We live in a culture that glamorizes grit and derides giving up. But in the world of climate startups, we can’t afford to spend time on projects that aren’t working.
So, how do you know when to quit?
Decision Strategist Annie Duke is a retired professional poker player and World Series of Poker bracelet holder. She is also the author of several books, including Quit: The Power of Knowing When to Walk Away.
On this episode of Reversing Climate Change, Annie joins Ross to explain how to apply decision-making skills from poker and her research in the context of a climate startup.
Annie walks us through her three strategies for deciding when to fold, challenging us to set explicit kill criteria and leverage the ‘monkeys and pedestals’ mental model to abandon projects that aren’t working.
Listen in for Annie’s insight on prioritizing moonshots that develop useful technology and learn how to persevere when projects are worthwhile—and walk away when they’re not.
More from the show
Ad-free episodes and other benefits come with a paid subscription.
Carbon Removal Newsroom, the news show that ran alongside this one, is over. Its episodes are still up, on the feed Climate Workers Anonymous now uses.
Carbon Removal Memes is still going.
Resources
Quit: The Power of Knowing When to Walk Away by Annie Duke
How to Decide: Simple Tools for Making Better Choices by Annie Duke
Thinking in Bets: Making Smarter Decisions When You Don’t Have All the Facts by Annie Duke
Superforecasting: The Art and Science of Prediction by Philip E. Tetlock and Dan Gardner
Grit: The Power of Passion and Perseverance by Angela Duckworth
Full Transcript
Ross Kenyon: You’re listening to the Reversing Climate Change podcast by the team at Nori, the carbon removal marketplace. This is a show about the innovators and entrepreneurs developing solutions to climate change. Hello and welcome to the Reversing Climate Change podcast with Nori. I’m Ross Kenyon, I’m one of the co-founders of Nori and the creative editor there. Today I have with me Annie Duke, retired professional poker player and World Series of Poker bracelet holder, the author of multiple books including Thinking in Bets: Making Smarter Decisions When You Don’t Have All the Facts, How to Decide: Simple Tools for Making Better Choices, and most recently, Quit: The Power of Knowing When to Walk Away. Welcome to the show, Annie.
Annie Duke: Thank you for having me, Ross.
Ross Kenyon: Yeah, you bet. Were you a little surprised to get an invitation from a show called Reversing Climate Change?
Annie Duke: Well, I was, although I will say, I sort of thought — I was like, well, to be fair, decision-making is about everything. So yeah, maybe that’s why.
Ross Kenyon: Yeah, I think people listening may see your work as somehow incongruent, but I can see a lot of continuity through your research and your playing career. How do you tell the story of your career?
Annie Duke: So I think that from the outside, my career looks disjointed, like lots of different windings, sort of abrupt changes in it, but the way that I tell the story of it, there’s actually a lot of continuity to it. I started off kind of my adult life studying cognitive science at the University of Pennsylvania. My advisor was Lila Gleitman, who is a really, really wonderful woman, intellectual giant, and then at the end of five years I was about to defend my dissertation and actually was going out for job talks to become a professor, which was kind of my plan, and I got pretty sick and ended up having to take time off.
And during that time off, I started playing poker. I had played kind of recreationally before then, but I started playing to actually make money at that point because I needed it. I wasn’t in school anymore, so I didn’t have a fellowship, and I really just loved the game, was lucky enough to be pretty good at it. Ended up doing that for 18 years. I actually never went back at that point, I didn’t go back to academics. About eight years into playing poker, I got asked to speak to a group about how poker might inform decisions about risk.
And so I put together a talk for that group and sort of remembered how much — like, one of the things that I really loved about being a graduate student was teaching, and I kind of really loved thinking about this super explicit conversation the cognitive science and poker could have, because they both go under the sort of decision-making under uncertainty, and what happens, you know, how does decision-making go awry sort of under the very challenging circumstances that most of our decisions are made. So I overlapped that with playing poker, teaching about poker, writing about poker for about 10 years.
Then in 2012 I retired from poker, started focusing much more on the speaking, and I was doing consulting by then. So I had a bunch of clients that I consulted with, helping them make better decisions, and had a book that I really wanted to write, which ended up being Thinking in Bets. I then followed that with How to Decide and now Quit, and then ended up back at the University of Pennsylvania. So I teach executive ed at Wharton, I work with some graduate students there, I work with Phil Tetlock and Barb Mellers on forecasting. And anybody who, by the way, is interested in anything having to do with climate should be reading Superforecasting, because any decision that you make is a forecast of the future.
And this is how we sort of think about what the impact of decisions we make now might have on outcomes that we might see in the future, of the way that things might turn out. So that’s kind of like the past, but it seems like, oh, you know, she was a graduate student, then she was a poker player, then she was a business consultant and she was an author, now she’s back to being an academic and also doing consulting and also being an author. So she sort of mashed everything up into one. But that’s actually not true, because if you look at everything that I’ve done in my adult life, it’s really thinking about this problem of decision-making under uncertainty.
That’s true when I was thinking about cognitive science and how do we build models of the world. It’s certainly true when I was thinking about this very high-stakes version of that problem with poker, and it’s what I do in my business, and it’s basically the only thing I write about. I just happened to attack it through different lenses.
Ross Kenyon: Yeah, I asked that question kind of knowing the answer, because it all does tie together. When you read your work — people oftentimes don’t associate poker with this high-level thought, thinking of chess as being the superior thinking person’s game. And I’m pretty sure I got this from you, the difference between these games, but it seems like poker is a much closer map to reality, and life is not actually a lot like chess. Do you agree with that? And if so, maybe you can unpack that for us.
