Philosophy of The 80,000 Hours Podcast
Rob Wiblin: So initially, as I described, the primary philosophy of the show was laziness and me trying to get as much content out there with as little effort as possible. But as we’ve done more and more, I have developed stronger views and stronger tastes, and a sense of “What’s a gap that we’re filling that needs to be filled?”
Rob Wiblin: The core aspect of that is, I perceive with interviews in general, folks think that the serious thing to do is to get someone and then have a slightly adversarial thing — where you can reveal the weaknesses in their argument, or you can reveal that they’re bad in some way that people didn’t realize. That’s viewed as “serious journalism” or “serious coverage” — whereas we do something that’s quite different, and I’ll explain why I think that it’s potentially underrated.
Rob Wiblin: We think of these interviews as collaborations with the guest. So they actually get to look at all the questions ahead of time if they want, and potentially cut out any that they don’t like. And after we’ve finished recording, we send them a transcript if they want to look at it, and they can cut out anything that they don’t like. So in that sense, it’s a very softball interview. But I don’t think that that is a problem at all, because we’re trying to find people who we think are really worth listening to, who have interesting opinions — a worldview that’s different from other people, that is instructive. Where the important work is getting them to fully explain themselves — getting them to communicate to us so that we can start to see the world through their eyes, and see the problem that they’re working on through their eyes.
Rob Wiblin: And that takes a lot of time, and it takes a lot of openness and followup questions, and people willing to be sincere, and to not feel threatened, and be willing to get into all kinds of details in a really honest, intellectual discussion. So even though we’re not holding anyone’s feet to the fire, I think you can potentially learn a lot more from these interviews — at least with the right guest — than an interview where you’re trying to pull out something from someone that they’re not actually excited to say.
Rob Wiblin: I guess we live in a time when people — and maybe this has always been the case — are very cagey, because they’re worried about saying the wrong thing. They’re worried about something blowing up on social media. Even folks who are working in not very controversial areas have learned, at least because of the way that we talk about that kind of risk, to reveal as little as possible — to think about exactly how you’re going to phrase things ahead of time, to check things with your media person.
Rob Wiblin: And potentially that makes sense. But by allowing people to just cut things out afterwards with the benefit of hindsight, it does mean that they don’t have to be cagey at the time. They can say their thing and then see afterwards — maybe it’s actually fine to put out; maybe it’s not going to be a problem. And they don’t have to be doing two things at once in their head — both trying to figure out how to communicate their actual views, and also thinking, “How is this going to be used against me at some point in the future?”
How The 80,000 Hours Podcast stands out
Keiran Harris: One very obvious thing is that we are a part of this effective altruism community; that makes us stand out. If we’re talking about podcasts, this is very unusual to be having a show that actually is, in a not-so-hidden way, actually really cares about concretely doing good in the world. And not even just doing good, but actually caring about the content of an episode, and being able to apply it to the real world.
Keiran Harris: Our focus is that we ultimately want people to be able to use a lot of this information to go out there and make the world a better place. And some of it is indirectly: sometimes it’s just keeping people engaged in these ideas, and staying part of the effective altruism community, or being introduced. But a lot of the time, we have very specific episodes, like, “Here is an episode all about how you can go and be a machine learning engineer.” That is just there for you. And we’ve had a lot of people who’ve changed their careers directly just from listening to the podcast, and that’s a big deal for us.
Aaron Bergman: Wow, that’s awesome. How do you guys think that the desire to really improve the world and change people’s careers affects how you produce the show?
Rob Wiblin: I guess the great majority of content that people read online, or podcasts that are produced — including most news — is basically there for entertainment. And that’s totally understandable — it’s the reason that I consume most content, because it’s like, “Oh, that’s interesting.” I enjoy learning stuff about the world: it feels semi-productive, and it’s just entertaining to feel like you’re keeping up with what’s going on.
Rob Wiblin: And though there’s absolutely nothing wrong with that, it does mean that very often, at the end of, say, a Planet Money episode about this or that topic, they’re not super inclined to close the deal and say, “And what does this imply about what someone ought to do about something specifically? What kind of job does this imply is more impactful?” Or “What kinds of research topics are most pressing within this particular area?” It’s basically a thing that’s completely absent, because people are not reading The New York Times in order to make actual decisions, almost ever. They’re reading it just because they get a good feeling knowing that the thing that’s important to know about — what’s going on in the world in general — even though it basically never changes anything that they do.
Rob Wiblin: I would say it’s partly just the business model: the business model of news is advertising or people subscribing. You’d probably shrink your audience by trying to have a focus on what a specific person should do. Because then, potentially only a tiny number of people are actually able to act on any specific piece of advice about some specific policy area, or some specific industry that’s doing something good or bad. So it’s a terrible business decision.
