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.

Today I’m delighted to bring you a conversation between Arden Koehler, and our CEO, Ben Todd.

Ben’s been doing a bunch of research recently, and we thought it’d be interesting to hear how he’s currently thinking about a couple of different topics.

This is all very off-the-cut compared to our regular episodes, and even though he’s our CEO the things Ben says here aren’t ‘official 80,000 Hours’ positions or anything like that.

It’s just 54 minutes long — a conversation so short which I’m sure regular listeners will barely believe it — and is split into two sections.

In the first, Ben and Arden talk about four different flavours of longtermism, and their distinctive implications for people trying to do good with their careers.

In the second part they move on to things 80,000 Hours might be getting wrong, including how much weight to put on personal fit, and whether we might be highlighting the wrong problems and career paths.

Given that we’re in the same office, it’s relatively easy to record conversations between two 80k team members — so if you enjoy these types of bonus episodes, let us know at [email protected], and we might make them a more regular feature.

Just before that though I wanted to let you know that our annual user survey is now open for submissions.

Once a year for two weeks we ask all of you, our podcast listeners, article readers, advice receivers, and so on, so let us know how we’ve helped or hurt you.

You can find that survey at 80000hours.org/survey

80,000 Hours now offers many different services, and your feedback helps us figure out which programs to keep, which to cut, and which to expand.

This year we have a new section covering the podcast, asking what kind of episodes you liked the most and want to see more of, what extra resources you use, and some other questions too.

We’re always especially interested to hear ways that our work has influenced what you plan to do with your life or career, whether that impact was positive, neutral, or negative.

That might be a different focus in your existing job, or a decision to study something different or look for a new job. Alternatively, maybe you’re now planning to volunteer somewhere, or donate more, or donate to a different organisation.

Every entry you write will be lovingly attended to by a team of expert survey readers from the Nepalese highlands, who will inscribe each submission in beautiful caligraphy, sing out what you’ve written from a cliff jutting out from the Himalayan mountains, and then ceremonially burn the paper in an ancient user-feedback-related ritual.

Then, totally separately from that, they’ll be carefully read by me and my colleagues as part of our upcoming annual review, and we’ll decide on some things 80,000 Hours should do differently next year.

So please do take a moment to fill out the user survey — those who do so will have forever won my love.

You can find it at 80000hours.org/survey

Alright, without further ado, I bring you Arden and Ben.

The interview begins [00:02:54]

Arden Koehler: Hi listeners. I’m Arden. I’m a researcher at 80,000 Hours.

Ben Todd: Hi, I’m Ben, the CEO and co-founder of 80,000 Hours.

Arden Koehler: And in this episode, I’m going to pick Ben’s brain about his recent research. So Ben, you’ve been doing a bunch of research recently for 80,000 Hours, and I know a little bit about what you’ve been thinking about, but just for the audience, what are the topics you’ve been thinking about and what are some ideas you’ve had?

Ben Todd: Well, there’s a bunch which I’m going to hope to write about soon on the website, but one I’ve been particularly interested in recently is what are the different types of longtermism? What do they imply about which careers seem highest impact? And then, what might that mean for what our advice should be?

Varieties of longtermism [00:03:35]

Arden Koehler: So, we’ve talked about longtermism on the show already, but for anyone who hasn’t heard of it, do you want to just say briefly what that is?

Ben Todd: Yeah. So, just very, very roughly, it’s the idea that what most matters about our actions is their very long term effects, and by that, I don’t just mean like effects over a couple of decades, which is often what people mean when they talk about long term thinking, but we really mean maybe over thousands or even millions of years. And so, roughly, if you were going to say, “Well, how high impact is this action?” You should be thinking about the question, “What might its very long term effects be?” And that would be like the key thing to focus on.

Arden Koehler: Okay. So what are these different varieties of longtermism?

Ben Todd: Yeah. So, pretty much all the researchers who focus on longtermism agree that we should focus on what’s sometimes called path changes rather than speed ups. So, you can think of a speed up as just getting us to where we’re going to go in the long term faster. Whereas, a path change is something which changes how the long term is going to be proportionally. So, an existential risk is an example of a path change because, in that scenario, if an existential risk happens, then the value of the future is much lower for the rest of civilization’s history.

Arden Koehler: So an existential risk, I think Toby Ord defines it like actually as something that just makes the rest of the future much lower in value, so that could be not existing, whereas we might have existed otherwise, or I guess it could also just be in a really bad state for a very long time, as opposed to a neutral or good state.

Ben Todd: Yes, exactly. Whereas like a speed up might be making an important discovery like a bit earlier than it would have been made otherwise, and that’s like getting us faster progress, but not necessarily changing where things end up in the long term.

Arden Koehler: So why is faster progress good at all?

Ben Todd: Well it means if the future’s going to be better, then you’re getting that sooner, so you’re spending longer at that good state, just very simply speaking. Yeah. So the best kind of treatment of speed ups versus path changes is in Nick Beckstead’s thesis, which we linked to from our article on longtermism. But anyway, that’s not the main topic I want to talk about. So assuming that it’s these path changes, or what I call them on my last podcast, ‘hinges’, is the key thing to focus on, we can then divide up different types of longtermism by which hinges they think are most important.

Ben Todd: So if you take this longtermist perspective, basically you’re trying to find key moments when what we do today can make a difference to what happens in the very long term. So maybe like most things we do, probably don’t really have very long term effects, or at least not ones that we know, things just get muddled and washed out. But maybe there’s some moments where what we do does have these ongoing effects or makes the difference between one way the world could have gone and another way the world could’ve gone.

Ben Todd: Yeah. We’re trying to find these influential moments, and then there’s different types of longtermists, who kind of want to focus on different types of influential moment, or they have different views of what those are. And so, yeah, broadly, we can categorize the different influential moments in a couple of different ways. So one would be whether we think the influential moments, the key ones, are going to happen, say, over the course of our careers in the next couple of decades, or are they going to happen beyond that period, like further in the future? And the more you think these influential moments are going to come later, the more you want to focus on the type of longtermism, which is called patient longtermism, otherwise, you’re going to want to focus on addressing these moments that are coming up in the next couple of decades.

Arden Koehler: So this is related to the conversation we had with Will on the podcast about the hinge of history, right? Will was suggesting maybe the hinge of history… Or that means, I guess, a cluster of really influential moments or maybe just one is far in the future, we should expect it to be, whereas other people think it’s maybe in the next few decades or the next century would still be considered less patient than some other varieties.

