1 00:00:15,450 --> 00:00:24,580 Speaker 1: Pushkin. Welcome to Cautionary Tales, and today's conversation is with 2 00:00:24,720 --> 00:00:29,220 Speaker 1: a very special guest, David Epstein. David, we are going 3 00:00:29,240 --> 00:00:30,640 Speaker 1: to get into the episode, we are going to talk 4 00:00:30,660 --> 00:00:34,530 Speaker 1: about your book, which is amazing, I've been enjoying it 5 00:00:34,550 --> 00:00:38,250 Speaker 1: very much, Inside the Box, but we can't really go 6 00:00:38,470 --> 00:00:42,950 Speaker 1: anywhere before discussing The Elephant in the Room, which is 7 00:00:42,990 --> 00:00:48,250 Speaker 1: that You and the Pushkin Podfather himself, Malcolm Gladwell, had 8 00:00:48,270 --> 00:00:52,250 Speaker 1: a bit of a beef. What happened? How did you 9 00:00:52,350 --> 00:00:57,560 Speaker 1: and Malcolm nearly come to blows or come to blows intellectually? 10 00:00:57,650 --> 00:01:02,140 Speaker 2: The Pushkin Podfather, I love that. My first book was 11 00:01:02,200 --> 00:01:06,130 Speaker 2: about genetics and athleticism. And when I pitched it, I 12 00:01:06,270 --> 00:01:08,770 Speaker 2: thought I was going to be writing about the 10,000 13 00:01:08,770 --> 00:01:12,130 Speaker 2: hours rule, basically, that Malcolm made famous, which embedded in 14 00:01:12,190 --> 00:01:14,030 Speaker 2: it is the idea that there is no such thing 15 00:01:14,070 --> 00:01:16,650 Speaker 2: as talent. Now, Malcolm doesn't believe that, but the scientists 16 00:01:16,670 --> 00:01:20,330 Speaker 2: whose work that he was making famous do believe that. 17 00:01:21,270 --> 00:01:23,150 Speaker 2: And that's what I thought I was going to write about. 18 00:01:23,470 --> 00:01:26,590 Speaker 2: And after about a year of research, I realized it 19 00:01:26,630 --> 00:01:28,330 Speaker 2: wasn't right. And I had to turn around and write 20 00:01:28,370 --> 00:01:30,530 Speaker 2: a different book. And so I ended up criticizing the 21 00:01:30,530 --> 00:01:34,390 Speaker 2: 10,000 hours rule quite stridently. Malcolm and I got invited 22 00:01:34,730 --> 00:01:37,850 Speaker 2: to debate athletic development at this kind of famous sports 23 00:01:37,890 --> 00:01:40,770 Speaker 2: analytics conference. This was the first time we had ever met. 24 00:01:41,270 --> 00:01:43,430 Speaker 2: I had no idea that I was going to get 25 00:01:43,450 --> 00:01:46,400 Speaker 2: on his radar. And it turned into a great discussion. 26 00:01:46,440 --> 00:01:48,520 Speaker 2: And we were coming off the stage. He said, you know, 27 00:01:48,560 --> 00:01:52,360 Speaker 2: you made some interesting points there. We had both been 28 00:01:53,180 --> 00:01:56,180 Speaker 2: nationally competitive runners. This was in Boston. He said, we'll 29 00:01:56,200 --> 00:01:58,520 Speaker 2: both be back in New York tomorrow. Why don't we 30 00:01:58,600 --> 00:02:01,040 Speaker 2: run together and we'll talk about it in our own time. 31 00:02:01,380 --> 00:02:03,940 Speaker 2: And we basically became weekend running buddies for several years 32 00:02:04,000 --> 00:02:05,620 Speaker 2: and would argue about this stuff on our own time. 33 00:02:05,880 --> 00:02:09,240 Speaker 2: So it became a wonderful relationship based in disagreement. 34 00:02:09,500 --> 00:02:13,450 Speaker 1: Amazing. Is it true that your book, Range, your second book, 35 00:02:14,030 --> 00:02:17,050 Speaker 1: which is another book I absolutely loved, was kind of 36 00:02:17,990 --> 00:02:20,650 Speaker 1: the result of that conversation or that disagreement? 37 00:02:21,310 --> 00:02:21,930 Speaker 3: Very much so. 38 00:02:22,210 --> 00:02:27,690 Speaker 2: So in that first debate, I sort of contrasted Tiger Woods, 39 00:02:27,790 --> 00:02:32,150 Speaker 2: the most famous story of modern development, probably, early specialization. 40 00:02:32,180 --> 00:02:34,700 Speaker 1: He's hitting golf balls when he's two or something ridiculous. 41 00:02:34,720 --> 00:02:36,440 Speaker 2: Yeah, I mean, he started showing off a swing when 42 00:02:36,460 --> 00:02:39,140 Speaker 2: he was seven months old. And by two, he was 43 00:02:39,200 --> 00:02:44,340 Speaker 2: on national television already. And I contrasted that to Roger Federer, 44 00:02:44,440 --> 00:02:48,080 Speaker 2: who played a dozen different sports and didn't specialize until 45 00:02:48,220 --> 00:02:51,380 Speaker 2: later than many of his peers. And my question was, 46 00:02:51,420 --> 00:02:53,380 Speaker 2: which is the exception, which is the norm? And what 47 00:02:53,400 --> 00:02:55,020 Speaker 2: the science showed was that the tiger is actually the 48 00:02:55,060 --> 00:02:56,780 Speaker 2: exception and Roger is the norm. 49 00:02:56,800 --> 00:02:57,940 Speaker 3: So there are lots of different paths. 50 00:02:58,220 --> 00:03:00,620 Speaker 1: I mean, he's a god, but he's a normal god. 51 00:03:01,140 --> 00:03:01,680 Speaker 3: Right. 52 00:03:01,770 --> 00:03:05,750 Speaker 2: His developmental path was representative of the typical path to 53 00:03:05,790 --> 00:03:09,670 Speaker 2: the top. And so when I made that point that 54 00:03:09,770 --> 00:03:12,170 Speaker 2: when we started running together, Malcolm and I, he would 55 00:03:12,210 --> 00:03:14,570 Speaker 2: refer to this as the Roger versus tiger problem. 56 00:03:14,610 --> 00:03:17,130 Speaker 1: That's a very Malcolm way of thinking about these things, 57 00:03:17,210 --> 00:03:17,450 Speaker 1: isn't it? 58 00:03:17,770 --> 00:03:18,770 Speaker 3: Yes, absolutely. Absolutely. 59 00:03:19,450 --> 00:03:21,470 Speaker 2: And that's sort of how the idea for Range began 60 00:03:21,510 --> 00:03:24,930 Speaker 2: to develop. And in fact, the title of the introduction 61 00:03:25,630 --> 00:03:28,650 Speaker 2: is Roger versus Tiger. And that came exactly out of 62 00:03:28,690 --> 00:03:30,500 Speaker 2: a phrase we were using when we were having our 63 00:03:30,540 --> 00:03:31,560 Speaker 2: own discussions while running. 64 00:03:31,920 --> 00:03:35,660 Speaker 1: Now, you have also, slightly unfairly, written a third brilliant book, 65 00:03:36,280 --> 00:03:40,640 Speaker 1: a massive bestseller. And that book is Inside the Box. 66 00:03:41,160 --> 00:03:48,830 Speaker 1: And that book inspired our episode about general magic. So 67 00:03:48,910 --> 00:03:50,940 Speaker 1: we are going to talk about general magic. We are 68 00:03:50,960 --> 00:03:52,920 Speaker 1: going to talk about the power of constraints. We're going 69 00:03:52,940 --> 00:03:55,780 Speaker 1: to talk about all sorts of things, but we're not 70 00:03:55,800 --> 00:03:58,840 Speaker 1: going to talk about anything until we have heard the 71 00:03:58,900 --> 00:04:27,070 Speaker 1: theme from Cautionary Tales. Today, I'm joined by David Epstein. 72 00:04:27,310 --> 00:04:30,790 Speaker 1: He's a journalist, a science writer, an investigative reporter, and 73 00:04:30,830 --> 00:04:34,150 Speaker 1: the author of three best-selling books. And we're going to 74 00:04:34,210 --> 00:04:39,460 Speaker 1: discuss the cautionary tales he's uncovered in the worlds of creativity, science, 75 00:04:40,000 --> 00:04:42,940 Speaker 1: and business. David, welcome to Cautionary Tales. 