Rob Wiblin: We have this incredible luxury, because we’re able to be funded through philanthropy as a nonprofit to produce this content. We can then actually do a bit at the end of the show, where we think for half an hour about what does this mean? Specifically, what’s to be done? Even knowing that the audience for that might be six people, we can still pay the bills, despite the fact that some people might tune out at that point. But it means that that section is quite different, or the mentality sometimes, when we’re like, “Okay, and what specifically should 80,000 Hours recommend to its readers in order to allow them to have more impact?” It’s different, more pragmatic, more practical-minded than most documentaries or most podcasts.
Keiran Harris: I like the idea that whenever someone sits down to listen to an 80,000 Hours Podcast episode, it’s like you’re buying a ticket to a raffle: that maybe this new cause that you’ve never heard of might be the thing you want to work on one day. So it’s probably not, but if you’ve never thought about, let’s say, voting reform, but now you’re going to sit there and listen to this episode on it. Maybe by the end of 90 minutes, you’re going to go, “Wait a minute, that sounds exactly like the kind of thing I would want to work on.”
What we try to avoid
Rob Wiblin: Over time, we’ve developed a bunch of red flags — either learned through experience or by reasoning things out — that this is not potentially going to be a super promising episode. Anything that’s super topical is a little bit suspicious. We would ideally want most of the episodes to be interesting in one, two, three, maybe even five years. There’s a bit of a social phenomenon where people want to talk about whatever other people happen to be talking about in any given minute. So through a somewhat random process, something becomes topical or trending, and then there’s this incredible suction towards everyone in the media talking about some topic right now — even though there’s other more pressing things going on and there’s no particular reason why this thing has gotten highlighted.
Rob Wiblin: We — and I guess a lot of other podcasts, because of the timelines involved — are able to avoid that topicality most of the time. I guess an exception was COVID, where we did do a bunch of fast turnaround episodes. In our defense, that was maybe the biggest news story of our life, so we make an exception once in a while. Despite the fact that we’re getting deeply into people’s worldviews, it’s not primarily a show about people or about people’s reactions to things, how they felt about something when it happened.
Rob Wiblin: The main reason not to do that is… Okay, imagine that you’re doing science journalism for The New York Times. Most people don’t understand biology very deeply, so if you’re talking to a biologist and you want to reach a broad audience in order to have a large advertising base and make enough money to pay the bills, then the thing that most people can understand is how the scientist felt when they made the discovery. And so even in interviews that naturally might be extremely technical, they tend to become quite personal because of the need for many people to take an interest in something that’s actually going to turn a profit.
Rob Wiblin: We tend to lean away from that and have something more of an engineering mindset. Some people don’t like that. It’s definitely a piece of feedback that we get, that people might like to get to know the guests a bit more on a personal level, or hear more about how I feel about things, or what struggles I’ve had in my life. And we do do some of that — it’s definitely not a complete no — but I feel like a more important gap that we can fill is actually being more technical and remaining focused on the practical, pragmatic question of how to solve problems, and what people have accomplished at an engineering level.
Keiran Harris: Another cool thing that we try to do is to not duplicate. We do this by prepping very carefully — we do way more prep than the average show does. We’ll typically have sometimes five to 10 pages of questions that we don’t end up using. And we really will try and ask guests things that they haven’t talked about before. If they’re talking about a book, we’ll try and put a lecture at the start of the episode, or we will direct people to something they’ve already done.
Keiran Harris: It’s a big reason why we go so long — three, four, five, sometimes more hours in. Because if somebody has their schtick — where people just go on every single podcast and they do exactly the same 60 minutes — if they’re used to that 60-minute schtick, and we’re like, “We’re just going to talk to you for three hours,” then in that third hour, they’re going to have to talk about new stuff. They don’t really have a choice. An essential thing that I think about as a producer of the show is that we want to make sure that our episodes are delivering content that you can’t get elsewhere.
Keiran Harris: Another thing that you want to avoid — and this is particularly a problem if you’re interviewing friends — is jargon. We really try to avoid it as much as we can in episodes, not always successfully. If you have someone that you’re really close with and you just have a way of speaking off mic, you will just naturally just start to talk using these terms that no one knows what they mean outside of your small subgroup, and I hate that. So you’re trying to have this balance where it’s very fun, it’s very accessible, and you can tell the people like each other — but at the same time, you really want to be bringing everyone alongside that journey of the conversation.
Topics we'd like to cover more
Keiran Harris: For me personally — both because I think it’s important and also based on the feedback we’ve gotten — I would love to cover more mental health topics, which we’ve tried to do this year. I recorded an episode with my colleague, Howie, which we released earlier this year, and it was phenomenally successful. It’s our most popular episode ever, based on user feedback and listening time. Based on our user survey, I was pretty sure that that was going to be a hit.