Ben Todd: Yeah, and so Toby Ord’s recent book… And Toby, we had on the podcast, he in his book, he’s actually arguing the next couple of centuries is the time of peril, so that’s an unusually hingey moment or I guess… Well, he’s arguing in the next couple centuries is ‘The Precipice’. Sorry about that. The time of perils; that’s Parfit’s term. Though I think he may also think the next couple of decades are like even unusually important. So Toby would be slightly more on the, “We need to act now” spectrum, whereas Will was pushing a bit more into the patient end of the spectrum.

Arden Koehler: So one example of a moment that would be extremely influential, I guess, is like, if you think that catastrophic biological risks are an existential risk, and you present existential risks and you’re in charge of some sort of bioweapons program or whatever, there could be a moment even in the next few decades in your career that you might be able to act in such a way that it actually prevents this existential risk from materializing, which would make a huge difference to the entire future of humanity.

Ben Todd: Yeah. So in Will’s post, he introduces three things that can make a moment influential. One is like, all this terminology is super undecided and often quite bad, but one is the degree of ‘pivotality’.

Arden Koehler: Pivotality. Oh my gosh. Okay.

Ben Todd: So that’s just like how much scope do we actually have to change what happens? And so, an example of something that could make a time really pivotal is if we discover a new technology, such as something that could create a new bioweapon. That moment right as we’re about to discover that, that would be a really pivotal time because maybe the details of how that technology is handled could make a big difference to whether there’s an existential risk or some other shift to the future.

Arden Koehler: Okay. What are the other two? Didn’t you say there were three sorts of properties that could make something influential?

Ben Todd: So the second one is how much ability we have to do anything about it, because the time could be very pivotal, but maybe we can’t actually change what happens. So there’s a really important event, but it’s not influenceable. And then, the third one is whether we know what those things are. So like, again, there might be these amazing opportunities around, so you might have the first two things satisfied. But if we can’t identify them, then it would also not be an influential time in Will’s definition.

Arden Koehler: Okay. So patient longtermists think that events or times or combinations of events and the state of humanity at that time where we’re actually going to be able to make a huge difference because of these three properties holding, is probably not going to happen for an extremely long time. Is that the idea?

Ben Todd: Yes. Or just now is not a special time. Like maybe it’s just as influential as other times that will come up in the next few hundred years.

Arden Koehler: Okay. That’s a really helpful clarification. So you could think that right now is, in some sense, there’s lots of existential risks right now, but if you thought that there was going to be even more existential risk in a couple hundred years or a thousand years or the same amount, you might still be a patient longtermist.

Ben Todd: Yes. Or you might think that we’ll understand them better in the future, and so we’ll be better able to deal with them.

Arden Koehler: Okay. So that’s patient longtermism. And impatient longtermism, does that have a name?

Ben Todd: Urgent longtermism.

Arden Koehler: Urgent longtermism. That sounds better.

Ben Todd: But then urgent longtermism then divides into several other forms. So, we’ll set patient longtermism aside and then just focus on the types of urgent longtermists. Okay. So then the next key question is, so we’ve now decided that there are these important influential moments in the next few decades, the next question is, do we actually know what they are? And so, you might think now is an unusually important time, but we’re very unsure about what are the particular influential things. And so that term of longtermism I call broad longtermism. And so, yeah, this is like a slightly odd position, but you can almost see Toby, in his book, is going towards this position. If you put AI to one side, I think he actually thinks… I think it’s either the second or third biggest existential risk is from an unknown risk. And so he’s relatively on the urgent end of the spectrum, but he also thinks there’s quite a good chance that we don’t actually know what the key thing is.

Arden Koehler: Okay. So one thing that feels a little confusing here is that you said before that something that makes something an influential moment is that we have enough knowledge to be able to take these opportunities, so I guess I’m thinking if you thought we didn’t have knowledge of what the opportunities were, then you should be a patient longtermist even if you think there are those opportunities in the near future. So how can somebody be a broad longtermist but also an urgent one?

Ben Todd: Well, you could think now is a pivotal time and we might be able to do something about it. You don’t have the third one, the knowledge, but those first two could still be enough to make you quite urgent. But yeah, I agree that the knowledge thing is pushing you maybe towards patience, but maybe the other two factors could outweigh.

Arden Koehler: I guess you could also think we don’t, in fact, know right now what the influential moments are, but you could think, “Well, when they come up, we might be able to recognize them.” So as long as we have people standing guard in the next 100 years or something, they’ll be recognizable when they come up, even if we can’t tell what they are now.

Ben Todd: Yes. Though that sounds a bit more like patient longtermism, though. Yeah. I mean, maybe it’s worth saying like all of these things are a spectrum. So with broad longtermism, what people then tend to focus on is… So, suppose one type of broad longtermism is you believe that now is an unusually risky time in terms of existential risk, but you’re not sure what the biggest risks are. Then what you can do about that is you can focus on reducing risk factors, so just things that generally increase existential risk, rather than trying to reduce specific risks.

Arden Koehler: Okay. And that’s things like trying to increase international cooperation or something, seems like robustly good for lots of different things.

Ben Todd: Yes. I think Toby mentions great power conflict in the book as potentially an important risk factor because it seems like many existential risks can be prevented if people can coordinate and work together, and it seems like one of the most likely ways that coordination isn’t possible is because there’s a war between important powers. So if we knew there was going to be no great power conflict, then we should probably think existential risk would also be a bunch lower over the rest of our careers.

Arden Koehler: So that’s broad longtermism, and then I guess the opposite would be narrow longtermism, or not the opposite, but the–

Ben Todd: Yes. Though I’ve been calling that targeted longtermism.

Arden Koehler: Targeted longtermism.

Ben Todd: Yeah. So that would be the idea that acting is urgent and we know what the key influential moments are. And that gets us on to, we can then divide targeted longtermism into two types, again. One where we should focus on existential risks as the main thing, and then the other is we should focus on some other type of path change. And those are like two broad types of influential moment, is an existential risk or a path change.

Arden Koehler: Can we also divide broad longtermism into the existential risk focused broad longtermism and other kind of path change focused type?

Ben Todd: So I’ve defined broad longtermism as we don’t know what the influential moments are, so it could be any of those.

Arden Koehler: I see. Yeah. Okay. So what’s an example of another kind of path change that’s not an existential risk — path change moment?

Ben Todd: Of the four positions, that one’s probably the least popular, but the thing that actually seems to be most supported, within that would be another AI focused longtermism. But some people think that the biggest reason to focus on AI is not because an AI disaster could be an existential risk, but rather because exactly how AI is designed might have some other kind of very long term effects. It might lock in a certain set of values or just change how civilization is for a very long time, and it might just change that in a proportional way rather than an all or nothing way.