Annie Duke: Yeah, so I mean, look, chess is an incredibly complicated game. You have to be able to look at the board, you have to figure out what the possible moves are that you have, then for any of those moves you have to think about what the moves are that your opponent might make. Hopefully, you know, you know something about your opponent, so you have some sense of the probability that they might make each of those moves, and so on and so forth. You have to go pretty deep in the game there, so it is very, very complicated, but it’s not complex, at least not in kind of the academic way that we think about complexity, which requires uncertainty, right? It requires a very heavy influence of uncertainty.
So in poker there’s really two main forms of uncertainty that are exerting their influence on the outcome of any given hand that you play. The first is just hidden information, right? Like, I can’t see my opponent’s cards, so I’m trying to guess at what the strength of my hand is in comparison to my opponent’s when I cannot see my opponent’s cards. Now notice that’s very different than chess. When I’m looking at my position, you know — and obviously it takes skill to be able to properly come up with the answer — I can look at my opponent’s whole position and understand something just sort of on an objective basis about how strong my position is compared to theirs. I’m not having to make guesses at that, so I can see where all the pieces are.
And then the second thing that’s very different between chess and poker is that in poker there’s a much, much stronger influence of luck. So, you know, in chess, if I move a piece, it moves by an act of skill, by my own skill, and it stays there until either it gets taken off the board by my opponent or I move it somewhere else. It’s not like you roll dice and if they land seven, you get an extra bishop, and snake eyes and I don’t know, right? Basically, right? So backgammon actually sits in between poker and chess. In poker, obviously, there’s a very strong influence of luck because there’s the luck of the draw, so I don’t have any control over the cards that are going to come.
So backgammon is a game that sits in between — like, literally right in between chess and poker — in the sense that I can see my opponent’s whole position, I can see where all of the checkers are, right? But there are dice in the game, so what moves might be available to me are going to be probabilistic. They’re going to be controlled by the roll of the dice. So basically, if you think about most decisions that we make in our lives, they’re much more poker-like than chess-like, in the sense that for most things that we decide, we know very little in comparison to all there is to be known, and there’s a strong influence of luck.
So, you know, I can go through a green light and get in an accident having done literally nothing wrong, because I just don’t necessarily — I don’t know what the other cars are doing. I might not be able to see them, you know, someone’s about to run a red light. There’s luck, like my tire might blow at that moment. And that’s really much more like what our everyday decision-making looks like, and it creates this really hard problem that’s somewhat intractable, which kind of goes like this. So, Ross, if I told you that two people play chess and I didn’t allow you to see any of the moves that they made — so you haven’t watched any of the game, all you know is the outcome, which is that Morgan beat Taylor — who made better decisions during the game?
Morgan, right? And that’s because, again, when you take away the uncertainty, the outcome of chess is so driven by skill that you could bet tons and tons of money pretty safely that whoever won that game had made the better decisions. Now, obviously something could happen, like the person could have gotten ill in the middle of the game, and that’s why they had, you know — but it’s a very, very safe bet. Let’s say a chess game takes an hour, and I now tell you that Morgan and Taylor played poker for an hour and Morgan ended up with more chips than Taylor. Who’s the better player?
Ross Kenyon: Well, how long? How many hands do they play? How long did this go on? An hour? I mean, oh God, I don’t know. If they’re playing live, maybe 30 hands. If they’re playing online, they might get like 150. [unclear] There could have been some bad beats in there, and it’s —
Annie Duke: Right, now it’s up to you. I mean, this is like — this is the whole shebang here, right? Which is, who knows, right? So in chess, you can look at the outcome and you can work backwards and say something pretty significant about what the quality of the decisions were that were made. But in poker, at least in the short run, you have to wait a lot longer, right? Like, in the short run you don’t really know very much from the outcome. You have to wait longer. So, like, if you get in a car accident, one car accident, I actually don’t know anything about the quality of your driving.
But if you’re getting in lots of car accidents and lots of near misses and getting tickets all the time, over time I can start to say, well, I mean, Ross is having really bad outcomes, it’s obviously not random, he must just be a bad driver. But you have to accumulate much more evidence than you do in chess. And this is where the problem really is for us as decision-makers, is that we draw this very tight correlation between the quality of the outcome and the quality of the decision, and we’re not good waiters. We’re not patient. We don’t wait around to get 10,000 coin flips. We like to say things after you flip the coin one time.
Ross Kenyon: Is that what gut is? Is that even a thing?
Annie Duke: So the first question is, is gut a thing? Gut is a thing. Gut is decisions that are made implicitly. So the things that drive a gut decision are going to be the same things that drive a more deliberative decision, you’re just not going to have any way to examine them. You haven’t made them — you can’t interrogate them in the same way. So look, if I make a decision by gut, I’m still thinking about, like, what are my options? What are the ways that this could turn out? What do I think the most likely way that it could turn out is? I’m just not doing any of that out loud, right? I’m just doing it reflexively.
Whereas if I make a deliberative decision, I’m going to think about those things in a very explicit way, which is what you need to do when you’re making a decision, right? Like, what are my options? For any of those options, how do I think they might turn out? What’s the probability that they’ll turn out well? What do each of them cost me? And then I can start to think about, you know, then you’re just thinking about payoffs and sort of what’s going to get you to where you want to go faster. And you can either do that, like, super fast without writing anything down or discussing it with anyone else or even thinking about what it is that’s driving your own decision in any kind of explicit way — that’s gut decision-making.