Keiran Harris: And the response to that makes me think that there’s this need for this content, and we are in a position to be able to take that on. So I would personally like to be able to do more of that, even though it is very non-standard, and Rob was saying earlier that we normally don’t talk about personal experiences much. I think we can lean into that a little bit more. And we have started asking guests at the end of episodes, actually — we’ve had a few good examples where we’ve asked guests how a certain moment in their life was to deal with emotionally, and if they have mental health tips and things like that.
Keiran Harris: I think it’s just great to be able to normalize that, to be able to just very casually talk amongst 80,000 Hours staff — where we do have this position in the community, where I think generally we’re pretty well respected — and to just be like, “Everyone at 80,000 Hours has either dealt with this themselves or loves people who have, and are very sympathetic to this.” And to try and promote this norm of, “That is okay and we’re here to support you” — I think it’s very important.
Keiran Harris: One more thing to add is that sometimes we have these gaps and we fully intend to fill them and just for practical reasons they fall down. So one example is, we haven’t done an episode on s-risks, on risks of astronomical suffering. It would be very reasonable for people to look at our library and say, “Why haven’t you done this yet?” But we do want to do that — it’s just a matter of finding the right guests. So we’ve been in conversations with people trying to find someone who is happy to come on, and basically those people are saying, “Actually, we would like to do that in maybe a year when our research agenda is more advanced.” So it can sometimes look from the outside that we don’t care about a topic, when actually there are just practical things about recording podcasts that get in the way.
Rob Wiblin: I guess there’s still lots of time, so there’s some things that I regret that we haven’t done yet, but maybe we will in the future. So I tend to be, as an economist, someone who has had training to be relatively positive about markets and perhaps sometimes a bit skeptical or a bit cynical about government interventions. I wish we’d found more guests who could make arguments that I think are persuasive and are likely to persuade the audience that markets are inappropriate in this particular area, or people perhaps from heterodox economic schools who can present perhaps a more socialist angle on things. I think realistically, the reason is that because of the views that I have, I don’t find those arguments super persuasive, so it can be a little bit hard to find a guest who I think, “Yeah, this person’s going to persuade me a lot.”
Rob Wiblin: We recently had an example where someone who has a lot of political opinions that I wouldn’t naturally agree with — Mushtaq Khan — we did a long interview with him. He has a more interventionist perspective than I would typically have, but I found him very compelling and very interesting. So hopefully we can do more of those in the future.
Rob Wiblin: Another thing we’ve never gotten to is doing a debate, where I significantly disagree with the guest and then we kind of structure it such that we’re going to try to get to the bottom of some issue where we don’t see eye to eye. Or an alternative structure might be that I host two different guests who have different perspectives, and then I try to mediate between them and reach some sort of agreement. I guess I’m a little bit scared of that format — I think there’s a degree of cowardice on my part, perhaps that it can be socially awkward to figure out how to navigate a situation where, in front of a large audience, you’re disagreeing with the person you’re talking to. And how hard do you push? But it’s probably a skill that I should try to develop a little bit better.
The post-production process
Keiran Harris: So it varies hugely depending on the episode. Sometimes we have episodes that just need virtually no editing. So one example was our episode with Cal Newport recently — he’s such a polished speaker that I made almost no cuts. It was crazy, but the whole episode was basically just what the conversation was. We only had him for two hours, and that’s basically what you hear. And in some cases, like we say, we’re very open to doing multiple recordings. Sometimes we can record for 10, 12, 14 hours or something, and it’ll end up being a two-hour episode.
Keiran Harris: And a nice compliment is that I feel like people don’t know when it’s been a real nightmare to edit — we’ve never gotten any feedback like, “Obviously that was from multiple sessions, and this was jumping around all the time.” Yeah, it just really depends on the guests. And we’re very open to having people who don’t have a lot of experience. It just means that you have this option of getting really brilliant people who have these incredible ideas, but they’ve never done any podcasts before, because that’s just not their kind of thing. But that will mean that there is a lot to be edited out.
Keiran Harris: So we have these multiple stages. We have a terrific audio engineer, Ben Cordell, who does the first stage for me now. I used to have to do this stuff in the early days, but Ben does it for me now, where it’s just taking out all of the filler words and the false starts. Often, people will start a sentence and they go, “Wait a minute. Let me start that again.” Or it’ll just be random jumbled words. And so we take all of that out to just basically clean up the audio.
