the decision more obvious. One example recently was, we came across an academic who was considering whether to do a sabbatical for a year in another location, and they thought about it a bunch, but they hadn’t considered just going to visit that place for a week, which would have probably made it a lot more clear whether they would actually want to spend a whole year there.

Ben Todd: We also sometimes see the opposite mistake, where someone has just quit their job and then they’re going to think full time about their career. That often seems a bit risky unless you’ve got a very clear plan worked out for what you’re going to do with this time and how you’re going to make sure you have answers at the end. In general, we’d encourage people to do a series of cheap tests or cheap ways to gain information which go in ascending order of cost. Often initially the most useful thing is just to go and talk to people about the job, and that’s the most useful information to gain. Then later you can go onto more involved things spending a week in the location, the sabbatical example, or doing some kind of trial work, applying to the job.

Ben Todd: These are bigger tests, and then you might even want to try the job for a year or something as a much bigger commitment. The key thing is to be thinking about what are actually the key uncertainties that are driving … Actually would help me decide between option A and B, and how might I figure those out, and how might I get information that answers those uncertainties?

Ben Todd: Yeah, the other most common mistake is probably just not considering enough options, which all the decision making literature … That’s maybe the biggest single piece of advice or biggest issue that’s uncovered in that literature, is when people are faced with decisions, they can often improve the decision by just considering a wider range of options than they’re initially considering. We come across people who’ve been following our advice for a while and just haven’t considered options that seem obvious for them.

Ben Todd: One example we mentioned in our annual review was Cullen who was a lawyer at Harvard, and someone with that background, that’s a great preparation to work on AI policy, but they hadn’t really considered that. They were thinking about earning to give or maybe environmental law, a bunch of other areas, and they ended up actually making a bunch of applications in that area along with their other jobs. They got a job offer from FHI, which is now where they’re a researcher.

Ben Todd: Just a final thing on decision making. Career decisions are a really messy domain where our intuitions are not necessarily a good guide. If you want to get better at making career decisions, a lot of that involves making these really difficult predictions about the future. Like how successful might you be in different fields? How high impact might this charity be? All the best advice we’ve found on how to get better at that skill of making better predictions is Phillip Tetlock’s work. We’ve got two other podcasts with him which provide a great introduction to all that kind of stuff.

Robert Wiblin: Another comment on decision making is that every so often I encounter people who are stressing a lot about their next career decision and want to work on it right away and make lots of decisions, especially decisions that they think will have ramifications over many decades. It seems like going through all this process, learning all of this information and then figuring out how to get yourself in a different career track is going to take probably at least months, possibly years. People shouldn’t feel an enormous pressure to reassess all of these things right away. It’s a marathon, not a sprint in my mind.

Ben Todd: Yeah, I think that’s right. Yeah, just in general, these kinds of decisions can be very overwhelming, but it’s always really useful to then try to bring it down to what specific options are on the table: what actually are my key uncertainties, how might I figure those out? Yeah, that’s what we hope the decision making process helps with. It’s just a checklist or a bunch of prompts and questions that can help you at least make sure you haven’t missed anything obvious and do the best job you can of making the best decision you can, which is all you can do.

Robert Wiblin: Are there any things we should highlight about the process of applying for jobs itself?

Ben Todd: Yeah, one article in the old career guide is all about how to actually apply to jobs. That’s a little bit out of date now. I wouldn’t write it exactly the same, but I think a lot of the specific advice applies, and there’s lots of further resources and reading in there. One point I wanted to highlight is the importance of making a lot of applications. One of our team members, Howie Lempel, when he was trying to get a job at a think tank, he applied to about 30 places, including some economic research institutes, lots of think tanks, and he got turned down by almost all of them, though he got a job offer from Brookings, which is pretty much one of the most prestigious think tanks. That just shows that even someone who’s qualified to get into some of the top jobs in a sector still needs to apply really widely and might get rejected a lot of times.

Robert Wiblin: Yeah, I don’t envy people who are applying for jobs. It’s something that everyone has to go through potentially many times in their lives, but it’s a very stressful thing to do. Even if you’ve been at the top of your game, you potentially have to just apply for so many jobs, and accept the fact that you’re just going to be rejected potentially again and again and again. It could be quite demoralizing.

Ben Todd: Yeah, and sometimes, especially early in our career, we might not have gone through a process of getting rejected 30 times in a row. It’s a pretty difficult experience to go through. Yeah, a typical job application process might only have a couple of percent acceptance rate, so you can see from that that as a ballpark, you’re going to be thinking that you need to apply to, say, 20 to maybe more than 50 places to be confident of having a job.

Robert Wiblin: One thing to keep in mind is just how random the job application process is. That very often organizations, especially in these very skilled jobs, are looking for someone with very specific capabilities and the ability to get along with a very particular manager or a very particular team, where someone could be qualified in lots of ways, but then if they were randomly missing one idiosyncratic factor, then potentially they may just not be suitable for it. It’s not a condemnation of them as a person by any means.

Ben Todd: Exactly. There’s your general purpose skills and then you need the specific skills needed in that job, but then you also need some very narrow … You need the fit with the team and fit with that specific manager. Yeah, so someone who’s really talented in many ways could easily get turned down from many jobs.