The thing that I want to say about that decision-making: sometimes it does get you to the right place, because meanwhile, a broken clock is right twice a day, right? Like, it can get you to the right place. The problem is that it’s very noisy. You don’t really know, it’s not very reliable, and it’s really hard to uncover what your mistake is, or if you made a mistake, because you haven’t made anything explicit that you could go back and look at and say, like, was I missing something? And our memories for these things are really, really bad. So you want to be thinking more explicitly about the decisions that you make, because for a thing like this, this really matters.
I mean, you could think about it for, like, let’s say that we’re talking about something like reversing climate change, right? There’s all sorts of different options. Well, first of all, you have to figure out what’s true, so that’s hard. Then, what are the different options for sort of stopping what’s happening, stopping the progress that seems to be sort of marching along and actually speeding up, right? And you could be thinking about things like carbon capture, you can be thinking about, you know, speeding up the adoption of electric vehicles or renewables or nuclear energy.
And as you’re thinking about those different options, in combination or alone, you know, what you’re thinking about is: what are the possible outcomes? What are the good things that come out of this? What are the bad things that come out of this? What’s the likelihood that people will actually adopt it? How big of an impact will it have on the thing that I’m trying to get it to have an impact on? And then also, what’s the cost? Because that, you know, that matters in a world that’s practical, which is the world that we live in. And so we want to be very explicit about these kinds of things so that we can actually be more rational in the way that we make decisions, and so that we can close the feedback loops better.
Ross Kenyon: Many people listening work in the nascent carbon removal industry, trying to figure out ways to pull carbon dioxide out of the atmosphere or ocean and store it in some sort of permanent or durable manner. But many of these are early-stage companies. Maybe they’re raising money for the first time, and even more, Series A, maybe Series B, most, or even before that, in the seed stage, pre-seed. What types of questions should they be asking themselves about making good decisions, and also knowing when to quit? Because I feel like we over-glamorize stick-to-itiveness and Teddy Roosevelt and this sort of, like, damn-the-torpedoes attitude. I find it really liberating to hear that there’s this kind of a story of people who just gave up and it actually worked out better. How do you know if that is you?
Annie Duke: I mean, it’s such a great question. So let me just say, like, if you’re smart enough and you care enough and you really want to create a better world so much so that you decided to start a company — which we know is, like, really hard, you don’t ever see your friends or family for a really long time, it’s a really high rate of failure — like, you have to be, like, a super greedy person whose time and intellect is, like, really valuable for solving the world’s biggest problems. So the key is that what you want to do is make sure that if you’re that type of person, that you’re actually spending your time on the things that are the most worthwhile, right?
So what you don’t want to do is get stuck in something just because, like, people say never give up, or, you know, winners never quit and quitters never win. You don’t want to get stuck on a path that isn’t actually going to bear the fruit that you want it to bear. You want to be able to switch, you want to get to something better. Because as Ron Conway told me, life’s too short to spend your time on something that isn’t going to actually change the world in the way that you want to change it. But you’re right, like, we really glamorize grittiness. We think of it as synonymous with character in a lot of ways, whereas, you know, quitting things is sort of a character flaw, right? It means that you’re weak-willed or capricious or you don’t have what it takes. And all of that is really bad if you’re working on a technology that ultimately is going to fail.
Well, because in the end, the thing about grit is that it does get you to stick to hard things, which startups are hard. It gets you to stick to hard things that are worthwhile. The problem is, it also gets you to stick to hard things that are not worthwhile. The key is that you have to be able to tell the difference between the two, and here’s where the rub is: when you’re thinking about starting something like developing a new technology, say in the carbon capture space, right? There’s a whole bunch of stuff you don’t know, and luck is going to have some sort of influence on that. So you’re making that decision to start under uncertainty. The issue is that the decision to stop is also made under uncertainty, right?
It’s a forecast of the future, that the chances that I succeed at this are not high enough for what my goals are compared to other things that I might be doing. Okay, so you have to be, like, sort of looking into a crystal ball and sort of seeing — guessing, well, making very good educated guesses about how things might turn out in the future in order to figure this out. So there’s kind of three things, I think, that you can do that can really help you with this problem, because it’s actually a really hard problem. Because, you know, you kind of alluded to this, right? Like, grit is a virtue, quit is a vice.
So most of the cognitive biases that we have, things like the sunk cost fallacy or the endowment effect. So sunk cost is, like, once you spend resources on something, you’ll escalate commitment even when you start to see bad signals, because you don’t want to feel like you wasted all your time, or the resources, or your investors’ money, or employees’ time, all the people who believed in you, all that. But of course, waste is not a backward-looking problem, it’s a forward-looking problem. Is the next minute worth it, right? Endowment is ownership over things and ideas, and if you started a company, you’re very endowed to the ideas and the thing that you’re pursuing. That’s hard to give up. As a founder, your identity is wrapped up in the company, and the hardest thing to quit is who you are, right?
I mean, there’s all these things that kind of make it really hard. So what I try to get people to do is to think about these things in advance, for a couple of reasons. One is, spend the least amount of time that you can on something before you get the signals that it’s not worth pursuing. And that means you have to approach problems in a very particular way, which we can talk about. The second is, make a plan. What are the signals that you might see in the future that would tell you it’s time to exit the road? That might seem trivial, because you’re probably thinking, but if I see those signals, I will obviously exit when I see them. And all the science tells us that that’s not true.
