The Myth of Deep Beliefs: Leading in a World of Improvisation with Nick Chater

Speaker A: Be happy with your improvised self. So I think I've spent a long time thinking. I don't really understand what I'm trying to achieve. I don't quite understand how I'm trying to do it. I don't quite know my values. I've got some broad idea, but it's all a bit vague. This is not good enough. And I think, no, it's okay. You don't have to feel an imposter or you're, uh, a failure for not having worked out all the details. That is the magic of human intelligence, that we can cope fluidity and uncertainty.

Speaker B: Can you achieve extraordinary results while making the world a better place? I'm Bruce Kempkin, and together we'll explore how. Welcome to Humanity at Scale Redefining Leadership Podcast. Join me as I speak with visionary leaders and top experts to uncover the secrets of humanity centric leadership. Ready to rethink leadership? Let's dive in. Hello and welcome back to Humanity at Scale. Leadership is often described as setting direction in making decisions. But in practice, it's really about something harder, coordinating how people think, feel and act, especially under pressure. When clarity breaks down and conflict shows up. Most leadership models assume people have stable beliefs and preferences that can be aligned. But what if that assumption is wrong? What if much of human judgment is constructed in the moment, shaped by context and social dynamics? That question sits at the heart of today's conversation. My guest is Nick Chater, professor of Behavioral Science at Warwick Business School and one of the world's leading thinkers on human decision making and rationality. In his book, the Mind is Flat, Nick challenges the idea that our beliefs and preferences run deep, showing instead how much of our thinking is improvised as we go. In this conversation, we explore what that means for leadership. Coordination, trust, and designing organizations actually work with human behavior. Nick, welcome to the show.

Speaker A: Well, thanks so much, Bruce. It's a real pleasure to be here.

Speaker B: Now, before we go into your fascinating research on the human mind and behavior, I'd love to get a little bit more about your background. You spent your career at the INSYS section of Psychology, Philosophy and Behavioral Science. What drew you into studying how people think and decide in the first place?

Speaker A: It's an interesting question. So I started off wanting, like so many people, to be a mathematical physicist. As a child, that was the glamorous, exciting thing to do. And I like doing maths. And I started reading philosophy because that's another one of those sort of big, mega question kind of disciplines. I went off to university, so I was at Cambridge Trinity College, which is the sort of epicenter of maths where, uh, all the maths undergraduates go. And soon realized that I wasn't going to be a star mathematician and also that I was kind of more interested in these more abstract questions, switched into philosophy and then I took a side course in psychology, thinking, oh, this would be very useful for thinking about these big philosophical questions. And suddenly thought, oh, this is really cool, I'd like to do more of this. Ended up doing more psychology, thinking I'd go back to philosophy. And in fact kind of ended up none of these places, right. I'm not really. And I guess if anything I'm a psychologist disciplinary identity. But in reality I'm a kind of mishmash between bits of psychology, bits of philosophy, little bits of maths, and indeed these days, a bit of behavioural economics at the edges as well. That was the beginning. And I've started off in a fairly jumbled fashion and I've continued just as jumbled ever since.

Speaker B: So it's interesting, you were looking for big questions, right? Even as a little kid when on that path, when you sort of started to look into the research, did you realize that, oh, my goodness, people just