Arden Koehler: So how we deal with AI could make the difference between having an okay future and an awesome future?

Ben Todd: Yeah. Or maybe you could even just think it might be possible that if it’s handled in one way, you get like 80% of the value, and if you handle it slightly worse, you get 79% of the value. And maybe there’s just like a whole load of gradations of how well it’s handled overall.

Arden Koehler: Okay, cool. Is that four types then of longtermism?

Ben Todd: Yes.

Arden Koehler: Okay. Can I ask you which one you’re most sympathetic to?

Ben Todd: I don’t know if we should recap the types, but we’ve got patient longtermism, then we’ve got urgent broad longtermism, urgent targeted longtermism, and then that… Well, then that divides existential risk focused longtermism and path change or trajectory change focus longtermism.

Arden Koehler: Okay. Actually, let me just, just because taxonomies are fun, let me ask like, so can patient longtermism be divided into broad and narrow? Could you think, “Well, the influential moments won’t be for a really long time, but we’re pretty confident that they’re going to be coming from brand new…” Okay, maybe this is like super implausible.

Ben Todd: Yeah. Yeah. No, I think it’s not super plausible, but you could imagine that there will be a further spectrum in there about how much knowledge we have about what these influential moments are going to be, and that could influence exactly what type of capacity building you want to do as a patient longtermist

Arden Koehler: Okay. But usually it’s going to be more plausible to be a broad patient longtermist. If it’s not going to be for a long time, probably we don’t have a great idea of what we’re going to solve.

Ben Todd: Yeah. So I’m not keen on that terminology because I think the most common way people confuse these different types of longtermism is between broad versus patient, and so they actually have pretty different implications because broad longtermists think… So sorry, I’ve defined broad longtermists as like urgent focused ones, so they want us to do stuff which immediately reduces existential risks or otherwise helps society better navigate whatever the influential moment is going to be in the next few decades. Whereas, instead, patient longtermists generally want to build capacity to empower future altruists to take on these influential moments when they come up. Yeah. Both of them will be keen on global priorities research. Probably the patient longtermists are going to be more keen on it because… Well, actually, that’s not clear, but yeah.

Arden Koehler: Maybe research takes a long time to give you returns.

Ben Todd: Yes. If you’re like super urgent focused, then it’s going to stop becoming attractive to do research because it’s quite a long-term investment.

Arden Koehler: Okay. So now that you’ve put off the question of which one you are most sympathetic to–

Ben Todd: Yeah.

Arden Koehler: Do you have a sense?

Ben Todd: Well, so I guess, for me, it’d be mainly between existential risk focused longtermism and patient longtermism. And then I put like a little bit of credence on the trajectory change one and the broad one. So, actually, overall, I’m a bit more focused on the urgent longtermism than patient longtermism but yeah, like still patient longtermism would be my second view out of the four.

Arden Koehler: And is that because you think the arguments that Will was making seem somewhat persuasive, but you’re not sure?

Ben Todd: Yeah. And I mean, also, Phil Trammell’s arguments for weighting seem pretty powerful in some ways.

Arden Koehler: Yeah. So we had Phil Trammell on the podcast, but just to very briefly say, that was more focused on how much we’re going to have power to affect these moments. Phil pointed out that if we invest resources now to grow them, meaning financial but also other kinds of resources, we could have a lot more influence in the future, which would push you toward patient longtermism.

Ben Todd: Yeah. I mean, I think, also, our understanding improving over time seems fairly convincing. And then it also seems like if you’re uncertain between the two views, the patient one maybe keeps your options open more, though–

Arden Koehler: Interesting. I would have thought… I mean, I feel like the intuitive thing is if you’re uncertain between patient and urgent, you should go for the urgent one because it seems like a bigger mistake or something to like–

Ben Todd: Yeah. I guess you’re right. I mean, I’m pretty unsure about which way that goes. But yeah, obviously, if there’s like an existential risk now and we could have prevented it, then that’s like the ultimate loss of optionality. I suppose there are other arguments where uncertainty puts you in favor of, so there’s like… Phil, I think, talks about Weitzman’s argument about discount rates where if you’re uncertain about what the discount rate should be, generally, that will push you towards using lower ones over time.

Arden Koehler: Interesting. I don’t remember why that was. Is it easy to explain?

Ben Todd: Not super easy, but if you… Well, I don’t know how many people will know what a discount rate is, but yeah, if you imagined like… Okay, so either we have a high discount rate, which means things in the future are not very important compared to things now, or we have a low discount rate, which means things in the future are very important compared to now. If you imagine taking the average of those two, you end up with things in the future being pretty important, and then if you do this more mathematically, it actually ends up that you should, as you get further and further in the future, your expectation value for the discount rate should tend towards the lowest conceivable value for the discount rate.

Arden Koehler: Interesting. Yeah. I don’t feel like I follow that last step, but we’ll just link in the show notes to where Phil explained this you said?

Ben Todd: Yeah. I think so, but it’s a famous paper by Weitzman, an economist who–

Arden Koehler: Okay. We’ll link to that in the show notes, too.

Implications for different types of longtermists [00:20:16]

Arden Koehler: Okay. So those are the four varieties of longtermism. So what’s the implication of these distinctions for what 80,000 Hours should be doing as an organization to help people do good with their careers?

Ben Todd: Yeah. So all the different longtermists basically agree that we should want some or all of these kinds of things, and so there should be some kind of portfolio of effort across them. And where people are more differing is just in exactly how much effort it would be ideal to go into each one. And we would take a similar approach in our advice, and among our users, we think it would be cool if people were pursuing all of these different types of longtermism.

Arden Koehler: Yeah. So can you say more about what it means to pursue the different types of longtermism?

Ben Todd: Yeah. So given each type of longtermism, you can then think about what priorities that would imply about cause selection and career capital and movement building. Those are the main differences. So yeah, I guess to give an example, the thing we have most clear on our key ideas page is the existential risk focused longtermism, and so there, we think we want people basically focusing on reducing the existential risks that are biggest, most neglected, and most solvable. And so there, we’ve tended to focus on AI safety, biorisks, nuclear security, and climate tail risks as the four key things to work on.

Arden Koehler: So if you were a patient longtermist, or you thought that, in fact, other kinds of path changes were more available to us, then you might focus on other priorities than the things we emphasize the most?

Ben Todd: Yeah. So the existential risk focused longtermist would focus on reducing those risks I just mentioned. They would also want to reduce risk factors as well. But then, if you’re a patient longtermist… Well, in practice patient longtermists would agree we would want still to invest a little bit in those things, but mostly we’d want to be building capacity to help people in the future do more good. And so then, that would tend to come down to some kind of movement building, because that seems to be one of the best ways to get there to be more people caring about these things in the future.