That’s why there’s a lot of dead people on the top of Everest, because they didn’t turn around even when the time was right. And we all are climbing Everest every day. There’s all sorts of stuff we should be turning around and walking away from, we just don’t. Right? So you have to kind of do some advance planning to think about what the signals are, and then set some pre-commitments around that. And then the other thing is to get coaches, people from the outside looking in, who can really help you to see the situation you’re in with a clear eye. So we can discuss any of those three if you want, just sort of laying out what’s available.
Ross Kenyon: Yeah, I was actually just doing some quote-unquote research before this. I was watching the pilot of Poker After Dark and you were in this hand. You got a couple of great folds in that pilot. I don’t know if you remember filming it, but you had pocket eights and the board came seven, nine, jack on the flop. Gus Hansen had ace-king of clubs, so you were out ahead, and you had the open-ended straight draw, and then he bet and you called it. And then he bet because he hit a king on the turn, not a club, and you just folded. And you get a couple of these — you were betting when you’re ahead, you’re calling when you’re ahead, and you knew with a gut sensation that Gus was up to something. He’s famously kind of a crazy, wild player, but you believed him. How do you do that? I don’t know, it’s sort of magical to me.
Annie Duke: Oh yeah, so actually, let’s talk about this in terms of sort of item number two, which is think in advance about, well, what might happen that would make you fold. So if you take the specific example that you’re talking about, when a king hits, there’s just too many possibilities, even if he was goofing around, for him to have me beat in that particular situation. And there’s not a lot of good cards for me to hit anymore. So if I hit a ten now, even if he just has a queen, he’s going to beat me, he’ll have a straight. All those overcards are really bad for me, and if I hit an eight, then a queen still makes a straight now.
So I have to really believe that mine is the best hand, and I have to be really worried about any more cards that are going to come. And in that particular situation, given the uncertainty, it’s just a bad board. So I can tell you what I would be doing in that situation: I’m calling, saying if certain cards come, then I’m going to do something like raise him or try to take the pot, and if another set of cards come, I’m going to easily fold. So this is actually a very good example of what I really recommend people do, which is set kill criteria.
So the best way I can actually kind of describe kill criteria is from an example, from what Ron Conway, who’s the founder of SV Angel, who’s probably the most successful angel investor, at least one of them, in the space. And so he’s working with the earliest-stage companies. These companies are pre-seed at the time that he’s investing with them, so that’s obviously when there’s the highest uncertainty. Now, why is that worth it? Because when they hit, they hit really big. So the payoffs are there despite the long odds. You know, he’s following the companies, and there’s some point at which he can sort of see as an outsider that the odds are too long now for the payoffs that might be available.
In other words, he can kind of see they’re probably not going to get where they want to go, which would be a venture-scale business. And so he employs this use of kill criteria in order to help the founders see that they’re at the exit. And this also brings up the third thing that I said, which is get somebody from the outside who can see the situation more clearly than you can. So Ron Conway is acting as that outside voice, or someone who I might call a quitting coach, helping the founder who’s really in it, you know, who wants to succeed. They don’t want to give up, they think that means that they’re a failure, nobody will ever invest in them again, and they, you know, owe it to their employees, and their parents will be ashamed of them, and all the stupid stories that we tell ourselves in our head, none of which are true.
And he can see it more clearly than they can because he’s not defending territory. So he’ll sit down with them and say, you know, so do you think things are going well? And they’ll, you know, they’ll say no, but they always say, I know I can turn it around, right? But I know I can turn it around. And the interesting thing is, he doesn’t really argue with them. Instead, what he says is, well, how long do you think you need to turn it around? Like, how long would it be before we could see that you’re turning it around? So let’s say that that was two months from now, right?
So he’ll say, okay, so you think that if you really sort of, you know, you’re turning it around, you’re about to take off, and within two months we should know what that looks like. So tell me, what are the things that we’re going to see from this product or this company that are going to tell us that you’ve turned it around? So he’s getting them to sort of think ahead to two months from now, and essentially just setting benchmarks, right? If you can hit this, yes; if you fall short, no. You know, across the set of benchmarks that just tell you, like, it’s a more mature company. It might be like net new revenue, a product that we’ve been trying to roll out that’s been delayed actually gets rolled out, or whatever.
So anyway, he sits down and they agree, this is what turning it around looks like, and we’re going to revisit this, and we’ll revisit it in two months. So having done that, he can then go back and sit down, because they’ve created the list together, in two months and say, you know, either, yay, you exceeded all of this, which is fine, but, you know, normally it’s, so you fell short by a lot, it’s really time to walk away now. Your life is too short for this, you’re too smart to be spending your time on this. So that’s a good use of kill criteria, which is, what are the things that I might see in the future that would tell me that I ought to quit?
And one of the best examples of kill criteria that I’ve seen is when actually people are climbing Everest, they have something called a turnaround time for any day’s climb. So on summit day, you leave Camp 4 at about midnight, and the turnaround time is that wherever you are on the mountain at 1 p.m., you have to turn around. So if you have not made it to the summit by 1 p.m., you must turn around, and there’s a reason for that, that has to do with — there’s a really hard part of the climb called the Southeast Ridge. It’s very narrow, and it’s very dangerous, and if you fall you either go, like, 8,000 feet into Nepal or 12,000 feet into Tibet, and you need to do that in daylight.