Speaker A: got this wrong, I think really late, actually. So an interesting and near approach to thinking like this. When I was an undergraduate in my second year at Cambridge, I went to see a lecture by Geoff Hinton, Geoff Hinton, the key figure in development of modern AI. And this was 1985, a, uh, long time ago. Uh, and he was talking about neural networks which were really tiny. Uh, they had 100 units that would be a big one. And they were doing kind of amazingly cool things, but on a tiny, tiny scale. And I can remember then thinking, well, this is all very well, but this can't really be the whole story about cognition because it's just too simple. And anyway, there's nothing that maps onto these everyday things we think of as like, knowledge and belief and desires. The everyday way we think about our minds, it didn't seem. Which is very linguistic. So we're very focused on the kind of things you can write down, kind of things you can put in mathematical notation and, uh, things you can do sort of reasoning over in a symbolic way. And I couldn't see how those were going to map onto the kind of thing that Geoff was talking about. So I thought, well, it's kind of very clever and very cute, but that can't really be how the mind works. The mind must work in a way that is somehow links in with our, uh, everyday conception of Ourselves as kind of reasoning beings, fundamentally linguistic in nature, so had a near approach to thinking quite differently. Because Geoff, I think, did think quite differently even then. I then ended up sort of gravitating towards thinking, no, we should be taking a symbol processing perspective on the mind. The mind's got these essentially, it's got a repository of knowledge which is written in a kind of language of thought of some kind. Just like a computer, uh, system has giant databases full of archived knowledge which is all encoded in a symbolic format. Well, mind must be something like that. Beliefs are, uh, you know, bits of that knowledge which you think those. I think those are true. And hopes are things where you think, well, that's not actually true, but I wish it was. And so this sort of perspective that the. And all of these seem to be quite stable. Of course, this is the way we tend to think. We think, well, I've got these beliefs, I've got these hopes, I've got these, uh, fears there. They're kind of, as it were, stably, uh, written in my mind. I can go in and inspect them. Maybe if I have a Freudian perspective, I might think, well, some of them might be tucked away and hidden. Or maybe there are several parts of my mind which, each of which have different ones. And that was a kind of mindset I carried around for decades, really. And there are people like Geoff Hinton saying, well, yeah, maybe I've got this totally wrong because actually, maybe that you build really rich intelligence which ultimately produces language and works in a symbolic way, but the machinery underneath is totally different. And so there aren't really inside you. There aren't really any of these beliefs and desires of any of these symbolic representations. But it took me really decades to buy that. And I guess it came about through just hitting the wall of trying to get this everyday perspective to work, try and get the everyday perspective to work. And you look at how we actually behave, how the experiments turn out, it just kind of just doesn't do the job. So I became frustrated on that cumulative way of thinking, God, it's just got to be right. But somehow it doesn't seem to fit. And then of course, alongside that, actually rather late, really, after I had changed or flipped my perspective, the modern large language models and the sort of whole revolution in AI, uh, came to fruition. And now instead of having to say to people, well, there must be some alternative that, uh, we have this. There must be some alternative machinery that miraculously generates all these linguistic improvisations that we create about our lives. We can say, well, There is one. It's not. It's like us. It felt different from us. But we can see it can be done because the technology now exists. I was a slow adopter. I really was. I didn't rush it.

Speaker B: So was there a research study that you did or that you were part of that highlighted, oh, my God, I got this, uh, they are wrong and I got it right and now I can prove it.

Speaker A: I think it was cumulative. But the kind of thing I think was some studies that we did, and these came actually originally from NOW idea by Eldar Shafir, who's a psychologist at, uh, Princeton. So he had these studies which we picked up and did quite a lot of work with. But Eldar, studies are the kind of classics, really. So what you do is you give people a bunch of choices. Say, it could be just two options, and you say, which of these two options or three or whatever would you prefer? And people say, well, I'll have a please. And then you give the same people, or actually a different group, but essentially identical people, you give them the same options. You say, which of these two would you disprefer? Which would you like to reject? And if you can set things up in a cunning way so that people will, on the whole, upon balance, both choose option A out of two, say. But also if you ask them which they'd like to reject, they also reject option A. You might think, well, that's just crazy. I mean, can't both prefer and disprefer the same thing? And the way to understand what's happening, which Eldar, uh, realized long ago, and again, I sort of dismissed this as a kind of interesting quirk. But now I think, no, no, it's really not. This is deep. The trick is to create options which have really good features and really bad features and compare them against something middling. So imagine we've got our extreme option. It's a good and bad stuff and middling option. And I say, which would you like to choose? And he said, oh, I've got to choose something. Well, what's the best? Oh, think about good features. And look, this extreme one has got these really good features. I'll choose it. On the other hand, if you ask the same person on a different occasion, which would you like to reject? And you go, I'm rejecting. Now I've got to find some really bad stuff. Oh, the extreme option has some bad stuff. Uh, I'll reject that. So by focusing on the good or the bad based on the way you ask the question, that flips Your apparent preference. And what that's telling you really, is that when you're asking people to give a preference, they're doing it in a really improvised way. They're thinking, I've got a question here. Help. What do I do? Ah, uh, I know I'll look for good features because you asked me which to choose, or I'll look for bad features if you ask me which to reject. But it's just kind of ad hoc. It's like, instead of thinking, let me just consult my stable set of likes and dislikes, uh, let me weigh these in the balance of those great likes and dislikes. Well, this one gets a five and this one gets a six and a half. I'm not doing that at all. I'm conjuring up a, uh, strategy, kind of ad hoc strategy to solve the problem, which is I'm accepting, I'm choosing, I better look for some good stuff. I'm rejecting, I better look for some bad stuff, whatever. And if you ask the question other, if you frame the question in many different ways, you just get answers just bouncing around all over the place. And, uh, we did this kind of stuff with really simple examples. So Eldar has these more rich examples about custody decisions. You have a parent who's really close to the child but really, really has some health issues versus a middling parent, whatever. That works beautifully. But you can also do it with really very, very simple things. Like, we had streams of numbers, and these are like amounts of money. And you're going to choose a stream of numbers and you're going to get randomly one of those sums of money. So it looks like five shoots up and a ten and a minus five and so on. And so this is like really basic sort of really primitive stuff, but you still get the same effect that the extreme option, which has really high numbers and really low numbers. You choose it and you reject it. And this just generally makes one feel, oh, hell, the idea that there's these deep preferences is a bit dubious. And once you start thinking that way, you just see it everywhere that basically the whole field of behavioral decision making research is all about you giving somebody a question which is logically the same in different ways, and they give different answers. And the fact that they give wildly different answers in a particular way. Now, I think, well, obviously that's going to happen because it's not puzzling. It's just that you give me a question, I improvise an answer. I think, well, how am I going to conjure up an answer for this one? And if you prompt me in different ways, I'm going to come up with a different answer. Uh, but in the old days I used to think, oh, there must be a real preference in there somewhere, where is it? And ah, that used to worry me and it worried the field, I think, but just the wrong way to think. The idea that kind of deep down solidity and it's kind of being distorted by sort of superficial characteristics of the question and context and stuff that's just the wrong way to think. The superficial stuff is kind of what's, what intelligence is all about. It's being able to cope with the surface level question you give me now and fire something back at you and then you'll fire something back at me. And that's the cleverness of thought. It's not hunting inside our minds for these deep beliefs and desires and so on.