76 00:04:43,380 --> 00:04:44,520 Speaker 3: Thank you very much for having me. 77 00:04:45,380 --> 00:04:48,720 Speaker 1: I love the book. It's brilliant. It's full of examples 78 00:04:48,840 --> 00:04:53,360 Speaker 1: of the way limitations and constraints are a good thing, 79 00:04:54,180 --> 00:04:59,180 Speaker 1: for our creativity and our happiness and our productivity, you 80 00:04:59,260 --> 00:05:04,920 Speaker 1: start quite early by talking about your own experience as 81 00:05:04,940 --> 00:05:06,140 Speaker 1: a football player at school. 82 00:05:06,580 --> 00:05:09,900 Speaker 2: Yeah, this would be as an American football player. And 83 00:05:09,920 --> 00:05:15,130 Speaker 2: I was playing quarterback when I was in eighth grade. 84 00:05:15,860 --> 00:05:18,090 Speaker 2: And essentially, I threw a ball as hard as I 85 00:05:18,150 --> 00:05:21,570 Speaker 2: possibly could. And on the follow through of the throw, 86 00:05:22,210 --> 00:05:23,290 Speaker 2: my arm snapped. 87 00:05:23,770 --> 00:05:23,830 Speaker 3: Ow. 88 00:05:24,400 --> 00:05:25,720 Speaker 1: That sounds really painful. 89 00:05:26,180 --> 00:05:28,560 Speaker 2: It was very painful, and it's a very unusual injury. 90 00:05:28,580 --> 00:05:31,170 Speaker 2: The only other time I've seen this was with a 91 00:05:31,450 --> 00:05:34,790 Speaker 2: Major League Baseball pitcher during a game, and he had 92 00:05:34,810 --> 00:05:37,990 Speaker 2: to have his arm amputated. We'll never know why it happened, 93 00:05:38,730 --> 00:05:41,890 Speaker 2: but the result was I had to have my arms 94 00:05:42,070 --> 00:05:45,770 Speaker 2: strapped to my torso from shoulder down to my hand. 95 00:05:46,830 --> 00:05:48,990 Speaker 2: My world revolved around sports at the time, right? So 96 00:05:49,010 --> 00:05:49,890 Speaker 2: I was out of sports. 97 00:05:50,430 --> 00:05:50,530 Speaker 1: Yeah. 98 00:05:51,150 --> 00:05:54,110 Speaker 2: But it had some pretty big long-term impacts on my 99 00:05:54,150 --> 00:05:57,450 Speaker 2: life that ended up being positive. So it changed both 100 00:05:57,510 --> 00:06:01,370 Speaker 2: my school and my sports life. So in school, which 101 00:06:01,410 --> 00:06:04,150 Speaker 2: I was not super interested in at the time, and 102 00:06:04,210 --> 00:06:09,510 Speaker 2: the bane of my scholastic existence was French class, where 103 00:06:09,530 --> 00:06:12,850 Speaker 2: we had these tests every week, where you had to 104 00:06:13,250 --> 00:06:16,090 Speaker 2: listen to a recording of a native French speaker, and 105 00:06:16,130 --> 00:06:18,330 Speaker 2: you had a worksheet that had a transcript that you 106 00:06:18,350 --> 00:06:20,570 Speaker 2: were supposed to follow along with. but there were blanks 107 00:06:20,650 --> 00:06:23,570 Speaker 2: in the transcript and you had to catch the words 108 00:06:23,610 --> 00:06:26,070 Speaker 2: as they were flying by and write the blank down. 109 00:06:26,750 --> 00:06:29,990 Speaker 2: And I was mediocre at this. And once I had 110 00:06:30,030 --> 00:06:33,030 Speaker 2: my writing arm, my right hand strapped to my body, 111 00:06:33,050 --> 00:06:36,730 Speaker 2: it was no chance. I realized I couldn't write fast 112 00:06:36,790 --> 00:06:38,390 Speaker 2: enough with my left hand to keep up. So I 113 00:06:38,430 --> 00:06:41,290 Speaker 2: started trying to memorize the words as they flew by. 114 00:06:41,850 --> 00:06:44,190 Speaker 2: And I did that by attaching the words to things 115 00:06:44,270 --> 00:06:46,910 Speaker 2: related to sports. I didn't know the term mnemonic device 116 00:06:46,960 --> 00:06:49,330 Speaker 2: at the time, But that's what I was doing. And 117 00:06:49,450 --> 00:06:51,470 Speaker 2: I found that I was able to memorize every single 118 00:06:51,530 --> 00:06:52,780 Speaker 2: one of them. And then I would just go back 119 00:06:52,850 --> 00:06:55,050 Speaker 2: very slowly with my left hand and write them all in. 120 00:06:55,110 --> 00:06:56,400 Speaker 2: And I started scoring perfect. 121 00:06:56,680 --> 00:06:59,280 Speaker 1: So that's an interesting example, because you could have done 122 00:06:59,320 --> 00:07:03,280 Speaker 1: that at any stage. You could always have used the mnemonics, 123 00:07:03,540 --> 00:07:06,920 Speaker 1: but it just never occurred to you to look for 124 00:07:06,940 --> 00:07:10,460 Speaker 1: that kind of option until the most obvious thing, which 125 00:07:10,480 --> 00:07:14,060 Speaker 1: was just to scribble it down as you go. Once 126 00:07:14,080 --> 00:07:16,260 Speaker 1: that was closed, you had to find another way. And 127 00:07:16,280 --> 00:07:17,400 Speaker 1: that turned out to be a better way. 128 00:07:17,860 --> 00:07:18,320 Speaker 3: Exactly. 129 00:07:18,340 --> 00:07:21,580 Speaker 2: I mean, only once it was so desperate that I 130 00:07:21,640 --> 00:07:24,960 Speaker 2: couldn't do it at all was I forced to look 131 00:07:25,580 --> 00:07:28,380 Speaker 2: for another way. So it was very much necessity being 132 00:07:28,400 --> 00:07:29,320 Speaker 2: the mother of invention. 133 00:07:29,340 --> 00:07:33,340 Speaker 1: And you said it changed your sports career as well. 134 00:07:34,100 --> 00:07:34,660 Speaker 3: Absolutely. 135 00:07:35,300 --> 00:07:38,140 Speaker 2: Because I was barred from contact sports for a year. 136 00:07:38,480 --> 00:07:40,900 Speaker 2: I couldn't do football, so I did cross-country running, which 137 00:07:40,920 --> 00:07:43,800 Speaker 2: I didn't even know was a sport before that, just 138 00:07:43,820 --> 00:07:46,860 Speaker 2: to stay in shape. And it turned out I kind 139 00:07:46,890 --> 00:07:50,040 Speaker 2: of liked it. And I've liked the sense of improvement 140 00:07:50,140 --> 00:07:53,010 Speaker 2: and the guys on the team. And I went back 141 00:07:53,050 --> 00:07:55,450 Speaker 2: to football, basketball, and baseball once I recovered. But long 142 00:07:55,470 --> 00:07:58,230 Speaker 2: story short, I ended up running at university, becoming a 143 00:07:58,270 --> 00:08:01,070 Speaker 2: university record holder. Running is still a big part of 144 00:08:01,090 --> 00:08:03,670 Speaker 2: my life today. And this is something I never thought 145 00:08:03,770 --> 00:08:06,700 Speaker 2: of as a sport that I would have participated in 146 00:08:07,180 --> 00:08:10,540 Speaker 2: until I was forced to experiment with activities that I 147 00:08:10,580 --> 00:08:13,700 Speaker 2: never would have tried otherwise. And it became one of 148 00:08:13,720 --> 00:08:15,600 Speaker 2: the most important things in my life. So to this day, 149 00:08:16,310 --> 00:08:21,410 Speaker 2: I use mnemonics to memorize hour-long keynote talks, and I 150 00:08:21,710 --> 00:08:24,510 Speaker 2: run almost every day of my life. 151 00:08:24,910 --> 00:08:27,330 Speaker 1: I find this absolutely fascinating. There's a classic study of this, 152 00:08:27,350 --> 00:08:29,370 Speaker 1: and we've both written about it, but I'll let you 153 00:08:29,430 --> 00:08:32,270 Speaker 1: tell the story about the strike on the London Underground, 154 00:08:32,290 --> 00:08:35,330 Speaker 1: I think back in 2012, and what this tells us 155 00:08:35,390 --> 00:08:38,070 Speaker 1: about this whole class of constraint. 