And so if you get to the summit plateau past 1 p.m., it’s just too likely you’re going to end up on the Southeast Ridge in darkness on the way down. So that’s a very simple example of a turnaround time, of a kill criteria rather, which is like, if it’s past 1 p.m., I have to stop, I have to quit. And that’s what Ron Conway is doing, so he’s really a good example of sort of strategy two and strategy three, which is thinking in advance, think about what the signals are that you should walk away, go through that with somebody else, and then ask them to help you make that decision at the point that you get the information that you need.
Ross Kenyon: Or to do this at work — I lead creative at Nori. We do a lot of creative communications work and you try a lot of things that are often described — our team is the special teams department of the company. So a lot of these are weirder projects that, you know, if they hit, they hit quite well, and sometimes you cringe and say, this never needs to see the light of day, and that’s totally okay. I try to go in pretty nimbly too and say, like, we’re looking for reasons to say no, we’re looking for reasons that this is going to fail. Because if it is going to succeed, it will take time away from some other project; it needs to justify itself. So I try to set that up ahead of time, that expectation where it’s okay for things to fall apart and to not work. Is that, like, how you would go into a business scenario like that?
Annie Duke: Yes. So that’s very close to what I would do. I mean, I would say the first thing is, it’s not enough just to think about how it might fail, but to be explicit about what the signals are that you are on a failing route, write those down, and then commit to actions that are associated with that. I know it doesn’t sound like much of a difference, except it’s very different, because we have the intuition, well, if I’m starting this and I kind of have an idea of what it might look like if it goes badly, then when it’s going badly, I will pay attention. And the answer is no, that you actually don’t pay attention.
So there’s a lot of value in just taking a very small extra step, which is, write down, like, what are the early signals that might occur that would cause me to know that I’m on a failing route, and then let me say what I’m going to do about those in advance. So sometimes it would be you quit, sometimes it’s get more information, but figure out what that is. But to your point about looking for reasons to say no and trying to get there really fast, that’s completely on the nose. Like, that’s exactly how you get good at this kind of problem.
And the reason why you want to do that is twofold. One, it’s a much more efficient way to figure out what you should work on and what you shouldn’t, because there are, as you just pointed out, huge opportunity costs to not having your best people on the best projects, on the things that are the most worthwhile. But it’s also, you want to get to no fast in order to reduce those sunk costs, right? The less that you spend on something, the less you think you’re wasting when you quit, which makes it easier to walk away, because you accumulate fewer of these costs that cause friction, that make it hard for us to walk away, and that actually get us to spend more money on something that isn’t worthwhile, which then makes it harder to walk away, which then makes us spend more money.
So one of my favorite ways to approach projects, kind of à la what you’re talking about, is using a mental model called monkeys and pedestals, which actually comes from Astro Teller, who’s the CEO of X, otherwise known as the Captain of Moonshots, over at X at Google, which is Google’s innovation hub. So exactly what you’re talking about: by definition, these are moonshots, right? I mean, these are things that are, like, way out there, strange stuff other people aren’t thinking about, and really trying to be world-changing. So their thing is, like, you know, 10x better, in five to ten years to commercialization. So that’s really what they’re trying to do, and they want to spend their time on the stuff that’s going to create the most change.
So he has people approach problems using this idea of monkeys and pedestals, and it’s a very simple mental model that basically goes like this. Imagine you decided you want to train a monkey to juggle flaming torches while standing on a pedestal in the town square, right? So, like, if you pass by me, you put a dollar in my hat, right? It’s pretty cool. So this is — it’s such a simple mental model, it’s so elegant. What should you do first? Should you build the pedestal, or should you figure out if you can train the monkey to juggle the flaming torches? You should try to figure out if you can train the monkey, because otherwise there’s no point in building the pedestal, right?
So not only is there no point in building the pedestal, but you already know you can build a pedestal, because we’ve built pedestals forever, right? You can buy one on Amazon, you can turn a milk crate upside down. So if you spend any of the time up front on building that pedestal, you’re not actually getting any new information. You’re not finding out whether you can accomplish the big thing that you’re trying to accomplish. So you shouldn’t do that, because it’s false progress. It’s the illusion of progress, because you already know you can do it, and you’re spending time and effort on it, which is starting to accumulate sunk costs. Instead, your plan of attack should be, let me address the bottleneck here, and the bottleneck is, can I get a monkey to juggle flaming torches? The bottleneck isn’t the pedestal.
Now notice that that plan of attack is really different. How many times have you been in a team meeting and people say, well, let’s go, let’s tackle the low-hanging fruit? It’s pretty common — very common. But the low-hanging fruit is pedestals. You already know that you can — that’s like, we just want to feel good because we want to feel like we’re making some progress, but it’s low-hanging fruit by definition. There is no progress being made, but you already know you can do it. And meanwhile, you’re getting your team invested in something, you’re giving them ownership over the project, you’re spending resources, including their time. That’s all going to make it harder to quit when you butt up against a monkey that’s actually pretty insoluble, right?
So what Astro says is, it’s always monkey first. And what that’s going to do is exactly what you said, you’re sort of attacking it saying, I need to be convinced that we can do this, which is what you want to do, right? It’s, don’t convince me — I’m standing from a place where I think this can’t be done, and you have to convince me otherwise. So that’s what you just said, how you approach projects, which is exactly what this is. It has to be monkey first, because if you show me that we can solve for the monkey, then I will know that we can unlock the whole system and it’s worthwhile to build all those pedestals, right?