Speaker B: So what about if someone is holding on to sort of the traditional view of the world and says, well, maybe we do have sort of these deep preferences, but we are spontaneously just choosing a different decision model that when you use in the time. So we're improvising our decision model, not necessarily our preferences. It sounds like there's a max min thing going on there in some decisions. Right.

Speaker A: The natural way people have thought for a long time about this is that there's an underlying kind of purified version of me which doesn't get pushed around by the details of the question and the time of day and the uh, what I've just said in my last answer and all of that stuff which actually seems to matter a lot, that's just noise. And deep down there's something really stable. And I suppose the revelation to me is you can get a much better understanding of humans if you throw away that idea. You just say no, no, the surface stuff is the most important stuff. And the momentary, that's the magic of intelligence is this ability to conjure up things in the moment. And it's true kind of whatever you look at. So a nice set of experiments came from a, uh, former postdoc of mine called Peter Johansson, working with Lars hall at the University of Lund. And this is work that Petra and Lars did before Peter worked for me. So I'd like to be able to claim credit for this. And I have done a bit of work along these lines with Peter, but he and Lars are the initiators of this work completely. Now they have this phenomenon they call choice blindness. So in choice blindness you give people a couple of things to choose or can be many and they make A choice. And then you actually perform a magic trick on it with literal magic. Usually you say, for example, there's a lovely experiment with jams so that people choose various jams and you say, which is your favorite jam? And by using a double ended jam jar, you can actually flip the jam jar over. So you give them a piece of the jam, they taste it very nice, you turn it upside down and they have a bit more. You say, well, just have a bit more and then tell me why you like it. And the trick is that you've now switched the jam. So now you thought you were given raspberry and you thought, I really like raspberry, now I'm given some strawberry. And now I'm going to explain why I like strawberry. And I will do that very happily. I say, oh, yeah, I like this. And the reason I like it is this. It's the wrong one, I don't think. Hang on, no, this isn't what I like. Or even I don't really like it that much now I think about it and that, uh, people will start to say things which are sort of strawberry specific too. They're saying things about the actual thing they're tasting now, saying, I like this because. And this is true for faces. The most famous experiment, which was in the journal Science, was with people will look at faces, which is your favorite face. Using a literal card trick, you choose one of the faces. The person in the experimenter then gives you the other face. And the faces can't be that different. I mean, they have to be like the same gender and same rough ages, otherwise you're going to notice. But people generally don't notice. And then they'll say things like, oh, I really like that hair, or I like those earrings. And that wasn't in the face they actually originally saw at all. And so in these cases, the jams and faces, you're being tricked about what you chose. You're then justifying it, and you're justifying it on the grounds of things that couldn't possibly have had any relation to your original decision. That kind of phenomenon is absolutely everywhere. And another lovely case, which I'll label you with too many, but another lovely case is a lovely work which is looking at split brain patients. So it's a great deal of very brilliant work over many years by looking at people who, because of very severe epilepsy, had the corpus callosum, which is the set of hundreds of millions of fibers that connect your two hemispheres together, had those seven. It's surprising that people can live pretty decent normal lives with relatively intact cognitive function, IQ and so on in most aspects of their life where the two hemispheres of the brain are disconnected. But the magic trick is that they seem to be able to explain their behavior using only the left hemisphere, because the left hemisphere is where language is. So they can give a story about what they're doing, even though only half of their body is being controlled by the left side. So rather weirdly, the left side of the brain controls the right side of the body, but the right side of the brain controls the left side of the body. So anything that the left side of the body is doing, for example, drawing a picture with your left hand, you can't possibly know. The side of the brain with language can't possibly know why it's doing it, but it produces lovely explanations, lovely little stories about what's going on which we know to be false. So the experimenter can give you a picture which goes into the non language side of the brain. Depending on, um, you can do that by positioning it in the left visual field. So you just get some information goes into the wrong side of the brain, Language system doesn't know about it. Then the left hand sets off and draws a little picture of the thing that's just been shown. And you ask the person, why did you draw that? And that just conjure up some crazy story. So they say, oh, I saw one of those on the way in. Or why did you draw a house? Well, it was a lovely house as I was driving in today. And in fact it's because there's a picture of a house which is being projected onto the part of your brain that the language center knows nothing about. Now the language center doesn't say, oh, you've got me there. I'm a bit puzzled. I'm going to see if I can cook something up. It's completely happy. It improvises a story. You can imagine a story about why you're behaving as you are. It sort of conjures it up and there you go, that's the story. The thing that sort of gives me pause about those sorts of examples, which again, I used to think they have kind of weird curiosities, but I'll just put them to one side. The thing that's shocking about them is that the language machine, which is sometimes known as the interpreter in this kind of literature, the language machine, is just conjuring up these stories completely fluently, completely naturally and very richly, even though they're completely false. So it's not as if it seems very unlikely that normally it's able to look back, look inside your head and see, ah, uh, this is what you really thought. This is why you did this. This is what you wanted. I'm going to read that off. But then if it's not there, if there's nothing to read, it kind of makes stuff up. It's much more plausible. That just makes stuff up all the time because it's just as fluent. It's unfazed by the fact that absolutely, uh, nothing whatsoever about what's happening. And of course, we're also very good at explaining other people's behavior too. So I see other people doing stuff. I can tell a story about why they're doing things. Really, it's the same with myself. I find myself doing things and saying things. Why did I do that? Why did I say that? I'll tell you a story, but I'm making that story up. At the time you asked me, I'm, uh, not able to look inside my head and read the story off.