Arden Koehler: And this is EA movement, or the effective altruism movement building in particular or–

Ben Todd: Yeah. It could be–

Arden Koehler: … Or is it supposed to be broader?

Ben Todd: Yeah. It could be any type of movement building that gets people interested in whatever the longtermist priorities will be in the future, in a very abstract level. But yeah, effective altruism movement building is like one that seems to be working. But yeah, I would wonder in the future whether we might want something like movements concerned with political representation of future generations, and that could be another independent movement that might be really good from a longtermist perspective.

Arden Koehler: Although, actually, I mean, it strikes me that if I’m a patient longtermist, then shouldn’t I think that a movement built around increasing representation for future generations is not going to be focused on grabbing the influential moments when they come up? Don’t I want a movement of people who are like scanning around all the time and looking for those influential moments and then grabbing them and acting when they come about?

Ben Todd: Yes, that’s kind of ideally what you want, but then you could imagine something like that broader thing might just get lots of people interested in the narrower thing. So patient longtermists might be focused on movement building. Yeah. We just released a blog post, actually, which has a quick summary of the types of things patient longtermists want to focus on. Another big one would be global priorities research to try and figure out what those influential moments are going to be. Yeah. And so another thing is patient longtermists would be more keen on people investing in their career capital. So yeah, by career capital, we mean skills, connections, reputation, credentials that put you in a better position to make an impact in the longterm. So it’s like how much you’ve invested in building your skills and abilities.

Ben Todd: Yeah. And so patient longtermists… Well all the different views will want people to focus on gaining career capital to some degree, unless you’re very, very urgent focused. So if you think the key influential moment is coming next year, then there would be no point in doing a PhD now. You should just do whatever you can do to help. But if you think it’s coming over a few decades, you probably would want to get some career capital if it can pay off within those few decades. But then, yeah, a patient longtermist would be the most keen on career capital because they’re really happy to wait for as long as needed to have more impact.

Arden Koehler: So one thing that feels a little confusing about that to me is the career capital question is like, “Okay, what should I do for the next couple of decades?” Whereas the thing that patient and urgent longtermists disagree on is, “How important is the next couple of centuries or even millennia?” So how much do they really differ on this question of career capital?

Ben Todd: Yeah. So I think it’s like maybe a bit less than it might first seem, but some of the more urgent focused longtermists do really think the next one or two decades is pretty unusually important. And so that effectively makes your discount rate a few percent higher probably, and so that increases the hurdle rates for getting career capital a little bit. And so, yeah, it wouldn’t make a huge difference, but if someone was on the fence about, “Should I do a Masters? Should I do a PhD?”, or some other thing they think won’t have impact but will get them some extra career capital, it could make the difference at marginal cases, and patient people would default towards the career capital and…

Arden Koehler: So it seems like if you thought the most influential moment or moments were going to be coming around the end of the century, it seems like you’d still count as an urgent longtermist. I mean, obviously, there are gradations, but you’re closer to the urgent longtermist than the patient longtermist, but you might still have the same view on career capital as the patient longtermist?

Ben Todd: Yeah, exactly. So another way of seeing how urgent or patient to be is in terms of what fraction of our resources we’d ideally give each year? And so this is easiest to see with money. So you can imagine like, “Well, we have a pot of money. How much should you spend each year?”

Arden Koehler: On doing good?

Ben Todd: Doing good. Yeah.

Arden Koehler: Yeah.

Ben Todd: So if you think the next 100 years is unusually influential, roughly you’ll want to spread out your money over that 100 year period, and so that will… Well, in Phil’s model, that would mean that you would give a couple of percent a year, something like that. And that’s not actually a very different answer compared to a patient longtermist. The patient longtermist would maybe give a percent less, and that’s because for the urgent person, they’re still spreading out over 100 years, so it’s still very spread out. But if you were an urgent longtermist who thought the next decade is unusually influential, then in Phil’s model, you would start giving something more like 10% a year, so that would be a really big difference. Yeah, it’s only once you get to these really short timelines that it really starts to make a big difference to how much we should be giving now versus investing, and then the same with how much we should be investing in career capital, but rather just trying to do the best we can right now.

Arden Koehler: And just to clarify, when we talk about spending resources or giving resources, we mean on things besides things like movement building, right? That don’t really count as investments of any kind.

Ben Todd: I categorize movement building as an investment.

Arden Koehler: Right. That’s what I’m saying. So you could spend 50% of your resources by funding programs, but if they’re mostly focused on things like movement building, that wouldn’t count as spending 50% of your resources on this model.

Ben Todd: Yes. It’s how much is going on object-level things.

Arden Koehler: Yeah. So just to make this really concrete, if a listener is hearing this and saying, “Yeah, I was really convinced by some of the arguments that I’d heard before for patient longtermism”, what should they maybe do differently because of that?

Ben Todd: Yeah. So career options for patient longtermists, perhaps the best… Well, one of the best things is any type of movement building that would pay off from that perspective. Second thing would be anything that helps with global priorities research. And then, a third category is earning to save, so basically earning to give, but then you would put the money into some kind of foundation or donor advised fund and try and invest that and have more money in the future to give. You can pursue that to more or less degrees. Like one could just be like you donated it all at the end of your life, or another could be you try and put it in some kind of trust that will donate it after you die. And then a fourth category is you could get any career capital that will help with those three priorities. So then that would be like career capital that helps you to do global priorities research, or that does movement building, or that helps you do earning to save.

Arden Koehler: So that would be like something that puts you into a position to make a lot of money?

Ben Todd: Yes.

Arden Koehler: In that last category.

Ben Todd: Yes. Another thing I think people sometimes mix up is they think that patient longtermists should be keen on really transferable career capital. And so, you can divide career capital into specialist and transferable. Specialist is like career capital that’s useful for just like a narrow range of paths, and then transferable is like useful in many different careers. Patient longtermists, I used to think as well, that they should probably focus on transferrable career capital because they’re not sure what the key moments are. But actually, they should focus on getting career capital that helps with those three priorities, but it could be specialist career career. The people who might be more tempted to focus on transferable career capital would be the broad longtermists because they might take a strategy of just doing something transferable and then trying to spot that key influential moment that will come up over their lifetime and work on that.

Arden Koehler: Although I guess you could also think, if you’re a broad longtermist, maybe you should focus on international relations. I know I keep using this example, but you could get really narrow career capital for working to help reduce the chance of great power conflict.

