The Exchanges

Q So now you're the co-founder for the Center for Applied Rationality. Can you, what does that do? I mean, what are you guys trying to accomplish in

A Yeah, so our mission is to develop, uh, strategies to improve on the default human processes for reasoning and making decisions, um, and to, uh, educate people in those, uh, improved reasoning and decision-making strategies, uh, and particularly with an eye towards having a positive impact on the world. So the timing of founding the Center for Applied Rationality, which we call CIFAR, Uh, is basically we had noticed that the science on human rationality, or more precisely human irrationality, had kind of hit this peak, um, over the last few decades. Um, if you've heard of Daniel Kahneman, uh, who wrote the bestseller Thinking Fast and Slow, uh, he won a Nobel Prize for his work on human irrationality, uh, and inspired a lot more work on similar topics. Um, and there was just, you know, the academic research had, had sort of been building up for the last 50 years, kind of culminating in Daniel Kahneman's research, uh, Nobel Award, Nobel Prize Award. Um, and then the public interest in the topic had been growing over the past 10 years or so, um, with books like Thinking Fast and Slow, but before that, Dan Ariely's, uh, Predictably Irrational, um, Nudge, lots of other books on behavioral economics and, uh, and Cognitive Science. And the thing was, there was all this research on how the brain is irrational, um, what kinds of systematic reasoning and decision-making errors humans ma…

AI assessment note: “our mission is to develop, uh, strategies to improve on the default human processes”

Q So in, in your, your view of rationality, then what's the role of, um, intuition?

A Well, so I know a common conception of rationality, um, in the public, in the media, is that rationality means, uh, dismissing or suppressing your intuition. There's often this dichotomy set up of, you know, reason and rationality on the one hand, and then intuition and, and emotion on the other hand. Um, and so, you know, This, this sort of perceived dichotomy gives rise to characters in movies or TV shows that are sort of explicitly the rational one or the logical one, and what that means in practice, you know, for the characters, uh, actions is that they, they're the one who goes around, you know, poo-pooing other people's, uh, thoughts and feelings and, and looking down their nose at people for, you know, having fun or, you Or falling in love, or, uh, or finding something beautiful. Um, and that is not what we or, or cognitive scientists, uh, deem rationality. Uh, the, the actual scientific model, which I think is, um, is also pretty common sense when you think about it, is our brains are, our minds are divided. We do have a more intuitive side of our mind, which, uh, cognitive scientists call system one. Um, it's kind of a dull name, I apologize for that. And then we also have this more recently evolved part of our mind, um, that allows us to do logical reasoning, um, weigh abstract trade-offs against each other, do math, uh, long-term planning, that kind of thing. Um, it'…

AI assessment note: “system one is, is indispensable. There's no way we could actually Survive”

Q Is there more than one type of rationality?

A Yeah. So, uh, sometimes we talk about two types of rationality. On the one hand, we have epistemic rationality. And sorry, these aren't just CFRS terms. These are terms in cognitive science and philosophy. Uh, right. So epistemic rationality is about, uh, using processes, processes for reasoning or processing information that systematically get you closer to an accurate model of how the world works. And, you know, we can never be 100% confident and get the, the Perfect, perfectly correct answer about everything, about how the world works, because, you know, we just have limited information and time, and there's a lot of uncertainty, but some reasoning, uh, processes are just more reliable than others. So, I mean, to take a dumb example, uh, making stuff up, or believing what a random person on the street tells you will be a less reliable process for having an accurate model of the world than synthesizing the opinions of top experts on a topic. Um, or looking at randomized controlled trials or something like that. And that's kind of an oversimplified, uh, description, but I'm just trying to sketch out the, the, what the spectrum of epistemic rationality looks like. Uh, should I go on to the second version?

AI assessment note: “Yeah. So, uh, sometimes we talk about two types of rationality.”

Q Did your parents do anything else to kind of encourage that in you? Or like when you changed your mind, did they... Uh, were they responding positively to that, or?

A I think so. That's less of a vivid memory in my mind, but I, if I had to guess, I would say yes. Uh, they, uh, I mean, my dad's, well, both of my parents, but my dad has more of a background in science. Um, my mom's background is in math, uh, and statistics, actually. Um, so my dad was the one I would more often ask questions about, you know, why is the sky blue? That kind of question about the world. Mm-hmm. Uh, and he rarely just gave me the answer. He would more often sort of lead me through a reasoning process in kind of a Socratic type way. Uh, and you know, usually I couldn't actually reason my way from first principles as a seven year old to figuring out why the sky is blue, but I could sort of come up with some guesses and maybe I could sort of notice, oh, maybe that guess doesn't make sense, et cetera. So when I finally got the answer, it was sort of more satisfying because I had tried to You know how when you try to solve a puzzle and someone tells you, tells you the answer, it's much more satisfying and it sticks with you more than if they just tell you the puzzle and then tell you the answer straight away.

AI assessment note: “He would more often sort of lead me through a reasoning process”

Q Do you think that that's a good strategy if you don't know what's likely to lead to success, but you do know what's likely to lead to failure is just eliminating the stupidity?