Speaker B: It's fascinating to me. So that whole description of the brain and how we operate mirrors how I would explain LLMs to people. I get the opportunity to talk to leaders and explain how LLMs work, right? And they're improvising as they go. They're drawing on what they do. And I say, our brains work like that, right? When we're having discussion, I didn't stop for 20 minutes and form the sentences I'm going to use. I'm sort of blurting as I go. So to what degree do you think that LLMs are actually more like us than traditional discussions about LLMs talk about?

Speaker A: I think they are a fantastic metaphor for how humans work. They're clearly going to be different in detail because we just created these particular engineering purposes. But the general principle, you can produce incredibly intelligent reasoning and sophisticated language, and for that matter, the ability to interpret images, because you can add those in as well, of course, these days. And, um, ability to create images. The fact that you can do all of this stuff using a system that doesn't have any kind of stable set of beliefs about how the world works or the principles of physics and so on. I mean, you might think classic perspective on computer graphics would be, well, you've got to have these sort of an understanding of the structure of the physical world and how optics works and all of that stuff to be able to generate images, say, or understand images. But it doesn't appear that that's true. At least some of that stuff must be implicit. But it's not codified anywhere. There's no place inside the network. You say, ah, there it is. That's the belief that somebody has about lap travels in straight lines, or things get smaller when they move further away. Oh, uh, there's none of that anywhere. And I think that's just the same with us. I think you're right. The book the Mind Is Flat was published in 2018. So this is before LLMs came along. I kind of wish they'd come along a bit sooner because then I could have said to everybody, well, look, here's a kind of system that works like this. But, yeah, I think you're right. I think we're much more LLM like than we'd like to imagine.

Speaker B: So let's, uh, talk about implications for leaders of a lot of the work you did in the Minus Black. If preferences are improvised rather than stored, what does that say about common leadership tools like surveys, engagement scores or values assessments, where we're asking for someone to answer some questions, are those meaningless? Or how should a leader think about that?

Speaker A: I think they're not quite meaningless, but we should view them as very provisional. And of course, we all know this when we're pulling in these things ourselves. So I think we tend to have a slightly blitz attitude to this, because while I'm giving other people a survey, I want to know what they really think. But of course, when they're giving me a survey, I think I don't really know what I think I've got these questions. What shall I say? Um, let's try a five for this one and a three for that one. So I find myself kind of flapping around, obviously slightly incoherent when I'm answering the questions. And of course, we should just recognize that that's true for us all. And so I think we should view responses people are giving as something to treat with skepticism. It may well be true that in aggregate, if people are saying that they find the cultures of their organization sort of repressive and too bureaucratic or whatever, that's telling you something that's not nonsense. Equally, if you're very happy in their work and they love their colleague, that's telling you something too. But it's really very sort of broad brush, trying to get detailed sense of what people are trying to achieve in their lives or what they believe the value of the vision of the company should be. Or this kind of stuff is stuff that we are all very, very vague about. And in a way, one of the objective aspects, well, one of the positive Things for a leader's point of view is not that you start from the sense that everyone's got these fixed views. They've got a fixed sense of who they are, uh, what they want the organization to be, and you just got to fight them tooth and nail. It's much, much more fluid than that. So if you can provide and indeed collectively build a positive, different vision or a more clear vision, how things should be, how the culture should work, what the organization's there for, it's quite likely people will think, oh yeah, I'll go with that, or maybe they'll add their own spin and add their own ideas. But it's not the case that everyone is kind of died in the world with their own fixed, um, viewpoint to start with because they're quite the opposite.