156 00:08:38,490 --> 00:08:41,490 Speaker 2: Yeah, so this famous study falls very much in line 157 00:08:41,510 --> 00:08:44,870 Speaker 2: with the necessity pushing people to experiment in productive ways. 158 00:08:45,190 --> 00:08:46,550 Speaker 2: And I think that's one of the things that useful 159 00:08:46,570 --> 00:08:50,200 Speaker 2: constraints do is they drive productive experimentation that people would 160 00:08:50,230 --> 00:08:54,140 Speaker 2: not have undertaken otherwise. And in this case, during the 161 00:08:54,160 --> 00:08:57,580 Speaker 2: underground strike, people had to find different routes to work. 162 00:08:58,660 --> 00:09:02,040 Speaker 2: And during the few days of the strike, a significant 163 00:09:02,100 --> 00:09:04,970 Speaker 2: number of people found a better route to work than 164 00:09:04,980 --> 00:09:07,750 Speaker 2: the one that they had been taking for weeks or 165 00:09:07,790 --> 00:09:10,800 Speaker 2: months or years and stuck with it. And I think 166 00:09:10,840 --> 00:09:14,220 Speaker 2: it saved, if I recall, like a collective 1,500 hours 167 00:09:14,380 --> 00:09:16,810 Speaker 2: of commuting time per day or something like that. 168 00:09:17,050 --> 00:09:17,250 Speaker 1: Yeah. 169 00:09:17,970 --> 00:09:20,690 Speaker 2: Simply because the path of least resistance was blocked and 170 00:09:20,710 --> 00:09:24,210 Speaker 2: people were forced into experimentation that turned out to be productive. 171 00:09:24,310 --> 00:09:26,590 Speaker 1: And it's fascinating because it was only a 48-hour strike. 172 00:09:26,680 --> 00:09:30,680 Speaker 1: I mean, it's annoying, but really not a big deal. Yeah. 173 00:09:30,800 --> 00:09:35,680 Speaker 1: And yet that generated, for some substantial minority of people, 174 00:09:37,360 --> 00:09:40,060 Speaker 1: generated information about the best commute that they could then 175 00:09:40,100 --> 00:09:42,220 Speaker 1: go on to use for years. So this sort of 176 00:09:42,320 --> 00:09:47,200 Speaker 1: forced experimentation generates potentially enormous benefits for what is actually 177 00:09:47,220 --> 00:09:50,570 Speaker 1: quite a small, a relatively small inconvenience sometimes. 178 00:09:51,290 --> 00:09:51,750 Speaker 3: Absolutely. 179 00:09:51,790 --> 00:09:54,470 Speaker 2: And so what this all suggests to me is that 180 00:09:54,510 --> 00:09:59,950 Speaker 2: we probably don't experiment enough and that useful constraints can 181 00:10:00,030 --> 00:10:01,370 Speaker 2: drive us to experiment. 182 00:10:01,650 --> 00:10:04,700 Speaker 1: So that's one kind of constraint. Your book's all about 183 00:10:05,130 --> 00:10:08,270 Speaker 1: the importance of different kinds of constraint your book also 184 00:10:08,330 --> 00:10:12,590 Speaker 1: has this amazing story of general magic which inspired our 185 00:10:12,950 --> 00:10:16,890 Speaker 1: recent cautionary tales episode and the issue with general magic 186 00:10:17,390 --> 00:10:21,420 Speaker 1: their problem was not that they weren't experimenting enough and 187 00:10:21,440 --> 00:10:23,460 Speaker 1: if only they had some kind of constraint they would 188 00:10:23,480 --> 00:10:27,380 Speaker 1: experiment more they were experimenting a lot their issue was 189 00:10:27,420 --> 00:10:31,770 Speaker 1: it was a a different failure to constrain what they 190 00:10:31,790 --> 00:10:36,050 Speaker 1: were doing. For those who didn't listen to the episode, 191 00:10:36,850 --> 00:10:38,610 Speaker 1: stop what you're doing now, go and listen to the episode. 192 00:10:38,710 --> 00:10:41,990 Speaker 1: But the gist is it was this startup in the 193 00:10:42,030 --> 00:10:47,490 Speaker 1: late 80s, early 90s. They had bags of talent, bags 194 00:10:47,550 --> 00:10:53,240 Speaker 1: of funding, almost infinite resources, and they kind of They 195 00:10:53,300 --> 00:10:56,200 Speaker 1: squandered it. They weren't disciplined enough. They weren't focused about 196 00:10:56,240 --> 00:10:58,740 Speaker 1: their goal. And in the end, when they finally produced 197 00:10:58,780 --> 00:11:02,230 Speaker 1: a product, it was very disappointing. They wanted to invent 198 00:11:02,250 --> 00:11:06,630 Speaker 1: the iPhone. They did, on paper, invent the iPhone. But 199 00:11:06,670 --> 00:11:11,130 Speaker 1: in practice, it was a very disappointing product because they 200 00:11:11,150 --> 00:11:12,290 Speaker 1: never focused. 201 00:11:12,330 --> 00:11:17,010 Speaker 2: Yeah, more startups die of indigestion than starvation. Because with 202 00:11:17,030 --> 00:11:18,810 Speaker 2: General Magic, not only did they have all the talent 203 00:11:18,850 --> 00:11:20,590 Speaker 2: and all the resources in the world, they had the 204 00:11:20,630 --> 00:11:25,610 Speaker 2: correct vision of the next... few decades of technology. But 205 00:11:25,730 --> 00:11:27,810 Speaker 2: because they could do anything, they did do anything. And 206 00:11:27,830 --> 00:11:29,410 Speaker 2: the project just grew and grew and grew and spiraled 207 00:11:29,430 --> 00:11:29,990 Speaker 2: out of control. 208 00:11:30,610 --> 00:11:32,790 Speaker 1: Yeah. But what do you think they should have done? 209 00:11:32,910 --> 00:11:35,670 Speaker 1: Because there is a view expressed in the documentary that 210 00:11:36,050 --> 00:11:40,070 Speaker 1: basically you can't invent the iPhone in 2007. The technology 211 00:11:40,130 --> 00:11:43,390 Speaker 1: wasn't there. There was nothing they could really have done differently. 212 00:11:43,430 --> 00:11:45,550 Speaker 1: But you don't see it like that, do you? 213 00:11:45,950 --> 00:11:48,460 Speaker 2: No, I think they should have started much smaller. and 214 00:11:48,520 --> 00:11:52,470 Speaker 2: had much smaller goals and touch points en route to 215 00:11:52,710 --> 00:11:57,550 Speaker 2: that longer vision. Since General Magic so vaguely defined their customer, right? 216 00:11:57,570 --> 00:11:59,090 Speaker 2: They had Joe Sixpack. 217 00:11:59,170 --> 00:12:00,110 Speaker 3: Nobody knew who that was. 218 00:12:01,350 --> 00:12:04,550 Speaker 2: That they really ended up building for one another instead 219 00:12:04,570 --> 00:12:07,340 Speaker 2: of having a specific problem they were trying to solve 220 00:12:07,360 --> 00:12:09,800 Speaker 2: for someone. So there's this famous saying that people repeated 221 00:12:09,820 --> 00:12:12,620 Speaker 2: to me in design a few times where people don't 222 00:12:12,660 --> 00:12:14,380 Speaker 2: want a quarter inch drill, they want a quarter inch 223 00:12:14,420 --> 00:12:16,720 Speaker 2: hole in the wall. So it's to focus on what 224 00:12:16,800 --> 00:12:20,070 Speaker 2: problem you're solving, not the cool thing that you're building. 225 00:12:20,230 --> 00:12:21,990 Speaker 2: And General Magic was so focused on the cool thing 226 00:12:22,010 --> 00:12:24,230 Speaker 2: they were building that they tossed in every good idea 227 00:12:24,290 --> 00:12:28,010 Speaker 2: someone had. Whereas even people who are internal, right? Like 228 00:12:28,730 --> 00:12:32,590 Speaker 2: one of their low level service engineers started eBay internally 229 00:12:32,630 --> 00:12:33,570 Speaker 2: and offered it to them. 230 00:12:33,750 --> 00:12:35,290 Speaker 3: And they said, no, no, too small. 