The Dangerous Rise of Scientism

Next, we've been promised another fireside chat, so you'll see the furniture coming back. But it sounds to me more like a master class with another global superstar. Here to discuss the art and science of decision making with Alec Ellison, the founder of OutVest Capital, please welcome the economist, author, and Nobel Prize winner, Daniel Kahneman. So welcome, Danny. Or should I say, welcome home to Jerusalem. So I want to provide a little bit of background for people on your work by telling a little story. So once upon a time, there were experts in fields ranging from medicine to finance to professional sports scouting. And then along came two professors at the Hebrew University, Danny Kahneman here and his partner, Amos Tversky. And they made observations about systematic mistakes we all make as humans. Mistakes you couldn't predict. Effectively, they became connoisseurs of error. Through their work, Danny and Amos became founders and pioneers in the field of behavioral economics. And as you heard, for this work, Danny won the 2002 Nobel Prize in Economics. I could spend our entire time here providing accolades for Danny's work, but I'll just personalize it. I've read his 2011 bestseller, Thinking Fast and Slow, three times. I know that makes you think I'm slow. And Michael Lewis did a piece, a book called The Undoing Project, which I've read twice, because Michael Lewis felt he had to understand who was behind the change in understanding of how, in the specific case, if you're familiar with, in professional sports scouting and money ball. So Danny, your work has been truly disruptive. People here talk about disruption all the time, but you may be the greatest disruptor we've ever had up on the stage here. So let's get into our discussion. We have an audience full of investors and entrepreneurs. They all need to make decisions with less than perfect information. Given all the systematic, cognitive biases you've documented, how can they make decisions when they are necessarily not fully informed and have to use some intuition? Well, all decisions are made with only partial information. And decisions don't have to be perfect in order to work. And most decisions are imperfect and they still work. So what we can do is improve things at the margin. And improving things at the margin can be done in multiple ways. And this is what we're studying now. So let's hear, how are some of the ways to improve decisions at the margin? You know, there is a question about intuition. You know, whether you're for intuition or against intuition, it's absolutely clear that intuition can be marvelous. And it's also absolutely clear that intuition is often wrong. And there are a few things that we know. We know about the conditions under which intuition is likely to be right. And I think we know something about how to improve it. And we know that it's likely to be right if you've had a lot of experience. And if the world is sufficiently regular for that experience to be worth something. So for example, I do not believe in intuition in the stock market. Because the stock market doesn't have the regularity that it takes. But where intuition is worthwhile, it's worth having. And it's worth having in many situations. What you really want to do, I think, is to delay it. It's to delay it until you have all the information. The problem with wrong intuitions is they tend to arise very quickly. They tend to be premature. And you're better off if you collect information first and collect all the information in a systematic way. And only then allow yourself to take a global view and to have an intuition about the global view. This applies in many domains. So delay as much as you can before making that judgment. Let's talk about optimism versus delusion. You view optimism as perhaps the most significant cognitive bias we humans have. People overestimate their abilities and they underestimate the odds they face. Yet you view optimism as such a blessing that, and I quote from your book, if you were allowed one wish for your child, seriously consider wishing him or her optimism. So, and you also refer to optimism as the engine of capitalism. So is there a difference between overconfidence and what I'll call healthy optimism? Well, you know, in the first place, when you look at great successes, great successes, when you look backwards, were always due to somebody being praisedly over optimistic. Or delusional, even. Delusional, actually. You don't get to big successes without taking unreasonable risks. And so if you look eccentric, the best advice to people is don't do it. But the few people who don't follow good advice, they tend to be responsible for the successes. Most of the people who don't follow good advice don't do very well. So on average, optimism, you know, the kind that leads to great successes, on average it tends to lead to failure. But the occasional successes, and that's where we speak of the engine of capitalism. For society as a whole, it's very good that we have crazy optimists. So we've got a bunch of delusional entrepreneurs in the audience, and maybe some delusional investors, but we all benefit as a society. It's kind of almost the opposite of moral hazard in a way. Yeah. So, you know, a related question is, given that there's so much entrepreneurial activity in the technology field here, how can humans properly assess technology which can move at an exponential pace, when we are really being conditioned through evolution to adapt to much slower changes? How do you think about that dynamic? Well, you know, I think there are many people, futurists like Ray Kurzweil, who believe that, I mean, the fact is that technology seems to be developing exponentially. Now people are really not exponential. We've adapted, we're linear, and we're not suited to exponential development. My guess is there's going to be a tremendous amount of dislocation, and the problems are likely to be social, not technological. It's how society is going to adjust to that level of technological speed. But do you think investors may sometimes underestimate the speed of adoption because of this exponential change, when you, when something really catches on, like the videos we saw a few minutes ago? Well, I mean, you know, it's not, I think that investors as a whole probably overestimate the speed of adoption, but the very successful investors are those who underestimate it, you know. It really depends whether you look at things from, you know, X-Anti or X-Post, and you get very different pictures. Okay, I want to turn to to Startup Nation, and I don't think many people here realize the role that Danny has played. How many hands here have been in the Israel Defense Forces? Quite a few. So over 60 years ago, when you were in the IDF, you developed personality tests to replace interviewing, to