Speaker B: If you're enjoying this episode, make sure to follow or subscribe so you never miss a conversation. We've got plenty more great stories coming your way. So if everything, if, you know, if we do go to the extreme and like everything is somewhat temporary and improvised in the moment, how do you run an organization without expecting some things to hold, like people like this or don't like this, but the next moment they're going to like this or don't like this. If they don't have, have deep pre existing preferences, then do we need to shape every moment with a different sort of context and just go moment to moment? As a leader, I think the thing

Speaker A: that we are very concerned about as social animals is having basically agreements, A uh, shared sense of we're trying to achieve roughly this in this kind of way. These are the ways we are happy to work together. We don't and don't expect anyone to behave like this and so on. Now these are all somewhat unstable and somewhat vague at the edges and that's okay. So trying to pin everything down in detail in a kind of giant rulebook is usually a doom strategy. Partly because the rulebook will be never endingly long, but partly because what's appropriate is continually evolving and changing anyway. But on the other hand, it's not the case that we are, um, wanting some kind of coherence. So in some ways the very improvised nature of our, uh, interactions means that we're trying to find some kind of structure. It's a bit like we're doing an improvised play. And if we do more and more of the improvised play, people start to develop characters and you think, well, you're playing the stern policeman and somebody else is playing the cheeky bank robber and I want you to keep playing that part, I don't want you to jump around and start playing different parts. Otherwise we're going to get into a tremendous mess. If we're trying to improvise, we need to have a sense of who's doing what rules are. And so our uh, objective as social agents is trying to, to have established these rules just in a way that's precise enough for us to be able to get along. So I think we shouldn't think as a leader. It's not that one's not trying to help create, I mean stable, ish sense of uh, these are the ground rules. But ground rules is probably all we do want really. It's not like we want detailed regulations for everything and uh, people really hate that. But also we have to be responsive to the world as it continually unfolds before us. So I think a mixture improvisation within a framework, but the framework itself can always be renegotiated.

Speaker B: So if I lead a very large organization where everyone is improvising, thousands or tens of thousands of organization, how should I think about the collective decision making of that group? Is there some way a leader of a large country, even a large number of people, how do I align their thinking or get them to work? Is it just making sure everyone has the right improvised roles? Or is there some other mechanism?

Speaker A: I think this is where thoughts about things like sort of vision and culture are uh, crucial. And now vague and often thinking about visions and cultures is somehow we find ourselves. What is the vision of the company? Let's write it down in 50 words or something like that or what kind of cultural. I've been certainly been in companies where people have been trying to write down the cultural values of the company in five key points or whatever. I'm not saying that's necessarily a bad thing, but the point is not trying to distill these things in a kind of completely crystalline form, but as a leader trying to give people a clear sense of this is what we're about, these are the kind of organization we are, these are the kind of goals we have. It doesn't have to be completely fixed and everyone's got exactly the same form of words and in fact it may be better if they haven't. But giving people a sense of this is what we're about that allows people to improvise in a constructive way so they're working alongside each other in a uh, relatively cohesive fashion, all pulling in the same direction. That high level vision is all you can do really. So extremely this is political leaders and Successful political leaders. A large part of what they do, clearly not all of it, but a large part of what they do is this really very broad agenda setting. This is the kind of thing we're trying to, society, we're trying to create. Here are a few key decisions. But trying to micromanage the whole of it is kind of hopeless. Clearly tens of thousands of people, you can't tell them all what to do. All you can say or you can do is set this vision which allows people to align with each other. But I think trying to set it excessively top down is probably usually not going to work terribly well. Um, because it's got to be setting a vision which is something that people are willing to be part of optionally. I remember somebody part of a big organization I used to know well saying, I just can't understand it. I mean, I'm setting out the vision and people just aren't getting on board. I mean, what are they doing? And that's kind of the completely the fatal error of thinking that the leader's role is to say, here's a really compelling vision about what's going to make our organization great. Go on. Then everybody just do it without thinking. Hang on. Why were all these people who are just kind of modeling through their lives, what's in it for them? Why should they be part of this? So the inspirational aspect of visions is really crucial I think, because if you're going to get people, if people are ah, fundamentally worried about just improvising their way through interactions, uh, at the moment that day, just getting through their lives, then any kind of high level insight about what we're trying to do as an organization or as a nation or whatever. There has to be something I can actually connect my actual daily life and think, oh yeah, that kind of makes sense. I'd do something a little bit differently, just a little bit differently because of that. And I kind of want to, otherwise it's just going to be ignored or indeed actively frustrated. I think it's big, nebulous, soft kind of thing trying to lead and there's no formula for it, but some people can do it much better than others.