231 00:12:35,450 --> 00:12:36,970 Speaker 1: This is never going to work. Yeah. I had a 232 00:12:37,270 --> 00:12:41,360 Speaker 1: slight disagreement with you about General Magic, David, although maybe 233 00:12:41,400 --> 00:12:43,980 Speaker 1: this is a rhetorical disagreement rather than a real disagreement, 234 00:12:44,040 --> 00:12:47,430 Speaker 1: but The thing about general magic is you describe them 235 00:12:47,450 --> 00:12:51,070 Speaker 1: as having no constraints. But actually, the problem they had 236 00:12:51,110 --> 00:12:53,590 Speaker 1: is they did have constraints. We all have constraints. There's 237 00:12:53,630 --> 00:12:56,290 Speaker 1: not infinite money. There's not infinite time. In the end, 238 00:12:56,330 --> 00:13:00,750 Speaker 1: we are all going to die. The problem was that 239 00:13:00,870 --> 00:13:04,370 Speaker 1: all the constraints were long term and they could pretend 240 00:13:04,410 --> 00:13:07,290 Speaker 1: to themselves they didn't have any constraints. And then, of course, 241 00:13:07,330 --> 00:13:08,990 Speaker 1: in the end, they did run out of money and 242 00:13:09,010 --> 00:13:12,440 Speaker 1: they did run out of time. So it's it's not 243 00:13:12,480 --> 00:13:15,710 Speaker 1: that the world without constraints is bad for you. It's 244 00:13:15,790 --> 00:13:18,750 Speaker 1: that the world without constraints is a fiction. And if 245 00:13:18,790 --> 00:13:20,850 Speaker 1: you act as though there are no constraints, they will 246 00:13:20,890 --> 00:13:21,890 Speaker 1: eventually catch up with you. 247 00:13:22,450 --> 00:13:24,050 Speaker 2: No, I think that's a fair assessment. I mean, I 248 00:13:24,110 --> 00:13:26,790 Speaker 2: would say that they didn't really run out of money. 249 00:13:26,810 --> 00:13:29,150 Speaker 2: They had more money than they could spend. They did 250 00:13:29,190 --> 00:13:29,770 Speaker 2: run out of time. 251 00:13:30,010 --> 00:13:30,230 Speaker 3: Yes. 252 00:13:30,970 --> 00:13:35,460 Speaker 2: And I think you're right that we all face constraints. 253 00:13:36,240 --> 00:13:39,860 Speaker 2: And I think it's hard for us to face those sometimes, right? Like, 254 00:13:40,590 --> 00:13:42,580 Speaker 2: like Oliver Berkman writes about this, if we just get 255 00:13:42,640 --> 00:13:44,920 Speaker 2: the right productivity tools in. 256 00:13:44,920 --> 00:13:46,900 Speaker 3: Place, then we can get everything done. 257 00:13:47,480 --> 00:13:49,620 Speaker 2: But maybe that's just a refusal to grapple with your 258 00:13:49,640 --> 00:13:53,040 Speaker 2: own mortality and realize that you have to prioritize ruthlessly. 259 00:13:53,540 --> 00:13:55,580 Speaker 2: And maybe they were a sort of corporate version of that, 260 00:13:56,120 --> 00:14:00,620 Speaker 2: where this illusion of unlimited time and unlimited resources got 261 00:14:00,640 --> 00:14:02,960 Speaker 2: them out of the hard work of having to prioritize 262 00:14:03,000 --> 00:14:07,040 Speaker 2: ruthlessly and decide what they shouldn't be doing. I would 263 00:14:07,100 --> 00:14:10,540 Speaker 2: say two-thirds of the former General Magic employees that I 264 00:14:10,550 --> 00:14:14,760 Speaker 2: interviewed said some version of, I just couldn't figure out 265 00:14:14,840 --> 00:14:18,070 Speaker 2: what not to do. Because there were no clear boundaries, 266 00:14:18,090 --> 00:14:20,830 Speaker 2: no clear problem they were solving, no at least short-term 267 00:14:20,870 --> 00:14:24,810 Speaker 2: money or time constraint. And so it was very hard 268 00:14:24,850 --> 00:14:28,010 Speaker 2: for them to have priorities. And by the way, I 269 00:14:28,310 --> 00:14:30,230 Speaker 2: think it's a parable for the AI age where the 270 00:14:30,250 --> 00:14:33,830 Speaker 2: cost of starting things essentially goes to zero. And I 271 00:14:33,870 --> 00:14:36,990 Speaker 2: think one of the reasons we've seen this unprecedented adoption 272 00:14:37,620 --> 00:14:41,940 Speaker 2: and not commensurate rise in productivity is that it's gotten 273 00:14:42,000 --> 00:14:43,900 Speaker 2: rid of the hard work of deciding what you shouldn't 274 00:14:43,940 --> 00:14:46,540 Speaker 2: be doing. And so I think general magic is kind 275 00:14:46,560 --> 00:14:47,920 Speaker 2: of a lesson for the current moment. 276 00:14:48,560 --> 00:14:52,060 Speaker 1: And I'm gonna pump you for more lessons because there's a, 277 00:14:52,760 --> 00:14:54,820 Speaker 1: you would never describe your book as a self-help book, 278 00:14:54,940 --> 00:14:58,910 Speaker 1: but it is helpful to the self. I just wanted 279 00:14:58,930 --> 00:15:03,290 Speaker 1: to ask for a nugget of advice about how to 280 00:15:04,210 --> 00:15:07,390 Speaker 1: constrain yourself to get more done? Because I know you've 281 00:15:07,410 --> 00:15:10,430 Speaker 1: been following your own advice in writing this book. 282 00:15:11,170 --> 00:15:14,110 Speaker 2: Yeah, I mean, one of the things for me, this 283 00:15:14,200 --> 00:15:16,920 Speaker 2: is going to sound really simple, but I've been using 284 00:15:16,980 --> 00:15:20,420 Speaker 2: something that I call for myself the Hemingway principle. So 285 00:15:20,460 --> 00:15:23,600 Speaker 2: Ernest Hemingway would stop his writing days in the middle 286 00:15:23,620 --> 00:15:26,950 Speaker 2: of a sentence so that he knew the important place 287 00:15:27,120 --> 00:15:30,640 Speaker 2: that he should return to the next day. And so 288 00:15:30,660 --> 00:15:32,300 Speaker 2: I do this at the end of my workday. My 289 00:15:32,360 --> 00:15:35,410 Speaker 2: last act is to designate the important place that I 290 00:15:35,450 --> 00:15:38,290 Speaker 2: will start the next day. And I think this guards 291 00:15:38,350 --> 00:15:41,970 Speaker 2: against the two main ways that workdays can go sideways, 292 00:15:42,330 --> 00:15:47,260 Speaker 2: which are falling into your feeds, some algorithmic feed or 293 00:15:47,280 --> 00:15:50,020 Speaker 2: your inbox in the beginning of the day, and what's 294 00:15:50,040 --> 00:15:52,540 Speaker 2: called the mere urgency effect, which is this finding that 295 00:15:52,580 --> 00:15:55,710 Speaker 2: people will prioritize things that feel urgent, like responding to 296 00:15:55,720 --> 00:15:58,430 Speaker 2: an email, over things that are actually important, even if 297 00:15:58,470 --> 00:16:00,650 Speaker 2: those important things are objectively better uses of time. 298 00:16:01,050 --> 00:16:01,390 Speaker 3: I love it. 299 00:16:01,930 --> 00:16:04,930 Speaker 1: Thank you, David. And we're going to take a handbrake 300 00:16:05,030 --> 00:16:10,360 Speaker 1: turn and discover how constraints are making scientists more honest 301 00:16:10,380 --> 00:16:22,180 Speaker 1: with themselves. We will do that after the break. We're back. 302 00:16:22,650 --> 00:16:27,230 Speaker 1: with me, Tim Harford, and my guest, David Epstein, the 303 00:16:27,410 --> 00:16:32,230 Speaker 1: author of Inside the Box. And we're currently battling against 304 00:16:32,530 --> 00:16:35,990 Speaker 1: some municipal workers who are mowing the lawn outside David's house. 305 00:16:36,030 --> 00:16:39,750 Speaker 1: So we're going to deal with that constraint while we talk. 306 00:16:39,770 --> 00:16:40,770 Speaker 3: Well played, sir. 307 00:16:42,030 --> 00:16:44,090 Speaker 1: But actually, the constraint I want to talk to you 308 00:16:44,130 --> 00:16:48,180 Speaker 1: about now, David, is a constraint on science and the 309 00:16:48,200 --> 00:16:51,680 Speaker 1: way scientists work. Because there seems to be this moment 310 00:16:51,700 --> 00:16:57,720 Speaker 1: where a couple of decades ago, the 1990s, where you 311 00:16:58,100 --> 00:17:02,540 Speaker 1: couldn't turn around without bumping into some exciting new scientific result. 