Ben Todd: Yeah, totally. So I think even in the broad longtermist case, there could be a strong case for fixing on specialist career capital where you would bet on some kind of broad intervention that you thought would help with many different potential risks or other types of influential moments for the long-term future.

Arden Koehler: Okay. So I want to move on here pretty soon to talk about your views on career capital and other research topics. But before we do that, is there anything else you want to say about the varieties of longtermism?

Ben Todd: Yeah. I mean, one interesting question is what are the biggest priorities within those four things right now and where the gaps are. And yeah, I tried to think a bit about, are we getting the patience versus urgency tradeoff right across the community and among our readers? And actually, I think that’s pretty unclear because, effectively, we are saving a lot of resources for the future, because lots of our users are young so they’re investing in career capital, and Open Philanthropy is saving lots of their money.

Arden Koehler: Open Philanthropy is our biggest funder.

Ben Todd: And it’s like the biggest effective altruism aligned foundation. Yeah. I mean, I think maybe we neglect the patient perspective a little bit, but overall, it’s pretty unclear. Yeah, and then within the more urgent forms of longtermism, what might be being neglected; I think most effort is going into the existential risk focused one and I think that makes sense. But I think there’s currently maybe a bit of a gap in the broad focused longtermism. One reason for that is it seems like most people don’t think that should be our biggest priority, but lots of longtermist people think maybe it would be reasonable to say like, “Put 10% of our resources into that.” But it seems like, currently, it’s much less than a couple of percent of people or money being spent on those style of things.

Arden Koehler: And that’s within the effective altruism community?

Ben Todd: Yes.

Arden Koehler: So, I mean, what if somebody said, “Okay, but in fact, the broad longtermist interventions are the things that most people outside of the effective altruism community focus on”, or like… Sorry, not most people, but a greater number of people. So that’s things like maybe increasing economic growth or trying to reduce the chance of conflict, which seemed a bit less weird and neglected than like some of the things.

Ben Todd: Yeah. The economic growth question is how that fits in is a bit of a complicated one.

Arden Koehler: Oh, okay.

Ben Todd: In some ways, economic growth is a speed up, so it’s not even one of these four varieties of longtermism. But then, in practice, some people think, “Well, if we have more economic growth, maybe we’d get through the time of perils faster”, and so it’s reducing existential risk.

Arden Koehler: I guess I was also thinking we just have better tools and better education or something like that at the point when something happens so maybe we’re a bit robustly better at dealing with it.

Ben Todd: Yes. But then, you have to set that against, if we have more technology, then maybe we have more risks in some ways as well. And then you have to quantitatively model the two different effects. And so, there’s this paper by Leopold that, I guess, we’ve mentioned before on the podcast, where he tries to make a quantitative model of those and he ends up including economic growth as overall probably going to be positive for reducing existential risk. And so that then could be a thing that more urgent longtermists would want, but maybe probably not as that top priority because it doesn’t seem like the very most leveraged way to reduce existential risk.

Arden Koehler: Okay. Okay. Well putting economic growth aside then, I guess–

Ben Todd: Yeah. Something like great power conflict would be a way of reducing risk factors and seems not as especially neglected. Again, this is maybe not super neglected, but we’ve had a previous episode with Glen Weyl about how might we design institutions to provide global public goods better. And being able to do that would probably reduce lots of other existential risks, so it’s an example of a risk reducing thing. If you might be interested in some of these broader longtermist things, then many of the things we cover in our new list of other problems, some of them are in that category, so we should link to those in the show notes.

Arden Koehler: Cool. All right, we’ll do that.

Things 80,000 Hours might be getting wrong [00:33:07]

Arden Koehler: Great. So let’s move on to some other research that you’ve been working on. So you’ve been thinking about ways that our advice to readers could be improved. So what are some of the areas where you think there might be the most room for improvement in 80,000 Hours’ advice?

Ben Todd: Yeah. So the last few months I’ve been trying to think about just very big picture, what might be the mistakes in our advice, or things we’re getting wrong. I think like the previous topic could be in that category. Maybe we should focus a bit more on broad longtermism, maybe a bit more on patient longtermism. But another one which we’re just getting onto is how to trade-off transferable versus specialist career capital and I think what you think about that trade off can have a really big effect on which careers seem best. Because if you think transferable career capital is where it’s at, then you should be focusing on just getting very general skills like management. Maybe you should just be really focusing on just being successful at something because it seems like being successful gives you lots of options in the future.

Ben Todd: It helps you get cool achievements and it helps you meet other high achieving people and often people who’ve done something impressive in a field can have lots of options in the future. Or it could mean things like becoming a journalist or maybe working in policy because these are platforms that let you support many different issues in the future. And so, if that approach is correct, then maybe our advice would be more just like, “Okay, just go and become really good at something”, and that would be the key message. Whereas if we think specialist career capital is correct, then instead, we should be trying to think, “Well, what are these highest impact things over the next couple of decades?” And then we should be trying to take bets on building career capital that’s really relevant for those. So maybe if we say, think like, AI technical safety is really important priority over the next few decades, then maybe you want to go and study machine learning and get skills that are really useful in that particular path.

Arden Koehler: So is the main thing that makes the difference between these two views or places is to emphasize how uncertain you are about what kinds of work is actually going to be most valuable in, say, 20 years?

Ben Todd: Yeah. So I think one really big dimension is just how uncertain you are about the priorities, which could be priorities about which problems are most pressing. And remember, again, this is priorities over the next couple of decades because that’s the kind of planning horizon for careers. So that does seem pretty uncertain, what will be the top priority in 20 or 30 years. But it could also be uncertainty about your own preferences and personal… Well, some types of transferable career capital, if you were very uncertain about what types of organizations you want to work in, then again, that might make you want to just get a transferable skill set which you could then use in many different organizations in the future.

Arden Koehler: Okay. So that would exclude transferable career capital. It’s just like, become extremely good at one narrow thing…

Ben Todd: That can be used in many different ways.

Arden Koehler: That can be used in many ways, because that’s still–

Ben Todd: You’re still focused on a particular skill set.

Arden Koehler: Right. But you might study management or something like that.

Ben Todd: Yes.

Arden Koehler: Okay. Is there anything else that influences how much you should focus on specialist versus transferable career capital besides uncertainty about what kind of work is most important and your own preferences?

Ben Todd: Yes. So I think the other big side of the equation is just how easy it is to get both of these forms of career capital. So one way of thinking about it is, suppose you got the transferable career capital, and then in 20 years there’s two possibilities. One is that the best guess of which priorities were most pressing that you had now is actually correct and so then you use that transferable career capital to work on those. Or there’s a new cause that’s arrived, which we sometimes call cause X and then you use that transferable career capital to work on that cause in 20 years time.