Speaker B: So if we, on one hand we're saying that people don't have these deep sets of principles and preferences that are driving their decisions. But you're also saying that it's not like all of those decisions are devoid of some amount of preference architecture. Right? Because if I'm sharing an inspiration for the vision, I'm um, trying to get them to factor that into their Improvisation. So it's sort of like. I'm thinking the model is more like a director. We're directors of a scene. Right. If you're a leader and you're trying to get the actors to have this in their mind, right. You're trying to be nice here or short scene sort of suggestions that then the actors go out and do as opposed to trying to change their core belief system. Is that the right mental model?

Speaker A: No, no, I think that's exactly right actually. So I think that theater or movies and so on is actually a very, very good metaphor. So the idea that good direction cannot be telling all the actors what to do, sort of hopeless. What it can do is giving a vision in which the actors can express themselves effectively and work together with each other. So I think that is exactly the right model. It's uh, kind of providing a sort of loose and supportive and inspiring framework within which people are actually able to do their best work.

Speaker B: So it's almost as if, uh, I take that model for leadership and really dig a little deeper on one regard. I think that leaders who tried to spend a lot of time embedding a deep vision inside of their people. Right. But if the model is we need to give them instruction for improvisation, it sounds like we need to connect with them much more often with much smaller bits of information so that we're there in the context of when they're making decisions and embedding it. Is that an implication for leaders?

Speaker A: I think it absolutely, um, is actually. Bruce. Yeah. Because I think the mistake with this sort of very deep vision approach is often that the vision expressed in a very abstract form is viewed as kind of the goal. But in an abstract form it may be very difficult for anyone to actually connect that to the day to day muddling through their day that they actually have to do. So therefore I think it's more like the director, so the director just saying in general, I have a vision for this, my version of Hamlet. Here it is in 50 words. But then someone's got to start up with uh, the opening scene. They've got the first few lines. What are they going to do with those? The goal of the director has to be to set some, give specific enough and open ended enough thoughts that the person who's actually got to deliver the first few lines has some clue that maybe I could try dismiss my work. But it has to be specific enough that it forces them down a particular track. But it gives them a sense of orientation that needs to be. Yes, more soft, small snippets which is kind of an ongoing process rather than here in tablets of stone is the vision for this organization. Off you go and deliver it.

Speaker B: Love to know, like if we think about then sort of the rise of AI and someone could be listening this going well, human beings are just sort of making stuff up as they go. Why don't we just replace them with AI? Right, because we talk to think that way. From your perspective, is there some value to having human beings in organizations even with this improvisation going on over just having everything automated with AI?

Speaker A: Yeah, I think there is. I think the punch of two things. One thing is that it's still the case though. AI can do a lot of smart stuff. Humans are uh, much more flexible and rich I think is the AI systems at the moment. Now of course, who knows, I mean, when we get to chatgpt, uh, 100, where will we be there? No, but I think that it would be a mistake to think that AI has reached artificial, uh, general intelligence or anything remotely like it. It hasn't yet. We don't know how far away that is. I think it might be quite far, but we don't know. But anyway, that's one thing. And it's easy to be fooled because the ability of AI to do something pretty kind of plausible looking quickly and easily is impressive. But most of the time in organizations we don't want something that's kind of decent looking and plausible looking. We want something that's really good. So it's a bit like if you ever get an AI to try and generate some PowerPoint slides, you can look at them and think, well, it kind of looks like a decent presentation, but kind of look through it closely, it's kind of slightly wrong and the images aren't quite right and the order's a bit screwy and it's not really very useful. I think an awful lot of getting close to something that looks kind of close to a decent human effort is really quite far from doing something that's really useful. And humans care about this stuff. We care about reading an article written by an intelligent, thoughtful human rather than something that looks a bit like such an article, but a bit vacuous. That's really a big difference. But the other thing is that of course the world we care about is composed of humans and humans want to have contact and interact with other humans. So even if we could create AIs as, as smart as us, very likely we don't really want to interact with them because they're not like us. There's no human contact from a Machine. So I think the human desire to interact with other humans is not going anywhere. And to create environments and businesses and governments which are oriented around serving the best interests of humans, that's going to be something that requires humans in the loop. Or at least if we don't have humans in the loop, we're in danger of creating a kind of nightmare for ourselves. We're social beings and humans. I'm a human first person. I'm very, very sure that we're going to have lose track of our society completely. We're going to want to have humans deeply abandoned everything we do. We've got to be calling the shots here.

Speaker B: Great. I have to say a lot of our discussion has been around your work in the book the Mind is Flat, which I suggest all of our listeners absolutely go and read. And so I'm wondering from your perspective, what are you most proud of the impact that came from the Mind is Flat and what's the next evolution of it for you?