So it’s not that you shouldn’t ever tackle low-hanging fruit — obviously it’s a necessity, I have to have the pedestal at some point — but I should know that I can unlock that big thing, that really hard thing. So you have to do the hard things first, so that does get you to no faster. And as Astro Teller says, if I can get to no having spent two million dollars on a project instead of nine, I don’t think about that as a waste of two million dollars, I think about that as a saving of seven million dollars, right? Which is the exact right way to think about it. And then the other thing he points out that I think is really important is that it just reduces the sunk costs. If I get to no at two million instead of nine, I’ve just reduced the friction for walking away.
Ross Kenyon: All very persuasive reasons. And the opportunity cost one is also — that I think people forget about, it’s invisible, there’s no counterfactual for it. I had a big project recently where we tried to put together a carbon removal kids’ book, and I really wanted to do this. No one else has done it, it’s a fun idea, it’s a novel way of getting this message out, they would be cute, people will like it, and I suspect more than a few would buy it. But we wanted to figure out how to get to no fast, and one way we did it was writing enough of a prospectus to send out and then also just, like, soliciting agents and seeing, are there any takers here? Or would any publishers even want us to proceed with this? And we got a bunch of no responses or rejections for it, despite that the writing was quite good.
But ultimately it doesn’t matter if no one’s going to be there to buy it, because they [unclear]. And so I felt a little bit foolish in front of my peers, where I took a risk and I spent some money to pay wages and to make this thing happen, to get to this stage. I hope maybe I didn’t flex this enough, but I feel like the leadership of that was also knowing when to pull the plug and say, like, all right, we’re just going to call this one and just move on. But I don’t really know.
Annie Duke: Which is very hard, because, you know, in that situation — I mean, as you know, I think one of Astro Teller’s — another of Astro Teller’s really good insights is that when you butt up against a monkey that you can’t solve, you tend to start building pedestals instead of quitting. So a lot of people in that situation would have said, well, it’s just a proposal. If I write the book with the illustrations, then people will really be able to see what I have in mind.
Ross Kenyon: That’s where I stopped, by the way. I was like, right, next step would be, we self-publish and we do this, and —
Annie Duke: Right. And that’s where most people go, you know. It’s like the equivalent of — I always sort of say it’s like you’re trying to put a slide deck together and you can’t quite figure out, like, how you want to order it, or how you want to organize it, or what the words are that you want to put on the page. So you start designing the slides, right? So it’s just so hard to walk away from something once you’ve put effort into it, that there’s all sorts of ways that we’ll convince ourselves that we should keep going. You know, it’s worth more than other people think it is, I see things that they don’t, right? If they could just see the product that I have in my mind, then they would love it and they would buy it.
So then you start doing the things like the illustrations, which, while necessary — and you need a good illustrator — are a pedestal, because, you know, you could make illustrations for it. You already know that. The question is, is there an audience for it? That’s the monkey. Will people buy it? Will they read it? You know, if we’re trying to raise awareness about carbon removal, is this the best use of our time, right? Like, those are the hard questions that you’re trying to answer, and it’s hard to abandon course once you’ve put time into something like that.
Ross Kenyon: Definitely. One of the questions I have too is, do you learn enough about what you’re doing moving incrementally like that to even make a decision? Here’s a bad example: in intelligent design debates it was always like, you couldn’t just have a stepwise ear. The ear had to come fully formed, like, by God, because it’s just too complex to emerge, you know, bit by bit, evolutionarily. Are there problems like that where you really can’t see until it’s finally complete?
Annie Duke: So I think there are, and then you have to sort of decide whether it’s worth it to or not to build the whole thing. So I’ll actually give you an example from X. They were thinking about whether to build a hyperloop. So for those people who don’t know, hyperloop would be, you know, high-speed rail, like, à la what they have in Japan. They did monkeys and pedestals and they realized that, you know, the pedestals were building the track. I mean, obviously we’ve been building track for hundreds of years, so that wasn’t going to be very hard — at least since the mid-1800s, I guess, so, you know, a little less than 200 years. So they knew they could do that, and they knew they could build the train.
The question, and the thing that they didn’t know — the monkey — was, can we get this thing, if it’s going that fast, to stop safely and to start safely, right? So that was the problem, and you can see why that would be, right? I mean, a thing’s going so fast, it’s like, are you actually going to be able to brake in a way that’s safe to load passengers off and on? That was what they wanted to know. So they went through this exercise, and what they realized was that because of the nature of the problem, they would have to basically build the whole thing in order to be able to find out, because in order to get it up to speed you have to build so much track, and then to get it to stop, you have to build more, right? So they were like, it seems like we’re going to have to build the whole thing in order to be able to figure out if we can stop and start safely. And so they decided not to.
Ross Kenyon: I was so worried you were going to tell me that then they did it and then failed.
Annie Duke: No, they didn’t in this case. But other people are trying to do it, like, other people may think that that’s worthwhile, but it just wasn’t really in X’s DNA to do that. What they prefer — and this actually goes to the question of the ear — is they really prioritize projects where you may not really know until the end. But if you’re not going to know until the end, you better have technology that you can take out of it, right? Either — there should be significant things that are going to be usable and workable that you’re going to get out of it along the way. So an example of that type of technology was — I don’t know if you’ve ever heard of Project Loon, but it was an X project to put big balloons up into the atmosphere that then would connect to each other to bring internet to the world’s most remote places.
Ross Kenyon: Oh yeah, I think I’ve heard of this. That’s cool.