A Uh, it's a, it's low hanging fruit, I think. Um, I mean, so often in my work at CIFAR, I guess I didn't get to fully finish explaining what we do, but the short rest of the explanation is that we, uh, in addition to sort of doing some of our own research on these topics, we run workshops where we, uh, run training sessions, teaching people some of the more promising techniques that we've been developing to help overcome some of these biases and make better decisions. Uh, and, and so often I find myself in this situation where someone is kind of, they're at some crossroads, uh, you know, they're, they're at their job and they don't really like it and they're not motivated. They're not really doing good work at all, but they don't really know what they should be doing instead. And they're having a lot of trouble figuring out what the best choice is. And so usually my answer in that case is, okay, I don't know what the best choice is. I don't think you have any way to know what the best choice is. But clearly the thing you're doing now is not the best choice. You're, you're, you know, if, if the best choice is staying at your job, then it means staying at your job and finding a way to make it actually fulfilling and to make yourself actually succeed at it. The best choice is not doing, you know, a half-assed job at your current job. So we can kind of eliminate your current strateg…

AI assessment note: “it's a, it's low hanging fruit, I think.”

Q and then one person in the meeting chimes up and says their grandmother, you know. Uh, has been smoking for 99 years and doesn't have cancer, and then it just gets dismissed and washed under the table, right? So, how do you, how do you get to a point where people, um, themselves are motivated, I guess, in some way, um, to change their minds, or be open, more open-minded?

A Right. So, I guess I have this, this mental model, uh, it's sort of a two-layer mental model where the, the bottom, the more fundamental layer is, um, Just abstractly, but sincerely wanting to, to be able to change your mind. And that doesn't mean that in any given situation, like if you, you know, read an article saying that caffeine, uh, well, that's a bad example because nutrition data is not very reliable, but let's say you read a reliable article showing that, a reliable study showing that caffeine was really bad for you, but you've cherished your morning cup of coffee or, you know, your three cups of coffee a day. That is going to be hard to hear, and the natural reaction is to find a reason to dismiss the study, and that's often going to happen even if you have that fundamental layer in place of believing in general that you would like to be the kind of person who changes your mind. The distinction I was trying to make, uh, with the, uh, in my response to the closed-minded phrase is I think a lot of people just don't have that fundamental bottom layer. Like, they don't even want to be able to change their mind in particular situations. I think it's, um, it's like, I'm pretty pessimistic about the ability to get it to work in a particular situation. So I tend to focus on the, that bottom layer first. And to get that in place, again, I don't think it's the kind of thing yo…

AI assessment note: “I don't think it's the kind of thing you can sort of do to someone else”

Q Um, so to kind of come back to it how do we, um, change other people's minds then?

A Yeah, so there's, there are different strategies to take depending on how intellectually honest you want to be, right? So, you know, a lot of the research that's come out, the books that I've been referring to on irrationality, um, that describe, you know, all of these hidden forces that affect our decisions that we're not even aware of, like, you know, how tall was the person making the claim? Or, you Was I holding a hot cup of coffee when, when the person asked, you know, made a request of me, that sort of thing. Right. Um, all these insidious aspects of, of our psychology that we're not even aware of, um, there's a real opportunity there for people to exploit those things and use them to change other people's minds, um, without actually making any good arguments or presenting, you know, any evidence. Um, and in fact, there's a good book about this I, I would recommend called Influence. By, uh, Cialdini. He's a psychologist. Uh, I can't remember his university now. It's been out for a few decades. It's been a bestseller for a long time. Um, but basically what he did is he went undercover, uh, at, uh, various, uh, companies who were in the business of persuading people, of influencing people. So I think that included a telemarketer, uh, some, some door-to-door sales. This was, you know, back in the seventies or eighties when that was more of a thing.

AI assessment note: “there's a real opportunity there for people to exploit those things and use them”

Q Yes. If only they saw the world through my eyes, right?

A Yeah. It's less interesting to me. Um, but I'll, I'll still give my best answer to it. Um, so to change your own mind, I mean, the first step I would say is just believing that that's a desirable thing to do, which many or most people don't really, uh, either would explicitly say is a bad thing to do, because changing your mind makes you kind of, uh, wishy-washy or weak or stupid, um, or they would explicitly claim it's a good thing to do, but they don't really believe it. Uh, on a sort of gut level, and so, you know, they don't have the motivation to do it, and therefore, they're not gonna do it. Uh, so, step one is actually believing, uh, yes, uh, I believe on a gut level that there's probably a bunch of things I'm wrong about that I, I'm not yet aware of, um, just because that's true of everyone, and also that whatever it is I'm wrong about, I would like to know about that, because I'll probably be able to make better decisions for myself and for other people, Um, I'll probably be able to avoid hurting other people if I have a more accurate model of the world, and so the implication of that is I would like to change my mind when, when I encounter new evidence. In fact, I want to seek out new evidence that might cause me to change my mind, not just sort of passively accept it when it comes.

AI assessment note: “It's less interesting to me... so to change your own mind”