Speaker A: Yes, that's a good question. So I think that I am most proud of, although it's not by any means unique to the Riley's Flat, but it is true there has been a kind of shift away from the belief that our, uh, beliefs and preferences and mental states generally are stable. I'm just part of that process. I think it helps. It's helped shift the starting point. I think it's something. It does help shift it. Probably, to be honest, much more is the vast rise in modern AI because then you've got the kind of demonstration which you can play with on your phone, as it were, day in, day out. You can see it before you. And that's amazing. But no, I think being part of that sort of reimagining of how the mind M works has been very exciting in terms of next steps. I think the thing that I'm most interested in these days in this area is the process by which people are able to improvise together. So I've written a book with my friend Morten Christiansen, who's a psychologist at Cornell, called the Language Game. And that's all about how humans are able to create languages, often very rapidly, from pretty much scratch. So for people who can't communicate at all with each other will be able. It's been shown that several generations of kids are able to create languages from nothing, as was done in schools in Nicaragua for kids who are deaf. They're able to create complex sign languages pretty much of the same complexity as English in a few generations from nothing. So the ability of humans to create conventions, systems of communication, rules, social norms, out of thin air is amazing. And I think seeing humans as kind of weavers of complex of social rules, that's something that I think is very interesting. Because we're not lone improvisers, we're kind of joint improvisers. And we continually try to think, what are the rules of this interaction? You know, what am I supposed to speak? When are you supposed to speak? What words are we allowed to use? What's the register of the conversation? And all of these things. And we're super sensitive and super good at this. So this process of joint construction I think I see as the kind of foundation for both culture, but also for organizations. When you think about how organizations work and when teams work effectively and when they don't and so on, it's a lot of it is about the fact that we've come to negotiate a set of rules we're all kind of happy with. These are good rules. We like these rules. We all have a common objective, a common sense of how we're going to achieve the objective. I know what I'm supposed to do, you know what you're supposed to do. And when some weird thing happens, we act in a kind of, in an appropriate way. You do what I expected you to do and I do what you expected me to do. Everything's fine. And when organizations are. Or teams aren't working well, then people are continually at loggerheads because they have different idea what rules are. Um, so I think this process of joint construction, that's just super interesting to me. And I think that's the, that's where my interests are in the future too.

Speaker B: Well, and I think you said that when people sometimes don't understand connect, we oftentimes blame the language. But it's not the language, it's the context around, because language is fluid. I love that part of the research you've been working on and publishing on as well.

Speaker A: I think that's dead, right? Yes. The language is a fairly inert thing. If you take away all the things we're trying to do with the language, it's not like the words aren't causing the problem. The problem is they're different perspectives. An interesting illustration of this kind of problem is when you have one organization taking over another. So if a friend of mine used to work for a largest consultancy was taken over by an even larger consultancy, it's a familiar problem to many people. I'm sure the kind of problems you have are, uh, that each side of this takeover are thinking the people on the other side, they're just being deliberately awkward. And what they think is sensible behavior is obviously ridiculous. And that's imagining that the other people are uh, misunderstanding things, deliberately flouting the rules. And in fact they're all really trying to do the best they can. But they got these different. They've been working for years, perhaps decades with a certain set of conventions and particular ways of using words and procedures and so on. And they're finding it baffling that other people are doing things differently. So it's very easy when people are different from us. It's very easy to think, oh, uh, you're just perverse and difficult and awkward rather than thinking, yeah, man. People with created a different set of cultural norms. It's going to take us a while to renegotiate something that makes sense for us all.

Speaker B: Yeah, that's almost like sounds like the fundamental attribution error in steroids. Right? They're being nasty because they're nasty. I'm being interpreted as being nasty, but I'm good and they're just misinterpreting me. Yeah, super. So this has been wonderful. I could keep going for hours. But I think we now need to go into my standard closing questions because I get to give you some of your time back. Let's start with what's one leadership skill you believe will become even more important in the future?

Speaker A: I think the thing that's going to become m most important is this very high level setting of objectives and goals. Because the more we're doing complex knowledge work, the less trying to figure out exactly where organization's going, exactly what its priorities are in any detailed way. That's hard. The knowledge workers at the frontier have to figure this stuff out for themselves. So this really sort of broad directorial sense of this is the kind of organization we are. Off you go. Flourish and figure this stuff out for yourselves. You're brilliant employees. And I think that's the loose directorial style I think is going to become increasingly important.

Speaker B: Great. Second question is if you could wave a magic wand and instantly change one thing in an organization to make it more humanity centric, what would it be?