Robert Wiblin: Another thing to keep in mind perhaps that might provide some comfort if someone’s applying to lots of jobs and not getting in is that if you’re not getting rejected most of the jobs you’re applying for then you’re almost certainly not being ambitious enough about the places that you’re trying to get into. Given how many hours you spend working at a job that you get, versus how many hours it takes to apply for a job, maybe it’s good to have only a few percent acceptance rate, because you should be aiming sufficiently high, or trying to be sufficiently ambitious that you should be getting turned down most of the time.

Ben Todd: Yeah, it’s a bit of a difficult question because we definitely come across people who are both. Some people are overconfident and some people are underconfident, and so we’ve found people who ended up working at great organizations in the community, but beforehand they had to be encouraged to apply. They didn’t think they would get the job, and then they ended up doing really well at it, kind of to their surprise. People like that need to be encouraged to apply more widely, though we’ve also found people who say … Recently someone was saying, my backup job is to work at an effective altruism organization, but many of those positions are actually really competitive, and that’s not going to serve as a good backup. But I agree with what you’re saying that, in general, I think the error is to not apply it enough, because as you say, there’s this asymmetry where spending several months making applications is unpleasant, but if you can get that even better role, that’s many years of having that better role. In general, the potential upsides are greater than the downsides.

Robert Wiblin: Yeah. The main reason people don’t do that, I think, is that as you’re saying, it’s unpleasant to apply for jobs which I suppose is a reason not to do it. Inasmuch as you find it unpleasant, that’s a reason to try to cut short the process a bit. But if you can find a way to grit through and apply for even more places and try to get something that’s even better and that doesn’t bother you too much, then maybe it’s worth having a crack.

Ben Todd: Yeah, in general, also try to have some stretch options where you think you maybe don’t have a good chance of getting them but you’re trying to maybe push yourself a bit in case you’re being underconfident, but also try to have a bunch of backup options to ensure that you have something solid at the end of the process. One other thing is, there is a lot you can do to increase your odds of getting a specific job, and we cover a bunch of advice on that in the article itself. But then even accounting for that, even if you can bring up your odds, it’d be pretty hard to bring them up to more than 10% per job, and so you’d still need to apply to quite a large number of places.

Ben Todd: One example of an area that’s a very tight job market where you might say it’s quite talent constrained is software engineering. But even software engineers would generally expect to interview at roughly 10 companies before they would land a position. That’s one of the most in demand jobs. Yeah, if you’re finding you’ve run out of specific options to apply for and you haven’t got enough, there’s a couple of things you could do. We’re trying to list a wider range of options on the job board. So yeah, check that out. One big thing I’d say is, you need to apply more widely than just core effective altruism organizations, because there’s just not enough jobs in there to make sure most people can make enough applications. It’s important to just consider lots of other good organizations that are in these problem areas but aren’t necessarily official effective altruism organizations.

Ben Todd: Yeah, I think a big area where there’s actually a huge number of jobs that many people could take is policy, as we mentioned earlier. I think that would be a big area to consider if you’re looking to come up with more options. Yeah, and then the other big things are just to really try and meet people in the problem area you’re interested in. That’s the biggest way that people come across good opportunities. Then, if you can’t find options with immediate impact, you can consider lots of ways to gain career capital as well, and that’s another way to get a broader range of options.

Robert Wiblin: Yeah. Is there any particular note that you’d like to end on? I suppose about how the research has come along over the years or how much people can get out of trying to piece all these different pieces of work together?

Ben Todd: Yeah, one thing I would say is, obviously all the topics we’ve considered are very complex and there’s still a lot to learn about them. But I think it’s important to not lose sight that there is concrete stuff to say about each of these issues, and actually it seems like many things you can do can increase your impact quite a lot. If you think through all the questions we’ve considered, if you can choose a more effective problem, you can choose a career path that gets you more leverage, you can find something with a slightly better personal fit, you consider the career capital trade off well, you apply good decision making advice, all of those added together could lead to really having a career that’s much higher impact than you might have been able to have before.

Ben Todd: This is the whole thing that motivates us to keep working on 80,000 Hours. I think another thing is, we’ve talked about how our advice has changed, and that can also be a bit disorienting. But partly that’s changing because the people interested in these ideas are actually responding and addressing some of these issues. We mentioned operations management. That was a path we promoted a few years ago, and then a bunch of people went and did it and made the bottleneck less pressing. It’s also really exciting to see our advice changing as people tackle these issues. Also, as far as we know right now, this is the only systematic attempt to try to start from first principles and work out which careers are highest impact, and we’re just a couple of people, so we’ve still got a lot to learn. But it’s a really exciting journey to keep learning more and discovering new considerations and finding out how to have a bigger impact.

Robert Wiblin: My guest today has been Ben Todd. Thanks for coming on the 80,000 Hours podcast, Ben.

Ben Todd: Thanks.

Rob’s outro [02:56:37]

I hope you learned a lot from that quick tour of what we know and don’t know about how to have a lot of impact with your career.

The key ideas page, which we’ll link to in the show notes, has plenty more specific advice that we didn’t manage to cover here. So don’t feel that this episode has completely substituted for reading it.

More importantly, it links to a wealth of articles that go into more detail about almost all the topics we covered here, including personal fit, career capital, accidental harm, coordination, each of our priority paths, the case for and against longtermism, and so on.

If you’re listening to this from the future, you can also go see how our advice has changed since November 2019.

The 80,000 Hours Podcast is produced by Keiran Harris. Audio mastering by Ben Cordell. Transcripts by Zakee Ulhaq.

Thanks for joining, talk to you in a week or two.