Annie Duke: Yeah, so it’s since been shut down. It actually turned out that they couldn’t make it happen, at least not at a reasonable cost. But the first approach they were trying to solve for was, like, how do you get the balloons to communicate with the ground and then with each other? And so they brought on a laser team to try to develop lasers that would solve for this problem, and it turned out it wasn’t a good solution for what they were trying to do. But then the laser team became its own team, because they developed really cool and new and useful lasers. So that’s kind of like the way that you can think about it, right?
Like, so some people — there are people trying to build hyperloops, right, who are totally fine with the fact that you get nothing out of it until you either know or you don’t. But X is not okay with that. And that’s why your example of the ear is such a good example. It’s like, every iteration of the ear had use. It had a purpose, it could work, it worked for whatever its purpose was, even if it wasn’t a fully formed ear. So, you know, had it stopped evolving at some point, it still would have served the purpose that it was serving at that time. Building an ear would have been worthwhile because even if you stopped at any moment, you would have usable technology essentially. And that’s the way that X tries to think about things. And also — can I just say that it’s really a dumb argument about the ear itself, or the eye, it sounds like.
Ross Kenyon: Then if you need to build the entire thing from scratch in order to determine if there is value, that’s almost like being all in preflop, right? You should probably have the top of your range.
Annie Duke: Or, yes, you would want to have a much higher probability that it will work. You’d have to have, you know — but you might have investors who are tolerant of that, right? So if you’re upfront, you say, look, it’s going to take us five years, we’re going to have to build the whole thing, the probability it actually works is like, you know, 3%, but man, if we do it, it’s really going to — that’s going to change the world. Are you in with me for this? You know, someone might be in with you for that. Okay, you know, that’s a question of what your values are.
Ross Kenyon: I’m sure some people will be okay with — maybe it’s a Gus Hansen holding and they’re going in on ten-two offsuit, and they’re cool.
Annie Duke: But that’s right. So, I mean, that’s the thing, like, it’s just a matter of, like, as long as you really understand what the payoffs are and what the problem is that you’re trying to approach, you know, instead of using your gut, right, to actually think about those things and say, is this the best use of my time? How am I going to — you know, is there something else that I could be doing where there’s going to be all sorts of usable technology that’s going to come out of it, even if we don’t ultimately reach the final goal, right?
And when you start approaching things that way — and I want to be clear, this is not an incrementalist approach. It’s not saying that, oh, we’re just going to incrementally improve on technology that we already have. No, this is for moonshots, because when you take moonshots, there is lots of technology that gets developed along the way, right? And you can just choose to prioritize moonshots where, you know, you’re going to be taking a lot of good stuff out of it, even if ultimately it doesn’t work. Or you cannot — I mean, again, like, that’s a value question.
Ross Kenyon: I want to end by, if you would, you just liberating people listening from the thing that they want to quit. But, like, I grew up and I wanted to play drums, and my parents didn’t want the drum sounds, so they told me go with your second choice, and I chose the saxophone. It didn’t take me very long to realize that I hated the taste of the reed in my mouth, and even today I can’t listen to, like, Coltrane or great saxophonists, because it sounds terrible to me, because I’m still traumatized by the reed. They wouldn’t let me quit for years. And in some ways I’m grateful, like, they forced me basically to get my Eagle Scout and stick it out with Boy Scouts long after I thought it was cool.
But I also feel like they could have been much more nimble by just letting me be like, okay, well, what if we sold the saxophone that we bought on bad advice, and how about a trombone? Trombone, [unclear], but also there’s no reed. But I probably would still be playing the trombone if I — yeah, if I had stuck with it. Like, was there a great lesson on — was it worth it? Or would you have given, like, a young me permission to quit?
Annie Duke: Oh, I would have totally, because the thing is, it’s like, what’s the goal? Is the goal for my child to be a saxophonist, or is the goal that I want my child to be doing some kind of enrichment exercise? Maybe I really want them to explore music, but maybe they’re a bad musician. So if they don’t enjoy it — if they’re bad and they enjoy it, I don’t care. But if they don’t enjoy it and they’re bad, maybe they try art, or maybe soccer, or, you know what? Like, you think about what the overarching goal is, and then there’s different ways that you can achieve that goal, right?
So I let my kids quit stuff all the time, I do, but they had to be doing something, right? So it’s sort of like, I was more like the one who didn’t want to decide where the tanks were going to go in the line, I was thinking more broadly, more strategically. Look, here’s the thing. I think that when we’re making decisions under uncertainty, it’s really easy to fall back on heuristics, right? Just sort of quick and easy rules of thumb, and in the battle between grit and quit, grit has won the day. And I think that part of that is because it reinforces a lot of biases that we have already against quitting, because we’re afraid of wasting things.
We have ambiguity aversion, meaning that we prefer something that we already kind of know isn’t going well to the unknown, to the vast unknown. We’d rather be on a failing course than move on to something we’re not sure whether — how it will turn out. You know, like I said, endowment and identity and all these things, and the nature of goals, which are pass/fail, right? Like, if I run 20 miles of a marathon, I failed. But I ran 20 miles — but I failed, right? And so there’s all these ways that we kind of process the world that make it hard for us to quit, and then this narrative around grit versus quit reinforces that, where grit is a virtue and quitting is a vice.