312 00:17:02,740 --> 00:17:06,460 Speaker 1: The newspapers were full of them. And then suddenly there 313 00:17:06,500 --> 00:17:10,170 Speaker 1: came a moment where all the cool results seemed to 314 00:17:10,210 --> 00:17:14,170 Speaker 1: dry up. And you describe this in your book and 315 00:17:14,210 --> 00:17:16,070 Speaker 1: you say it's a good thing. So what happened? 316 00:17:16,570 --> 00:17:19,790 Speaker 2: Yeah, that was exciting times, right? Where it was these 317 00:17:19,830 --> 00:17:23,010 Speaker 2: very counterintuitive findings of like, If I say Florida to you, 318 00:17:23,050 --> 00:17:26,650 Speaker 2: you'll walk more slowly because you unconsciously associate Florida with 319 00:17:26,710 --> 00:17:28,010 Speaker 2: old people and things like that. 320 00:17:28,030 --> 00:17:32,830 Speaker 1: Yes, all of those. The power posing can make you 321 00:17:32,870 --> 00:17:37,920 Speaker 1: feel more confident and adjust your testosterone level. Hurricanes named 322 00:17:38,070 --> 00:17:43,240 Speaker 1: after women are more dangerous than hurricanes named after men, 323 00:17:44,500 --> 00:17:47,980 Speaker 1: because people take hurricanes more seriously if they have a 324 00:17:48,300 --> 00:17:50,560 Speaker 1: scary male name rather than if they have a soft 325 00:17:50,620 --> 00:17:51,240 Speaker 1: feminine name. 326 00:17:51,530 --> 00:17:53,330 Speaker 3: Right, so they evacuate more, right, if it's a male? 327 00:17:53,350 --> 00:17:55,970 Speaker 1: Yeah, and they evacuate. All of this is nonsense. So 328 00:17:56,210 --> 00:17:58,590 Speaker 1: how did this sort of nonsense get published? What was 329 00:17:58,630 --> 00:17:59,450 Speaker 1: going on? 330 00:17:59,650 --> 00:18:02,390 Speaker 2: And let me just add to that, that beyond the... 331 00:18:02,370 --> 00:18:05,030 Speaker 2: Because these kinds of psychology findings were the most famous. 332 00:18:06,030 --> 00:18:08,440 Speaker 2: But if you look at medical findings prior to the 333 00:18:08,460 --> 00:18:13,620 Speaker 2: year 2000, all these big national institutes of health-funded drug 334 00:18:13,660 --> 00:18:18,120 Speaker 2: trials for cardiovascular health were finding positive, amazing, miracle drugs. 335 00:18:18,960 --> 00:18:22,370 Speaker 2: And then starting after the year 2000... almost all of 336 00:18:22,410 --> 00:18:25,170 Speaker 2: those trials, the new trials were negative. 337 00:18:25,670 --> 00:18:26,650 Speaker 3: Drugs are no longer working. 338 00:18:27,110 --> 00:18:28,690 Speaker 2: So it's like if you look at results before 2000, 339 00:18:28,690 --> 00:18:32,480 Speaker 2: it's like most new drugs work. And then after 2000, 340 00:18:32,480 --> 00:18:36,300 Speaker 2: almost no new drugs work. So what happened in 2000? 341 00:18:36,300 --> 00:18:39,640 Speaker 2: It was that in that case, the NIH forced researchers 342 00:18:40,820 --> 00:18:44,480 Speaker 2: to constrain their behavior. They forced them to write down 343 00:18:44,740 --> 00:18:47,320 Speaker 2: their predictions for what they were testing ahead of time. 344 00:18:47,970 --> 00:18:49,370 Speaker 2: And so they had to make a plan ahead of 345 00:18:49,430 --> 00:18:52,490 Speaker 2: time that they had to stick to And that led 346 00:18:52,530 --> 00:18:57,490 Speaker 2: to more negative results, meaning those prior positive results essentially 347 00:18:57,530 --> 00:19:01,050 Speaker 2: were false positives because what researchers were doing was probably 348 00:19:01,110 --> 00:19:05,770 Speaker 2: making some prediction and then it wouldn't come true. But 349 00:19:05,810 --> 00:19:08,400 Speaker 2: they would have this big set of data that they 350 00:19:08,420 --> 00:19:12,060 Speaker 2: would start sifting, looking for correlations. And when they would 351 00:19:12,100 --> 00:19:16,900 Speaker 2: find something, they would retroactively make a prediction. And then 352 00:19:16,940 --> 00:19:18,580 Speaker 2: when they published the paper, it looks like that's what 353 00:19:18,600 --> 00:19:20,380 Speaker 2: they were looking for all along. And I don't think 354 00:19:20,400 --> 00:19:22,380 Speaker 2: there was any ill intent in this because I did 355 00:19:22,420 --> 00:19:26,140 Speaker 2: something like this for my own master's project unwittingly, just 356 00:19:26,180 --> 00:19:29,820 Speaker 2: not understanding. Like it's fine for exploration deciding what to 357 00:19:29,920 --> 00:19:33,560 Speaker 2: then test. But the problem is when you're sifting a 358 00:19:33,580 --> 00:19:37,940 Speaker 2: bunch of data retroactively, you're essentially doing almost an infinite 359 00:19:38,060 --> 00:19:41,120 Speaker 2: number of tests. And for each one of those tests 360 00:19:41,160 --> 00:19:44,320 Speaker 2: you do, the chances of finding a false positive goes up. 361 00:19:44,800 --> 00:19:48,620 Speaker 2: So you're basically guaranteeing that you will find false positive associations. 362 00:19:48,920 --> 00:19:52,950 Speaker 2: So it's really important in order to draw true conclusions 363 00:19:53,450 --> 00:19:56,370 Speaker 2: to make a prediction ahead of time and then. 364 00:19:56,280 --> 00:19:56,820 Speaker 3: Stick to that. 365 00:19:57,100 --> 00:20:00,830 Speaker 2: So if you're testing the drug to, lower cholesterol and 366 00:20:00,870 --> 00:20:02,770 Speaker 2: it turns out that doesn't work, but you're like, well, 367 00:20:02,790 --> 00:20:05,010 Speaker 2: I have all this data. Let's see if it lowered 368 00:20:05,050 --> 00:20:07,420 Speaker 2: something else, heart attack or stroke or blood pressure. Oh, look, 369 00:20:07,820 --> 00:20:08,840 Speaker 2: one thing was lower. 370 00:20:09,140 --> 00:20:10,560 Speaker 1: Let's say we had an experiment where I'm going to 371 00:20:10,600 --> 00:20:13,080 Speaker 1: roll three dice and I've got a red, yellow and 372 00:20:13,120 --> 00:20:16,000 Speaker 1: blue dice and I'm rolling them. and I don't have 373 00:20:16,080 --> 00:20:18,080 Speaker 1: any hypothesis in mind. I just roll them and roll 374 00:20:18,100 --> 00:20:20,400 Speaker 1: them and roll them and roll them. And then I 375 00:20:20,420 --> 00:20:22,540 Speaker 1: start looking and I get my computer to look as well. 376 00:20:23,440 --> 00:20:26,100 Speaker 1: It'll be like, well, it's interesting that the blue dice 377 00:20:26,140 --> 00:20:29,050 Speaker 1: seems to roll odd numbers more and the yellow dice 378 00:20:29,070 --> 00:20:32,510 Speaker 1: seems to roll even numbers more or the green dice 379 00:20:32,710 --> 00:20:35,950 Speaker 1: usually rolls more than the blue dice as long as 380 00:20:35,970 --> 00:20:38,050 Speaker 1: the yellow dice is an even number. But if the 381 00:20:38,090 --> 00:20:40,610 Speaker 1: yellow dice is an odd number, then that pattern is reversed. 382 00:20:40,630 --> 00:20:42,930 Speaker 1: I mean, there'll be something and we know it's chance 383 00:20:42,950 --> 00:20:45,180 Speaker 1: because we're just rolling dice. But there'll be something. 384 00:20:45,300 --> 00:20:47,340 Speaker 2: This is the kind of thing that led to my 385 00:20:47,400 --> 00:20:52,000 Speaker 2: favorite hilariously dark paper that people sometimes refer to as 386 00:20:52,030 --> 00:20:54,410 Speaker 2: the everything in your fridge causes and prevents cancer paper, 387 00:20:55,070 --> 00:20:57,810 Speaker 2: where it gathered up all the results of nutritional studies 388 00:20:57,870 --> 00:21:00,690 Speaker 2: and showed that essentially every food you can think of 389 00:21:00,730 --> 00:21:02,990 Speaker 2: has been found in various studies both to cause and 390 00:21:03,030 --> 00:21:05,800 Speaker 2: prevent cancer, except for bacon, which was found only to 391 00:21:05,840 --> 00:21:06,420 Speaker 2: cause cancer. 