Ben Todd: And then we then want to contrast that with instead you could have just betted on whichever priorities seem most pressing now, and then you’ve built specialist career council relevant to those priorities. So then the question is, if you did the specialist path, how much of a better position are you in to work on the priorities compared to if you’d done the transferable route, if it’s still the same priorities? Or we’ve got the other scenario where it’s cause X, in which case your specialist career capital is presumably less useful than having the transferable career capital.

Ben Todd: So you want to think about what’s the delta in these two different scenarios between the two different things you could have done, which I now just realized is really hard to follow.

Arden Koehler: Well, okay. Let me try to summarize it; something like, so you want to know two things, 1) How… Maybe I’m just going to put it in an even more confusing way, how transferable is specialist career capital? So if it turns out that you’re wrong about what the best kind of work is for you, how much can you still make that transition? Because you became an amazing machine learning researcher or even just a somewhat good machine learning researcher. And then the other question is, how useful for the things that you like would be your best guess, are the most–

Ben Todd: Is the transferable career capital.

Arden Koehler: Yeah, is the transferable career capital.

Ben Todd: Yes. Also another thing is, how good is cause X compared to those other priorities in the future? A new cause X could arrive, but maybe it’s only a little bit better in which case it doesn’t really matter if you can’t switch into it.

Arden Koehler: And by better, of course, we mean worse because oftentimes causes are bad things in the world that need fixing.

Ben Todd: We mean how pressing it is, yeah.

Arden Koehler: How much better it is to work on this; to work on cause X.

Ben Todd: How much more value you have per year of work on the cause.

Arden Koehler: Okay, yeah. Great. So have your views shifted at all on those questions over the years?

Ben Todd: Well, so I mainly see this as just a big uncertainty because I still feel pretty unsure about, in general, whether people should prefer specialist or transferable career capital and it also has a big effect on what our recommendations should be. Where the specialist one would be like, “Well, there’s these priority parts, just do those.” And the transferable one would be like, “Just be successful at something generally impressive or generally useful.” And that seems like two pretty different emphasis.

Arden Koehler: So on that, for the second one, that could also mean just to clarify, it could mean, just do whatever you’re most passionate about or something or best at.

Ben Todd: Yeah. I mean, I would focus on best at rather than passion, but passion is relevant insofar as it’s making you more likely to do well at something.

Arden Koehler: Right, okay. So I guess some of this uncertainty is on the user’s side, because some of it is about what they in fact will be good at or something. And then some of the uncertainty is on our side, of course, I mean, we want our users to also help us think through these things in some ways or do some thinking for themselves. But questions about, how likely is it really that AI presents a big existential risk?

Ben Todd: Yes, totally. And maybe in terms of what people should do, I think either of these two strategies can make sense and a lot of it will come down to the particular opportunities that are open to you and how good a fit you seem for them.

Arden Koehler: Yeah, or if they disagree with us. If they think, “Oh, they’re wrong about these things being the most important issues to work on.” They might decide to go a more transferable route.

Ben Todd: Yeah. Or maybe they should take the specialist route, focus on whatever they think the priorities are.

Arden Koehler: Yeah, right. Okay, so what else is there to say about specialist versus transferable career capital?

Ben Todd: Yeah. Just guessing which one seems best, it does seem like a lot of the highest impact routes, you do need to get some specialist career capital eventually and there seems to be pretty big gains from that. So it is definitely worth considering specializing, though one thing is if you’re unsure between the two or if you need to get some transferable career capital and some specialist career capital, you should do the transferable one first because then you can still switch.

Arden Koehler: Oh yeah, okay. Keep your options open.

Ben Todd: Yeah. But I mean, in general, because the uncertainty pushes towards transferable.

Arden Koehler: And presumably we’re going to be less, I mean, everyone will be a little bit less uncertain in the future, if we are doing a good job at trying to figure things out.

Ben Todd: Well, yeah. Then you’re just really getting into almost the general patience versus urgency debate.

Arden Koehler: Oh, interesting. Okay. So I guess that’s one way that these two debates are linked.

Ben Todd: Yes. I think another interesting framing is what would be best from a community perspective. And if we think in 10 years, what do we want effective altruism to look like? Having 10,000 McKinsey consultants doesn’t seem as useful as having 10,000 experts in loads of potentially different interesting areas. We definitely want a bunch of specialists. To be a bit more serious, we would want some of both. But I think thinking about it from a community perspective makes it more attractive to have people bet on lots of different, interesting, potentially relevant things.

Arden Koehler: Right. So, does it seem accurate to say that we seem pretty pro-specialist career capital in our content? So we could be making a mistake by being as pro-specialist career capital as we are.

Ben Todd: Yeah. Most of the priority paths are on the specialist end of the spectrum. And so, I could imagine maybe we should put a bit more emphasis on the transferable career capital. Though maybe just worth clarifying, as I did in my key ideas podcast as like, I think when people think of transferable career capital, the first thing that comes to mind is something like consulting. But actually it could be something much more like just going to become a great civil servant because civil servants can work on many different problems in the future. Or we mentioned the example of a journalist earlier as someone who actually has very transferable career capital with respect to causes. And so it’s those kinds of things I’m most excited about.

Arden Koehler: Yeah. I wonder, if you’re really unsure about the transferable versus specialist trade off, maybe there are some kinds of career capital that seem to be both pretty good for if you are pretty sure that one problem is the thing that you want to be working on, but also more transferable than some other kinds of career capital.

Ben Todd: Yeah. I mean, I guess one big thing is, if you can focus on something specialist but you have good backup options, then you’re doing something that’s pretty good on both ends of the spectrum. And I think a lot of our priority paths do have good backup options, which is one reason why we haven’t been as worried about recommending them despite a lot of uncertainty about which causes are going to be best in the future.

Arden Koehler: That makes sense.

Ben Todd: Also, we’ve now released this longer list of other interesting career paths and a lot of them are these slightly more transferable things. Those are now on the key ideas page as well.

How much weight to put on personal fit [00:43:21]

Arden Koehler: Cool. Okay, so are there any other areas where you worry that maybe 80,000 Hours could be getting the balance wrong or places where you think we’re most likely to be mistaken?

Ben Todd: Yeah. I think there’s a broad category of things around… One way you could see it is how much weight to put on personal fit. And so something I worry about-

Arden Koehler: Can you just define personal fit for us?

Ben Todd: Yeah. Loosely, personal fit is just how good do you expect you’ll be at a certain career path. But often we’re particularly interested in what’s the chance of having like a really outsized success in a particular path.

Arden Koehler: Okay. And that’s compared to other people in the path, right?