But if you read Grit, the book by Angela Duckworth, it’s very different than what kind of the popular narrative about that work is. What she’s saying is you should stick to things that are worthwhile, even if they’re hard. That’s the character, right? Is that you don’t give up just because it’s hard. So you figure out what is my passion, what are the things that I really think are worthwhile in my life, and if I have passion for those and they’re worthwhile, I ought to be able to persevere even if they’re hard. And what people have turned that into is just, when, you know, or even before her, winners never quit and quitters never win, quitters are losers, I give up, that’s what I say, you know, all of this kind of stuff that really just sort of sets it up that if you quit, you’re a failure and you’re a loser.
And that translates to even parents who don’t want their children to lack character by quitting. But when we do that, it’s such a disservice, because obviously there are contexts under which quitting would be the much better choice. And if people just take a moment to think about, like, people that I know, they know — who’s in a toxic relationship where you can see really clearly they would be better off not in it? Who’s in a job that they hate where they would be better off if they left, right? Like, who’s climbing a mountain and now there’s a snowstorm that came in, and why aren’t you turning around? I mean, it’s so easy to come up with very clear examples of people who ought to have quit and didn’t, and from the outside looking in you can see that so clearly. And yet we call quitters losers.
So instead, what we ought to do is say, look, we don’t want to think about grit as just a rule of thumb or heuristic — just stick to things and you’ll build character because perseverance wins the day. It’s stick to things that are worthwhile even if they’re hard, but then the opposite has to be true, which is quit things that aren’t worthwhile, because otherwise you have no way to free yourself up for the things that are worthwhile. That’s the problem. Every minute you spend doing something that’s going to go nowhere, that isn’t going to bring you happiness, that’s maybe going to traumatize you like the saxophone, is time you can’t spend on other things like playing the trombone. And that is so sad, when life is so short.
That’s the tragedy of it. We have such a limited time on this planet. We have to be spending time on the things that bring us joy, that we’re passionate about, that are actually going to get us to where we want to go. And in order to do that, you have to quit a lot.
Ross Kenyon: I can respect that, but you just sound like such a poltroon to me.
Annie Duke: I love that. Okay, so, well, let me just say that just proves that Ross read the book. So, poltroon — look, let me ask you this: if someone called you a quitter, would you be insulted?
Ross Kenyon: Yes.
Annie Duke: Right, it’s an insult. Now, if you called me a poltroon, I might be like, I sure am, but, you know, it would be because I was being funny. Calling someone a quitter is an insult, and it’s a sufficient insult that it used to be legal grounds for a duel. So the word that you mentioned, which is poltroon, is an old word that was a synonym for a quitter or a coward. It’s in that [unclear]. So when you look up synonyms for quitter, it’s like coward, and, although — it’s just the horrible words — but poltroon is one of them.
And Alexander Hamilton, I think, was called a poltroon and, you know, dueled somebody over it, and it was totally fine within U.S. law. Andrew Jackson, I think — oh, Andrew Jackson, sorry, not Alexander Hamilton. Excuse me, Andrew Jackson. So Andrew Jackson was called a poltroon and immediately challenged the person to a duel. Like, I think shot him on, like, Fifth Avenue, because — well, whatever. But yeah, anyway, that was legally okay because he was insulted, right? Calling someone a quitter was a sufficient insult that you could challenge somebody to a duel. That’s how bad it is.
But again, the thing is, think about how absurd this is. Like, what if the first person you ever dated you had to marry? I mean, think about that, it’s so stupid. When it comes to dating, it’s quitting kind of by definition. It’s like, I’m going to date a bunch of people, I’m going to quit most of them, and then I hope maybe I’ll find the person that I’m going to commit to. So, you know, we can see this value of quitting all over the place, and yet when you ask people, they’re like, don’t be a quitter.
Ross Kenyon: Yeah, I wish we could change that. That was a lot of my childhood too. I’m grateful — I feel like I do have a lot of grit now — but also, was that actually good for my character or my relationship with parents, when they were forcing you to do things that you clearly hate for years?
Annie Duke: Sounds like — the other thing is, it’s not clear that that developed grit. And that’s the other thing, making someone stick to something they hate, it doesn’t necessarily — that doesn’t necessarily develop grit. Making someone stick to something that’s hard, that helps develop grit, because they can see that it doesn’t stay hard forever, right? And again, this is what the difference is. It’s like, what’s the difference between hating something and it being hard? How do you tell that difference? And it goes back to that question that you asked me. Like, there’s lots and lots of founders who are working on technologies. How do they know when it’s just hard, and how do they know when it’s not worthwhile?
And that’s very much what the book is about. And that’s why I talk about kill criteria and monkeys and pedestals and getting coaches, and, you know, making sure you have outside voices to help you try to figure those things out. Because sticking to hard things is good, as long as those hard things are worthwhile and they’re going to get you to the goals that you actually have. But if they’re not, then it’s just blocking you from achieving your goals.
Ross Kenyon: Well, thanks for being on the show. I’ve read your work for so long. It’s a true pleasure to speak with you, and thanks for spending time with me.
Annie Duke: Well, thank you for having me on.
Ross Kenyon: Yeah, links to all those things, they’re in the show notes. Go get her nice new book, Quit, as well as her other works. Links are in the show notes. Give us a great rating and review on a podcast app or Spotify. Thanks for listening, tell a friend, and have a great day. Thank you so much for listening. If you could please subscribe and give us a great rating and review on Apple Podcasts, or a rating on Spotify, that’d be much appreciated. It helps us get our content out to more people. You can sign up for our newsletter at nori.com, follow us on social media. We will catch you next time.