392 00:21:06,440 --> 00:21:07,780 Speaker 3: Which was depressing. 393 00:21:07,980 --> 00:21:09,660 Speaker 1: That's a shame about bacon. 394 00:21:09,700 --> 00:21:10,900 Speaker 3: Yeah, important for mental health. 395 00:21:11,100 --> 00:21:13,840 Speaker 1: And there was a famous example of this. which was 396 00:21:13,960 --> 00:21:18,160 Speaker 1: the researcher Brian Wonsink. So tell us about him. 397 00:21:19,100 --> 00:21:23,900 Speaker 2: Yeah, Brian was a star in the field of nutritional research. 398 00:21:24,020 --> 00:21:28,280 Speaker 2: I mean, his papers would make headlines. He was involved 399 00:21:28,580 --> 00:21:31,970 Speaker 2: in creating the guidelines that are given to Americans for 400 00:21:32,350 --> 00:21:36,510 Speaker 2: healthy eating. And he would do these incredibly intriguing experiments. 401 00:21:36,530 --> 00:21:40,030 Speaker 2: The most famous one was the bottomless bowl study, where 402 00:21:40,710 --> 00:21:43,090 Speaker 2: people were given bowls of soup. and just told to 403 00:21:43,170 --> 00:21:46,100 Speaker 2: eat until they were satisfied and then stop eating. And 404 00:21:46,160 --> 00:21:49,560 Speaker 2: some of those people had bowls that were being filled 405 00:21:50,140 --> 00:21:52,800 Speaker 2: secretly from a tube in the bottom slowly as they 406 00:21:52,820 --> 00:21:55,880 Speaker 2: were eating. And the finding was that those people ate 407 00:21:55,980 --> 00:21:59,300 Speaker 2: way more than the people who didn't have the refilling bowl. 408 00:21:59,660 --> 00:22:01,260 Speaker 2: And the idea being that we don't actually know when 409 00:22:01,280 --> 00:22:04,760 Speaker 2: we're satisfied. So things like that. Or we'd have people 410 00:22:04,840 --> 00:22:08,030 Speaker 2: eat way more calories when watching an action movie versus 411 00:22:08,070 --> 00:22:08,670 Speaker 2: a talk show. 412 00:22:09,270 --> 00:22:12,530 Speaker 3: All these kind of intriguing findings. So he was a superstar. 413 00:22:12,570 --> 00:22:15,130 Speaker 2: And then he wrote a blog post about how he 414 00:22:15,170 --> 00:22:17,930 Speaker 2: does his research. And what he was saying was, oh, 415 00:22:18,050 --> 00:22:21,330 Speaker 2: you know, so we do these studies and it'll be like, whatever. 416 00:22:21,410 --> 00:22:23,640 Speaker 2: One example would be, does the cost of a price 417 00:22:23,680 --> 00:22:27,580 Speaker 2: of pizza affect your satisfaction? And, you know, and maybe 418 00:22:27,600 --> 00:22:31,060 Speaker 2: the answer was no. And he framed that as being disappointing. 419 00:22:31,860 --> 00:22:33,480 Speaker 2: But he said, but we've got all this data. So 420 00:22:33,500 --> 00:22:35,980 Speaker 2: I give it to this grad student and I say, well, 421 00:22:36,020 --> 00:22:38,730 Speaker 2: find something, find something positive. And she never fails to 422 00:22:38,770 --> 00:22:41,290 Speaker 2: find something positive. And what he was basically doing was 423 00:22:41,350 --> 00:22:46,720 Speaker 2: describing flawed methods in research where your initial prediction didn't 424 00:22:46,760 --> 00:22:48,720 Speaker 2: work out, so you go start sifting the data to 425 00:22:48,760 --> 00:22:53,040 Speaker 2: find something positive, and you always will because of statistical chance. 426 00:22:53,320 --> 00:22:55,660 Speaker 1: Yeah, and this is the moment where you go, oh, 427 00:22:56,520 --> 00:23:00,770 Speaker 1: maybe these guys really, really didn't understand that this was 428 00:23:00,870 --> 00:23:04,090 Speaker 1: bad research methods, because if they really understood that this 429 00:23:04,150 --> 00:23:07,620 Speaker 1: is not okay... probably wouldn't have written a blog post. 430 00:23:07,960 --> 00:23:11,510 Speaker 3: No, absolutely. And again, I did this without knowing it. 431 00:23:11,560 --> 00:23:13,820 Speaker 2: And then a few years later, as an investigative reporter, 432 00:23:14,280 --> 00:23:18,120 Speaker 2: writing about bad science, I started saying, some of this 433 00:23:18,140 --> 00:23:21,200 Speaker 2: sounds very familiar. Like, I think I did some of this. 434 00:23:22,780 --> 00:23:24,920 Speaker 2: And so obviously, he was unwitting. But that led to 435 00:23:24,960 --> 00:23:27,960 Speaker 2: this scrutiny that, you know, basically, most of his life's 436 00:23:28,000 --> 00:23:32,280 Speaker 2: work was eventually retracted. I mean, he was, his studies 437 00:23:32,300 --> 00:23:35,260 Speaker 2: were cited in like, books by Nobel Prize winners and 438 00:23:35,280 --> 00:23:38,020 Speaker 2: things like that. So he was a big deal. But 439 00:23:38,320 --> 00:23:41,960 Speaker 2: I think the good analogy is it's like what he 440 00:23:41,980 --> 00:23:44,390 Speaker 2: was doing is like a sharpshooter who fires a bunch 441 00:23:44,410 --> 00:23:47,690 Speaker 2: of bullets randomly at a wall and then finds a 442 00:23:47,710 --> 00:23:50,410 Speaker 2: group that are clumped together and draws a target and 443 00:23:50,430 --> 00:23:52,490 Speaker 2: a bullseye around them. And then somebody who comes by 444 00:23:52,590 --> 00:23:54,590 Speaker 2: afterward will say, wow, what a great shooter. But really 445 00:23:54,630 --> 00:23:58,190 Speaker 2: it was just sifting for random chance. And that's what 446 00:23:58,230 --> 00:23:58,900 Speaker 2: Brian was doing. 447 00:23:59,380 --> 00:24:02,320 Speaker 1: So let's just go through then what what has happened 448 00:24:02,520 --> 00:24:07,190 Speaker 1: in science to prevent this it is a problem because 449 00:24:07,210 --> 00:24:10,730 Speaker 1: there is pressure to find results just saying well we 450 00:24:10,750 --> 00:24:13,930 Speaker 1: looked at this and we didn't see anything should be 451 00:24:13,970 --> 00:24:16,490 Speaker 1: publishable not actually very easy to publish that kind of 452 00:24:16,530 --> 00:24:19,650 Speaker 1: result so so instead people are looking for interesting results 453 00:24:19,690 --> 00:24:22,070 Speaker 1: they're looking for eye-catching results and it's a lot easier 454 00:24:22,110 --> 00:24:26,890 Speaker 1: to find them if you hark if you hypothesize after 455 00:24:26,930 --> 00:24:31,900 Speaker 1: the results are known so what constraints were then imposed 456 00:24:31,960 --> 00:24:34,200 Speaker 1: or self-imposed by scientists to prevent this? 457 00:24:35,260 --> 00:24:39,600 Speaker 2: Yeah, I think the biggest one is so-called pre-registration, which 458 00:24:39,760 --> 00:24:43,290 Speaker 2: is writing down your hypothesis ahead of time. So you 459 00:24:43,310 --> 00:24:47,030 Speaker 2: just mentioned Harking, H-A-R-K, hypothesizing after the results are known. 460 00:24:47,890 --> 00:24:50,490 Speaker 2: You have to hypothesize before the results are known and 461 00:24:50,530 --> 00:24:53,110 Speaker 2: then stick to that hypothesis. It doesn't mean you can't 462 00:24:53,150 --> 00:24:57,280 Speaker 2: explore later, but that exploratory analysis gives you ideas for 463 00:24:57,320 --> 00:24:59,740 Speaker 2: something you then have to test. It doesn't give you 464 00:24:59,780 --> 00:25:00,180 Speaker 2: a result. 