assess recruits in order to channel them into the proper units and roles, and indeed, as I understand, there was a so-called Kahneman score that people were assigned. And I also understand it was so successful that it survived to the present day with some relatively minor adjustments. So now within the Israeli technology ecosystem, there's a tendency often to recruit people based on which army units they were in, which means that people are often being recruited based on a test that you developed 60 plus years ago. Yeah, well, actually, it wasn't the tests. What I did then was to modify the interviewing system, and the key idea in the modification, and that was about 63 years ago, was to delay intuition, to make the interviewers collect information and attributes one at a time, and really try not to develop a global view of whether this is going to be a good soldier or not until all the information was in. And these days, I'm engaged in writing a book about decision making, where actually our motto is that options and decision making are very much like candidates, and that there ought to be a way of applying what we know about personnel selection, and we know a lot, to decision making in other domains. Now, as I understand, one of the things you were trying to address was interviewers hiring in their own image, correct? So this was a way of turning the judgment into more of what we might call an algorithmic process today, not let the interviewer know exactly. I mean, actually, you know, that interviewing story that you're telling, and my very early experience in psychology, that was my first experience in psychology, is, again, people wanted to have intuitions. The previous interviewing system that we replaced was a system where people just talk to the individual and try to form a global impression. And the problem is, when you're interviewing in that way, you form a global impression much too quickly, and the global impression is likely to be wrong. And if you delay forming a global impression until you collect information on specific topics, you end up with an intuition which is far more accurate. And that was the lesson of 60 years ago, and it turns out that it's been widely accepted in the domain of human resources and personnel selection. And applying it to decision making more broadly is an interesting exercise, and that's the one in which I'm engaged these days. So your message to entrepreneurs who feel a rush to build their teams is to really step back and make sure that they're hiring based on, whether it's formal tests, but real data, delay the decisions, and not the five-minute interview. Well, I mean, you know, that advice of being quite systematic about building a field, it turns out that the most successful companies are really doing that, and they're investing a great deal. Google, I think, is the best example that I know about, about personnel selection, and they are very systematic. They spend a huge amount of money on it. There are multiple interviews. The interviews are all organized and structured and independent. That's a very important part of good decision making, and it's keeping your sources of information as independent as possible. So let's elaborate on this. I think some of your current work is on different ways of tuning out the noise. Yeah. Could you elaborate on that? Well, there is an enormous amount of noise in decision making. I mean, we became aware of that, of the extent of that, actually, in an insurance company in asking, by how much do underwriters who look at exactly the same risk differ in their assessments? And we compared that to the expectations of the executives, and the differences among underwriters were about five times as large as expected by the executives. And it turns out that this is true wherever you go. We now have a sort of saying, wherever there is judgment, there is noise, and there is more of it than you think, because people underestimate the extent to which others agree with them, and underestimate the amount of, the extent of differences very consistently and very systematically. I would suggest, in entrepreneurial environments, taking the time, or in corporate environments, taking the time to almost write down areas of agreement, to avoid this cognitive bias. Well, I mean, you know, the one danger in all of this is you don't want to paralyze yourself by too much analysis, and you don't want to paralyze yourself by too many bureaucratic procedures. So, finding the best way to combine a disciplined approach to decision making with something that is not too bureaucratic, and the decision makers will feel is a help to them, rather than a sort of a bureaucratic constraint, that's a tough exercise. So, changing taxes, Slater-Mao. So, the implications of your work have driven the use of algorithms, again, most famously, or most in popular fame in sports, but again, in finance, medicine, we talked about zebra technologies earlier, there's been a lot of talk of AI here, using of algorithms, algorithms, algorithms, and of course, there's a potential downside in terms of loss of employment in certain fields, but greater employment in others. Do you have a perspective on how fast this might occur, or which fields may be the most vulnerable? Well, I mean, you know, this is happening at a tremendous rate, and it's, and the people are being displaced, and we're at greatest risk of being displaced, I think, are in white-collar professions. So, there are disciplines that are disappearing. Dermatology, dermatological diagnosis is going to be done better on the phone than when people do it. There are forms of cancer that are far better detected by AI than they are by radiologists. A lot of the legal profession, and legal profession, the legal, in the legal field, collecting precedents and collecting relevant laws is something that is going to be automated. So, the extent of this, very likely, I think people are underestimating it, and people in the professions involved think they're unreplaceable, but actually, more people are going to be replaced than think they are. I have one last question. We talked at the outset of your being one of the greatest disruptors of the last generation. What do you want your legacy to be? Is it as a disruptor, or one who gave us insights that we never understood before? The work is still far-ranging. You know, it's a question I've never asked myself. The one thing I would actually like to leave as my legacy is a way to change the way the controversies are conducted in my field, and I would like the controversies, I invented the term adversarial collaboration, which is the collaboration between people who disagree on a way of doing things, and not that there is much hope for it, but that's what I would wish my legacy to be. Agreeing disagreeably. Thank you so much, Danny. What a pleasure.