Ben Todd: Yeah, if you want to get more formal about it.

Arden Koehler: Okay.

Ben Todd: Or the average person. So something that I would worry about is, well, we’re making these particular lists of career paths and global problems quite salient to people and so I worry about someone who goes into one of those things and then becomes pretty good, but not amazing, when they could have been amazing at some other thing that we haven’t listed or maybe just one of our lower priority things that we think are just, on average, a bit less pressing but would have actually been an amazing option for this particular person.

Arden Koehler: So is the worry then that because we aren’t emphasizing thinking about where our readers, in particular, can excel enough they will end up doing something that they excel less at. And then, in fact, that’s actually worse for the world because they would have had more impact doing something they were much better at even if it was a less pressing area in general.

Ben Todd: Yeah. So, for instance, we tend to slightly push people towards studying economics because, all else equal, having an economics PhD is really good for global priorities research and policy and a bunch of other things we’re interested in. But then, you can imagine someone who stretches to do economics and they just about succeed but actually they would have just been amazing at psychology because just their skills slightly matched it more or they would have turned out more interested in it. And we don’t emphasise psychology quite as much as economics. And it would be better to have someone who’s right at the top of psychology and do something really innovative and interesting compared to someone who’s a bit above average but not amazing at economics.

Arden Koehler: Yeah. So I guess it’s all kind of a matter of degree, right? Because if somebody was just a little bit more above average in psychology versus a bit above average in economics, we might still think it’s better to do the economics and it just depends on how much better you’re going to be in that other path.

Ben Todd: Yeah, exactly.

Arden Koehler: So yeah, I mean, I might have just answered the question I’m going to ask, but, so why can’t we just emphasize personal fit more in the content and solve this problem?

Ben Todd: Yeah. Well, so we do try to emphasize it quite a bit, but just because the particular list of things we say are so much more concrete and salient, it’s very easy to focus on those too much and not think enough about your individual circumstances. And we’re going to try and add more content to the key ideas series on the website about how to think through your individual situation to try and help with this problem.

Ben Todd: But you can kind of imagine normal careers advice is the opposite to us, where it would just start with like, “Okay, what are your strengths?” That would be the number one question and then, well, that would be it actually. But then with us, we kind start from, well, what does the world most need? Because I think that’s a really neglected perspective and also, well, if you want to have a big impact, it’s really important to be actually thinking about what will actually help people. And then we work back from there and think, now, how can my strengths fit into that? So we do cover both ends of the spectrum, but we tend to emphasize what the world most needs more.

Arden Koehler: Yeah. I wonder if in fact it’s fine, because maybe people are more naturally inclined to think in terms of their personal circumstances and what they’re good at because they’ve been being told since they were 11 that that’s the way to choose your career. And emphasizing what the world needs even more than is actually justified, if all else is equal, is maybe a good corrective.

Ben Todd: Yeah. That’s how I mainly feel about it. But I guess I worry that, I mean, people who are really into 80,000 Hours and then maybe hear the perspective too much.

Arden Koehler: Yeah.

Ben Todd: But hopefully that’s not like the typical reader, but–

Arden Koehler: Hopefully the typical reader isn’t into us very much.

Ben Todd: Yeah. No, I mean it’s, in general, this is a very difficult challenge of writing for multiple audiences, and tends to be people who’ve been listening for a long time need to hear like pretty different things from what a new person would ideally hear.

Arden Koehler: Okay. So maybe we’re not emphasizing personal fit enough. Is there anything that you would recommend to readers to either read of ours, or somebody else’s, or to do that would help them think better about their personal fit?

Ben Todd: Yeah. Well, one answer is, I’m working on a how to plan your career process, which we’re going to put onto the website. And then the aim of that would be to actually lead you through everything you need to think about, including your personal fit. And so hopefully if you work through that process it will mean you’ve significantly thought about personal fit. Another thing is just whenever you’re making a career decision, obviously, do consider personal fit and bear in mind that it could outweigh which things seem most pressing in general. It could easily be that it’s better to do something that you’re good at, especially if you’re focused on this more transferable career capital focus thing that we were just talking about. And then I mean the third thing is, I think it is worth people at some point reflecting on their strengths and trying to really clarify what they are.

Ben Todd: And there’s a lot that could be said about how to figure out what your strengths are and also just how to predict what you’re going to be good at. That’s a whole difficult question in forecasting. But one perspective that I think gets a little bit neglected even in the mainstream talking about what your strengths are and what you’re going to be good at, is not only thinking about broad strokes: the area, what might be a good fit, but also really thinking about the nitty gritty day to day of the particular jobs. And trying to build up a picture of what does this job actually involve day-to-day and could I see myself doing that and being motivated in it.

Ben Todd: And one exercise I found personally useful, it’s got a slightly cheesy name, of ‘The Energy Audit’. But basically what you do is you look at the last two weeks in your calendar and then you try and categorize things by energizing or not energizing. And then you try and think about how can you do the energizing ones more often. And you can use that within your role, to try to design your role a bit better, but you can also use that to try to understand what actually are your strengths. What are the types of activities, people to work with, skills that you find most energizing, which is often a good sign of something that you might have good fit with.

Arden Koehler: I guess the other thing to say here is that oftentimes people can try the work or at least talk to somebody in the domain so that they can get a better sense of this day to day and a better sense of their fit with it. And I mean, this is a lot more applicable when people are early in their careers and they can spend the summer doing an internship. And this is of course common sense, but it’s one thing to maybe emphasize.

Ben Todd: Yeah, we could have a huge discussion about what are the best ways to predict personal fit. And one perspective is trying to predict it from the armchair and then you’ve got a whole question of forecasting and which predictor is the most powerful. And then another perspective is how can you get more information and test things out and learn about them. And I think often people focus quite a bit on the armchair stuff and often just by going to talk to someone in their career, you can learn so much that’s obviously really useful, that that’s a really good use of time.

Arden Koehler: Yeah. So I just listened to “How to Measure Anything” by Douglas Hubbard and one thing he hammers home is the more uncertain you are, the actually easier it is to reduce your uncertainty. If you know practically nothing about an area, just talking to somebody for half an hour is probably going to be a huge deal for making you understand what that path is like. So people who feel super uncertain, in fact, be hopeful because it’s easier to reduce your uncertainty.

Ben Todd: Yeah, totally. And we have this article on the site called “How to make a career decision” and then that tries to lead you through a process where you first try and assess things a bit from the armchair. Then you identify your key uncertainties and then it asks, how can you actually go and figure out these uncertainties in the world? And we have this thing called “The Ladder of Tests”, where you can start with really easy ways to get more information and then do more and more costly things like an internship would be a relatively big commitment, but could still be really worth if it lets you test out a whole career path.