465 00:25:00,560 --> 00:25:05,080 Speaker 1: So to come back then to this moment when all 466 00:25:05,120 --> 00:25:08,210 Speaker 1: the cool stuff seemed to dry up, that sounds like 467 00:25:08,240 --> 00:25:12,950 Speaker 1: it's bad. Actually, that's the moment where all the imaginary 468 00:25:13,010 --> 00:25:15,850 Speaker 1: results that were just illusions, all the drugs that we 469 00:25:15,890 --> 00:25:19,380 Speaker 1: thought worked and don't work, that started to dry up. 470 00:25:19,770 --> 00:25:22,350 Speaker 1: And that's good news because that's when we stopped believing 471 00:25:22,410 --> 00:25:26,610 Speaker 1: things that were published in peer-reviewed journals that turned out 472 00:25:26,630 --> 00:25:27,570 Speaker 1: to be completely untrue. 473 00:25:28,630 --> 00:25:32,070 Speaker 3: It's great news. And I think one danger is. 474 00:25:33,760 --> 00:25:38,100 Speaker 2: Because of this so-called replication crisis, what it's called, the 475 00:25:38,140 --> 00:25:41,580 Speaker 2: failure of a lot of famous results to replicate, people 476 00:25:41,600 --> 00:25:44,120 Speaker 2: have said, some people have taken that to say like, oh, see, 477 00:25:45,780 --> 00:25:49,980 Speaker 2: don't trust science. All this stuff was wrong. Whereas these 478 00:25:50,050 --> 00:25:54,990 Speaker 2: problems were surfaced by the scientific community. In some cases, 479 00:25:55,050 --> 00:25:57,990 Speaker 2: people criticizing their own work, if not their peers' work. 480 00:25:58,910 --> 00:26:03,690 Speaker 2: And learning lessons and leading to better practices. And in 481 00:26:03,730 --> 00:26:06,730 Speaker 2: that sense, I think it's incredible what the scientific community did, 482 00:26:06,990 --> 00:26:10,910 Speaker 2: take this very hard, painful reckoning to get to a 483 00:26:10,930 --> 00:26:14,090 Speaker 2: better place. And so to me, in that sense, it 484 00:26:14,490 --> 00:26:18,290 Speaker 2: breeds more trust in the scientific community over a decent timeline. 485 00:26:19,270 --> 00:26:20,550 Speaker 2: And the one other thing I think that should give 486 00:26:20,590 --> 00:26:22,990 Speaker 2: people some solace is, in the midst of all this, 487 00:26:23,460 --> 00:26:28,980 Speaker 2: there was a study done where people were given descriptions 488 00:26:29,360 --> 00:26:32,810 Speaker 2: of psychology studies and asked, do you think this result 489 00:26:32,850 --> 00:26:34,910 Speaker 2: will hold up basically? And sometimes they were given some 490 00:26:34,930 --> 00:26:37,450 Speaker 2: of the results also. And these were just like randomly 491 00:26:37,490 --> 00:26:41,010 Speaker 2: selected people. And people were reasonably good at saying like, 492 00:26:41,030 --> 00:26:42,700 Speaker 2: that just doesn't sound like it's going to be true 493 00:26:42,840 --> 00:26:46,760 Speaker 2: and that matching reality. So we kind of score one 494 00:26:46,780 --> 00:26:49,880 Speaker 2: a little bit for common sense and skepticism. 495 00:26:50,140 --> 00:26:52,440 Speaker 1: And I think one of the things that has improved 496 00:26:52,480 --> 00:26:56,540 Speaker 1: in our consumption of science, both scientists and I think 497 00:26:56,740 --> 00:27:00,480 Speaker 1: the wider community of kind of voracious nerds who like 498 00:27:00,520 --> 00:27:03,010 Speaker 1: to read about popular science, we've got a little bit 499 00:27:03,050 --> 00:27:07,390 Speaker 1: more like, hmm, really? And just that moment is very powerful. 500 00:27:08,050 --> 00:27:08,530 Speaker 3: Definitely. 501 00:27:08,610 --> 00:27:11,770 Speaker 2: And especially now, this was a subtext of something you 502 00:27:11,790 --> 00:27:14,950 Speaker 2: said earlier, but now when you see studies, like some 503 00:27:14,970 --> 00:27:17,390 Speaker 2: of the sauna studies, I'm not saying sauna isn't good 504 00:27:17,410 --> 00:27:17,970 Speaker 2: for long-term health. 505 00:27:18,010 --> 00:27:18,530 Speaker 3: It may well be. 506 00:27:18,570 --> 00:27:20,730 Speaker 2: I think there are some real results there. But some 507 00:27:20,770 --> 00:27:24,140 Speaker 2: of the papers will say, you know, sauna is good 508 00:27:24,180 --> 00:27:26,650 Speaker 2: for dementia, but But when you're doing it like between 509 00:27:26,750 --> 00:27:28,350 Speaker 2: four and nine times a month, if it's any more 510 00:27:28,390 --> 00:27:30,940 Speaker 2: or less than nine, then the result doesn't hold up. 511 00:27:31,230 --> 00:27:34,410 Speaker 2: And it's usually the sign that someone has been slicing 512 00:27:34,450 --> 00:27:38,890 Speaker 2: and dicing the data until they've found something positive. So 513 00:27:39,160 --> 00:27:44,060 Speaker 2: I think you're right that that skepticism is useful, even 514 00:27:44,100 --> 00:27:46,180 Speaker 2: if it's been kind of painful at times. 515 00:27:46,880 --> 00:27:50,380 Speaker 1: And just to return to our self-help corner for a moment, David, 516 00:27:51,360 --> 00:27:56,270 Speaker 1: we've learned something about being appropriately skeptical. about scientific claims. 517 00:27:57,210 --> 00:28:02,980 Speaker 1: But is there something else in this idea of pre-registration 518 00:28:03,120 --> 00:28:06,880 Speaker 1: or hypothesizing early that people might use in their own lives? 519 00:28:07,400 --> 00:28:12,780 Speaker 2: So I think the personal takeaway is that we should 520 00:28:12,910 --> 00:28:18,590 Speaker 2: all kind of be pre-registering hypotheses more often. So making predictions, 521 00:28:18,650 --> 00:28:21,570 Speaker 2: say we're trying a behavioral change, we want to exercise more, 522 00:28:21,770 --> 00:28:24,690 Speaker 2: whatever it is we want to do, making a prediction 523 00:28:24,750 --> 00:28:27,890 Speaker 2: about finding a thing to try, making a prediction about 524 00:28:27,910 --> 00:28:31,720 Speaker 2: what it's going to do, and then proceeding and see 525 00:28:31,760 --> 00:28:33,200 Speaker 2: if your prediction was met or not. 526 00:28:33,320 --> 00:28:35,280 Speaker 3: That will improve your learning. 527 00:28:35,860 --> 00:28:38,560 Speaker 2: And this was actually, I wrote about some research where 528 00:28:38,580 --> 00:28:41,310 Speaker 2: this was applied to entrepreneurs, where they were trained basically 529 00:28:41,330 --> 00:28:44,930 Speaker 2: in the scientific method, making a hypothesis for what they 530 00:28:44,990 --> 00:28:48,270 Speaker 2: think their theory of their market is, finding a way 531 00:28:48,310 --> 00:28:53,190 Speaker 2: to test it. And in this study that randomized entrepreneurs 532 00:28:53,230 --> 00:28:55,550 Speaker 2: and startups either to that kind of scientific training or 533 00:28:55,570 --> 00:28:59,200 Speaker 2: just normal market research, those who were trained in the 534 00:28:59,230 --> 00:29:03,340 Speaker 2: scientific method ended up falsifying their beliefs much more often, 535 00:29:03,420 --> 00:29:05,240 Speaker 2: finding out that the value they thought they were adding 536 00:29:05,260 --> 00:29:08,660 Speaker 2: to the market was not right. And the result was 537 00:29:08,680 --> 00:29:10,400 Speaker 2: that they ended up pivoting a lot more often and 538 00:29:10,440 --> 00:29:12,600 Speaker 2: succeeding a lot more often. So I think this idea 539 00:29:12,660 --> 00:29:15,580 Speaker 2: of making a concrete prediction ahead of time, writing it 540 00:29:15,650 --> 00:29:18,570 Speaker 2: down so that you can compare the result to what 541 00:29:18,590 --> 00:29:22,170 Speaker 2: your intuition was, is how you end up educating your intuition. 