Speaker A: That's a very good question. I think the thing I would try to do is to various candidates fighting for control of me here. I think I'd probably try to reduce excessive rigidity in organizations so that ideas. Let's try this so that the insights from people uh, across the organization can actually impact what's done. So there's something I see over and over again in organizations, I'm sure we all do, is that, uh, people are actually doing the job. They kind of know what's wrong, they know how to fix it. But the people who are actually trying to fix the problem are miles away in some different corner of the organization or far above, or they're in a consultancy. And so try to allow organizations to fluidly take insights from the people who are actually at the coalface. I think that's my biggest thing.

Speaker B: And what's one final piece of advice you'd like to leave with our, uh, listeners?

Speaker A: I would say be happy with your improvised self. So I think I've spent a long time thinking I only understand what I'm trying to achieve. I don't quite understand how I'm trying to do it. I don't quite know what my values. I've got some broad idea, but it's all a bit vague. This is not good enough. And I think, no, it's okay. You don't have to feel an imposter or uh, you're a failure for not having worked out all the details. That is the magic of human intelligence, that we can cope fluidity and uncertainty. And we can do it by keeping, keeping loose and keeping improvised. So feel okay about your own flexibility and uncertainties and instability of your own mind is actually a plus, not a minus.

Speaker B: And maybe it'll make everyone feel good to know that we're all improvising. You're not alone.

Speaker A: Yes, you're not alone, but we're all at it all the time.

Speaker B: Great. Now finally, where can our listeners go to follow more of your work and find out more of what you're up to?

Speaker A: So there are, uh, three books I'll point you to. So there's the minded flat you've mentioned, which is all about this idea of the mind as an improviser. Uh, the language game is my friend Morten Christiansen on how people improvise language. And a new book coming out with my other friend George Levenstein from Carnegie Mellon University, who's a very well known behavioral economist called it's on youn, which is about how we've been persuaded that many of the deep social problems we suffer from are actually problems of us as individuals rather than deep social challenges. So, uh, watch out for that one coming on January 27th.

Speaker B: Well, I'd love to have you on next year and talk about that book, your Game. That'd be fun.

Speaker A: I would be absolutely great.

Speaker B: Nick, that was fantastic. Thanks for joining the show.

Speaker A: Thanks so much, Bruce. It's been a real pleasure.

Speaker B: Well, I hope you enjoyed our conversation with Nick Chaytor. Now let's dive into the After Show, a segment where I'll reflect on the key takeaways from this episode. Welcome to the After Show. One thing this conversation quietly dismantles is the idea that there's a deep, stable version of ourselves calmly running the show. Nick's argument is that we're much closer to improvisational Systems, more like ChatGPT and other large language models than we'd probably like to admit. We take in a situation, generate a response that fits, and then explain it afterwards as if it was all carefully thought through. It's not that we're shallow, it's that cognition works at the surface. That idea can feel unsettling, but can also be incredibly clarifying, especially for leaders. Here are five insights from the conversation that really stuck with me. The first is that preferences aren't hidden, they're assembled. Nick described studies where people both chose and rejected the same option, depending on how the question is framed. Ask people what they want and they search for good features, ask what they reject and they search for bad ones. The preference flips. The key insight isn't inconsistency, it's that the preference didn't exist until the question forced it into being. The second insight is that the mind explains itself after the fact that some of the most striking examples Nick shared, like choice blindness with jams or faces, show people confidently explaining decisions they never made. They're not lying. The brain's language is simply doing what it always creating a plausible narrative. It's a reminder that when people explain their behavior at work, those explanations are often sincere and still constructed in the moment. The third insight is that, uh, improvisation is the mechanism, not the failure. Nick pushed back on the idea that context and framing distort some deeper, truer preferences. The surface response is the cognition. Humans are built to handle what's in front of them, not to retrieve answers from a mental archive. Once you see that, a lot of so called irrational behavior starts to look like adaptation. The fourth insight is that leadership is about orientation, not installation. One of the most useful frames Nick embraced was leadership as directing an improvised play. Leaders aren't installing beliefs or downloading values, they're creating a shared sense of direction that people can use while improvising. Abstract visions fail when they don't help people in the moment when real decisions are made. Final insight is that culture works when people improvise from compatible rules. Nick talked about why organization, especially during merges, break down each side assumes the other is being difficult or uncooperative when reality they're improvising three different unspoken norms. Language isn't the real problem. Context is. Culture holds when the rules are shared enough to allow coordinated improvisation. So here's the question I want to leave you with. If people are improvising instead of acting from fixed beliefs, how consciously are you directing the environment they're performing in? That's a leadership question worth sitting with. That's it for now. I hope you enjoyed the episode. And remember, you are the flames that can spark humanity at scale. One decision, one conversation, one moment at a time. Thanks for tuning in to humanity at Scale. Redefining Leadership Podcast Remember, true leadership isn't only about achieving short term results. It's about creating a lasting, positive impact on all of the people you touch. Join me next time as we uncover more insights and strategies for humanity centric leadership. Until then, I'm Bruce Temkin encouraging you to lead with purpose and empathy and to be the spark that elevates humanity at scale.