Are we highlighting the wrong problems and career paths? [00:51:19]

Arden Koehler: Okay. So before we wrap up here, any other ways we might be wrong that listeners should be aware of?

Ben Todd: Yeah. I mean, maybe in a way the most obvious one is just the list of problems and career paths that we highlight the most could just be the wrong ones.

Arden Koehler: Are there any that you’re particularly worried about?

Ben Todd: Well, I mean, just thinking about it, big picture, one thing would just be maybe we’re wrong about longtermism, just because that is a very new philosophy and there’s still a lot to be figured out about it. Then you can try to ask, “Well, if we were wrong about longtermism, what would be recommended instead and how different would it be?” Which I think is a really interesting question. So there’s alternatives to longtermism, one alternative to longtermism is currently called near-termism. Which you can think of as, well, what are just some common sense ways we can help people in a really high return way today. And people there tend to focus on helping the world’s poorest people, especially through global health or if they think animal suffering is a pressing priority, then they might work on reducing factory farming.

Ben Todd: But then I’m not sure if I would end up being a near-termist if I rejected longtermism. Another framework might be the conventional economics framework, which is maybe not a very catchy title, but I think it’s pretty different from near-termism because economists will think of benefits and costs over hundreds of years, but they’re just discounting them quite highly. So then they end up being like medium-termists, and I could imagine becoming a medium-termist and then I don’t know what I would work on then, but I could imagine easily ending up working on something like climate change or I mean maybe even still like pandemic reduction and biorisks. I could imagine that still turning out to be a top priority from medium-term perspective.

Arden Koehler: Yeah. So maybe some of our recommendations are a bit more robust to this choice.

Ben Todd: Yeah. And I should have also mentioned AI alignment. I could easily imagine that also seeming like one of the top things from a medium-termist perspective, either because there could be a big risk from it or because maybe you just want to… Maybe the medium-termists end up trying to speed up progress and do economic growth. That’s like Tyler Cowen, I kind of think of as a medium-termist sometimes.

Arden Koehler: Yeah. Interesting. Okay. So there’s the period of time termist choice point.

Ben Todd: Yeah, and those are not great terminologies. Because I think most near-termists, what actually makes near-termism distinctive is not really about when you think it’s good to help people. Most near-termists agree that future generations matter, what’s actually making the difference is something more about what methodology they think is best. And so they might want to put more weight on things that seem common sense good as opposed more speculative priorities.

Arden Koehler: Or they want to do things where they can get feedback about how well it worked. And then one thing about long-termism is, “Well, if I’m really trying to make a difference in the next 1,000 years from now if that’s my main priority, it’s pretty hard to get feedback” and they might think you’re just likely to go so wrong doing that, better to focus on the here and now.

Ben Todd: Yes. So I hope there will be a different name for these different philosophies at some point.

Arden Koehler: Yeah. So I guess we could probably talk forever about, okay, let’s go through all of the problems and paths and decide how uncertain we are that they are in fact as good as our best guesses at the moment. So we probably shouldn’t do that now, but is there anything else you want to leave listeners with about where your uncertainties lie about these recommendations?

Ben Todd: Well, we’ve already covered a bunch, which is like supposing longtermism is correct, then we’ve got these different varieties of longtermism and how much emphasis to put on each one. That would be another area.

Arden Koehler: So if patient longtermism is correct, we probably would recommend somewhat different things.

Ben Todd: Yes. We shouldn’t think of it as correct, but if it’s the main thing we should focus on then–

Arden Koehler: What do you mean we shouldn’t think of it as correct? If its…

Ben Todd: Well, the things are really a matter of degree. Even the people who are most into patient longtermism still think we should spend some on object-level things today. It’s just maybe they would only give half a percent of the portfolio as opposed to 4%.

Arden Koehler: I feel like the most natural way for me to think about it is if patient longtermism is correct, then it says you should spend a tiny bit per year now, I guess, to keep things running or something until the–

Ben Todd: And to get information because of diminishing returns.

Arden Koehler: Wait, sorry, what diminishing returns?

Ben Todd: If there’s some diminishing returns each year, then you’ll want to like give a little bit each year.

Arden Koehler: Right. But why would patient longtermists think there were diminishing returns each year?

Ben Todd: Well, so the patient longtermist, they might still think there’s like a small number of really good things around now, and so they they would still want to take those, but then…

Arden Koehler: They want to save a decent amount. Okay.

Ben Todd: So I think just thinking of those spectrums is more accurate in most cases.

Arden Koehler: Okay.

Ben Todd: Yeah. I see what you’re saying.

Arden Koehler: Okay, cool.

Ben Todd: And I mean then we could also get into the whole discussion of just, if we’re uncertain about all these views, we should probably take some kind of portfolio approach as well. But I suppose you were saying suppose we just knew perfectly what the situation was.

Arden Koehler: Right, yeah. So I guess when it comes to just how highly should we be recommending people work on certain issues or take certain career paths, that’s something that we’re just continually working on and trying to figure out.

Ben Todd: Yeah, totally. And there’s a lot of other career paths that we haven’t looked into yet and there’s a lot more research we’d like to do. So the idea that we’ll discover more paths that we think are really promising or downweight some of our existing ones seems fairly likely.

Arden Koehler: Yeah. I know Ben Garfinkel was recently on the podcast and articulating some reasons to think maybe AI safety was a bit less pressing than some people had thought. And it’s something we’re thinking about. Although, it’s worth saying that he still thinks it’s a big deal.

Ben Todd: Yeah and we still think it’s one of the most interesting areas to focus on.

Arden Koehler: Cool. All right. Well thank you, Ben.

Ben Todd: Cool, thanks.

Rob’s outro [00:56:58]

That’s the first piece on a series of chats we’ll be releasing between Ben and Arden over the coming months.

If you’d like some further reading on these topics, there’s Arden’s recent publications:

Ideas for high impact careers beyond our priority paths
Global issues beyond 80,000 Hours’ current priorities

Ben recently published a related blog post called The emerging school of patient longtermism.

And if you’d like to go deeper on patient longtermism I suggest listening to Phil Trammell’s episode on that topics which is interview number 73.

We’ll link to them all in the shows notes.

Also, here’s a reminder that our annual user survey is now open for submissions.

You can find that survey at 80000hours.org/survey.

The 80,000 Hours Podcast is produced by Keiran Harris.

Audio mastering by Ben Cordell.

Full transcripts are available on our site and made by Zakee Ulhaq.

Thanks for joining, talk to you again soon.