Keiran Harris: Then it’s sent to me, and I make the editorial cuts, which is a series of things. Sometimes, questions just don’t go anywhere — so we’ll ask a question, the guests will just say something like, “I don’t really have a good answer for that,” or they’ll give an answer that’s kind of a non-answer, where they probably should’ve passed, but they didn’t because you kind of feel obligated to say things. Sometimes, we have full sections that didn’t quite work, so I’ll do a lot of rearranging sometimes. We really try and make sure that the most engaging section comes at the start. So actually, sometimes the section that you hear at the start of an episode happened three hours in, and that was the best bit. And so as long as we haven’t created these contingencies, where people keep saying, “Well, as we were talking about earlier…” —
Rob Wiblin: Yeah, that’s another piece of advice that Keiran will have. I can hear him feeling unhappy on the other end of the line when we’re recording and we’ve created a dependency, where we’ve referred back to something in the past — as we’re doing right now — which is then going to make his editing job enormously harder, potentially, if we want to rearrange.
Kearney Capuano: So Keiran, you’re doing all of that behind-the-scenes work? That’s all you, or do you have a broader team that’s helping you with all this?
Keiran Harris: Yeah. I lead the team. Basically, as Rob said, we’re a four-person team. We have Rob as the host, I’m the producer and the editor, then we have Ben Cordell (who’s on audio engineering), and then we have Katy Moore — she basically does all the work with the transcripts. That used to be Sofia Davis-Fogel — people have probably heard her name a lot. So basically, I’m managing those two, and between the four of us, we’re taking care of all this.
Is there an optimism bias in the effective altruism world?
Keiran Harris: So there’s a couple things going on here. One is that I do think, typically, the effective altruism community is made up of very cheerful people, and we do focus on this longtermism stuff. But within longtermism, you can be very worried about suffering, as I am personally, and I know Rob is as well. A lot of us really do care about animal welfare and we have several episodes on that. It is just that this is a matter of the core of effective altruism, where you can’t focus on everything equally. And so the fact that we don’t focus as much on one cause doesn’t mean that we don’t care about it. I mean, both Rob and I are just absolutely horrified by the horrors of factory farming, viscerally. It is something that really bothers me in a way that few other things do.
Keiran Harris: But if I sit down and just actually think about these ideas — for me, it’s difficult to make the case that it is of comparable importance to AI, in my opinion. And then Rob has fairly similar views. And so the question is, in making The 80,000 Hours Podcast, should we follow what we think is best? Should we focus on these problems that we personally think are the most important? Or do we try to lean back and say, “Well, this is a more accurate representation of the entire community”? We try and do a little bit of both…
Keiran Harris: One thing is that just focusing on longtermism isn’t inherently positive. More so than I think maybe anyone else at 80,000 Hours, I really care about or am interested in mitigating s-risks — so these are these risks of astronomical suffering. It’s often framed as, “Well, there are these longtermists, and there’s other people who care about s-risks.” And it’s like, “Well, no. If you care about s-risks, you are a longtermist.” This is over very long time periods. And so you can be entirely motivated by suffering, and still be a dyed-in-the-wool longtermist. These things go together, potentially.
Rob Wiblin: Just backing up, let’s pause and think through this question of, “Is there an optimism bias in effective altruism?” In general, it doesn’t seem like there’s an optimism bias among human beings. If anything, it seems like people tend to be biased towards remembering bad things that happened to them, and being annoyed by that, and then worrying about things that could go wrong in future. This would be going against the stream, I think, if effective altruists had a different thing, and were generally focused on only thinking about ways that things could go well.
Rob Wiblin: And if you look at the stuff that the effective altruism community is worried about, it’s true that factory farming isn’t the number one intervention in terms of people’s efforts or donations. But we talk more about terrible suffering on factory farms, or terrible suffering in the wilderness, than basically almost any other group other than communities that are specifically focused on animal rights and animal welfare.
Rob Wiblin: Even within longtermism — where perhaps people are a bit more optimistic about how well the future will go — the main focus for the last 20 years has been ways that we could go extinct, and the ways that things could go horribly. And there’s many people who think there’s more than a 10% chance that this century, all of us are going to be destroyed and everything we care about will be ruined — which, actually, that’s a very common view among the general public. So it actually isn’t that different than the background rate, but it doesn’t seem to suggest that there’s a kind of Panglossian bias among people involved in ineffective altruism.
Rob Wiblin: On the poverty side, I think most people focused on that are probably optimistic that, in the fullness of time — over hundreds of years, or possibly even over the next few decades — we can largely eliminate extreme poverty, or we’ll get a long way towards massively improving the wellbeing of people in countries that are currently very poor. But if you look at the evidence, there’s pretty good reasons to think that. Yeah. I don’t see that much reason to think that the community is biased towards naive optimism.
Keiran Harris: I think that’s a really good point. I was just going to say, let’s say you take people who work on extreme pandemics or preventing nuclear war, and they’re thinking through these scenarios. These are very grim scenarios. This isn’t like, they’re waking up with a sunny disposition and a song in their heart, just thinking, “Oh, nuclear war. Oh, what an incredible thing to work on.” No, this involves all sorts of incredible suffering, and it’s just in these different avenues.