542 00:29:22,770 --> 00:29:25,630 Speaker 1: Amazing. Well, here is a concrete prediction for you, David. 543 00:29:25,910 --> 00:29:29,690 Speaker 1: You and I will be back after the break to 544 00:29:29,730 --> 00:29:41,600 Speaker 1: discuss creativity and happiness. We're back. You're listening to Cautionary Tales. 545 00:29:41,960 --> 00:29:45,340 Speaker 1: I am Tim Harford, and I am talking to David Epstein, 546 00:29:45,480 --> 00:29:48,720 Speaker 1: who is the writer of Inside the Box, and he's 547 00:29:48,760 --> 00:29:53,920 Speaker 1: also a a self-professed lover of limitations. So, David, I 548 00:29:53,980 --> 00:29:56,860 Speaker 1: am going to start this part of our conversation by 549 00:29:56,940 --> 00:30:00,960 Speaker 1: reading from what is undoubtedly the greatest work of literature 550 00:30:01,640 --> 00:30:05,290 Speaker 1: in human history. I do not like them in a house. 551 00:30:05,910 --> 00:30:08,110 Speaker 1: I do not like them with a mouse. I do 552 00:30:08,130 --> 00:30:10,190 Speaker 1: not like them here or there. I do not like 553 00:30:10,230 --> 00:30:13,210 Speaker 1: them anywhere. I do not like green eggs and ham. 554 00:30:13,910 --> 00:30:18,160 Speaker 1: I do not like them. Sam, I am. So why 555 00:30:18,180 --> 00:30:20,940 Speaker 1: are green eggs and ham important for your ideas, David? 556 00:30:21,700 --> 00:30:25,480 Speaker 2: Green eggs and ham was written by Theodore Geisel, a.k.a. 557 00:30:25,520 --> 00:30:25,820 Speaker 3: Dr. 558 00:30:25,940 --> 00:30:30,190 Speaker 2: Seuss, on a bet that he couldn't write a book 559 00:30:30,610 --> 00:30:31,710 Speaker 2: using only 50 words. 560 00:30:31,810 --> 00:30:31,930 Speaker 3: Yeah. 561 00:30:33,590 --> 00:30:37,650 Speaker 2: And it was that limitation that forced him to experiment 562 00:30:37,690 --> 00:30:41,710 Speaker 2: with rhythm because he couldn't experiment with words. And in fact, 563 00:30:42,790 --> 00:30:45,510 Speaker 2: what led to that bet was a different assignment where 564 00:30:45,550 --> 00:30:47,950 Speaker 2: he was asked to write a children's book using only 565 00:30:47,990 --> 00:30:51,160 Speaker 2: about 200 words from a kid's vocabulary list. And he 566 00:30:51,180 --> 00:30:53,140 Speaker 2: looked at the list, saw there were very few adjectives, 567 00:30:53,940 --> 00:30:57,280 Speaker 2: and made this Seussian complaint to his wife where he said, 568 00:30:57,700 --> 00:30:59,980 Speaker 2: it's like trying to make a strudel without any strudels. 569 00:31:00,000 --> 00:31:00,480 Speaker 3: Yeah. 570 00:31:01,000 --> 00:31:02,260 Speaker 2: And then he just threw up his hands and said, 571 00:31:02,300 --> 00:31:04,180 Speaker 2: I'm going to use the first two rhyming words on 572 00:31:04,200 --> 00:31:06,390 Speaker 2: the list. And the first two rhyming words were cat 573 00:31:07,050 --> 00:31:09,910 Speaker 2: and hat. And that's what forced him to develop this 574 00:31:10,010 --> 00:31:13,090 Speaker 2: rollicking rhythm that he became famous for because the typical 575 00:31:13,310 --> 00:31:18,550 Speaker 2: storytelling path was blocked. And so as some psychologists prefer 576 00:31:18,600 --> 00:31:22,240 Speaker 2: to the green eggs and ham model of creativity, meaning 577 00:31:22,280 --> 00:31:26,620 Speaker 2: that because our brains are wired for convenience, or as 578 00:31:26,640 --> 00:31:29,170 Speaker 2: the cognitive scientist Daniel Willingham says, You may think that 579 00:31:29,210 --> 00:31:30,870 Speaker 2: your brain is made for thinking, but it's actually made 580 00:31:30,890 --> 00:31:33,470 Speaker 2: for preventing you from having to think whenever possible because 581 00:31:33,510 --> 00:31:39,060 Speaker 2: thinking is energetically costly. So without limitations, we will do 582 00:31:39,100 --> 00:31:41,920 Speaker 2: the convenient thing, the path of least resistance. So in 583 00:31:41,980 --> 00:31:44,200 Speaker 2: order to be creative, the best thing you can do 584 00:31:44,300 --> 00:31:48,040 Speaker 2: is block the typical solution that you would reach for. 585 00:31:48,400 --> 00:31:50,750 Speaker 2: And that's what happened with Dr. Seuss and why it 586 00:31:50,760 --> 00:31:52,790 Speaker 2: became known as the green eggs and ham model because 587 00:31:52,810 --> 00:31:55,770 Speaker 2: this shows up widely in creativity research. 588 00:31:56,450 --> 00:31:59,440 Speaker 1: Now- I would love to talk about Keith Jarrett and 589 00:31:59,480 --> 00:32:03,720 Speaker 1: the unplayable piano. You write about this story beautifully in 590 00:32:03,760 --> 00:32:07,420 Speaker 1: your book, but I'm not going to talk about Keith 591 00:32:07,960 --> 00:32:11,370 Speaker 1: with you, David, because loyal listeners can go and listen 592 00:32:11,390 --> 00:32:15,990 Speaker 1: to our classic episode, Bowie Jazz and the Unplayable Piano. 593 00:32:16,290 --> 00:32:19,330 Speaker 1: They can read inside the box. They can hear all 594 00:32:19,370 --> 00:32:23,890 Speaker 1: about it. I also discussed the example in my book, Messy, 595 00:32:24,710 --> 00:32:27,140 Speaker 1: There were movies about it. I mean, it's an amazing example. 596 00:32:27,580 --> 00:32:34,300 Speaker 1: I'm totally sold on your ideas, David. I'm very sympathetic 597 00:32:34,320 --> 00:32:38,420 Speaker 1: to them. But the question that came up again and 598 00:32:38,460 --> 00:32:40,660 Speaker 1: again in my mind as I was reading the book, 599 00:32:41,960 --> 00:32:44,260 Speaker 1: and it's always good to have a book that raises questions, 600 00:32:44,540 --> 00:32:48,540 Speaker 1: is how do we know what a good constraint is 601 00:32:49,840 --> 00:32:54,420 Speaker 1: and what is an unhelpful or unproductive constraint? So in 602 00:32:54,520 --> 00:32:58,290 Speaker 1: our episode about the making of Jaws, the shark that 603 00:32:58,350 --> 00:33:04,050 Speaker 1: nearly ate Hollywood. So the useful constraint that they had 604 00:33:04,170 --> 00:33:07,510 Speaker 1: was that the shark models didn't work. They were trying 605 00:33:07,530 --> 00:33:09,090 Speaker 1: to make a movie about a shark and they didn't 606 00:33:09,170 --> 00:33:13,470 Speaker 1: have a shark. And that forced various innovations. They really 607 00:33:13,570 --> 00:33:16,920 Speaker 1: leaned on John Williams to produce this score. And John 608 00:33:16,960 --> 00:33:20,500 Speaker 1: Williams kind of becomes the shark. The camera would become 609 00:33:20,540 --> 00:33:24,070 Speaker 1: the shark that shoot images. scenes from the perspective of 610 00:33:24,110 --> 00:33:27,370 Speaker 1: the shark. So you don't, the shark doesn't see itself. 611 00:33:28,070 --> 00:33:30,450 Speaker 1: The shark would be hidden. All of these things just 612 00:33:30,510 --> 00:33:32,110 Speaker 1: made a much better movie than if you had this 613 00:33:32,150 --> 00:33:35,640 Speaker 1: great big, you know, gnashy rubber shark in every shot. 614 00:33:36,250 --> 00:33:41,180 Speaker 1: So that was a really useful constraint. But a constraint 615 00:33:41,320 --> 00:33:44,520 Speaker 1: that they were laboring under that I think nearly destroyed 616 00:33:44,540 --> 00:33:48,380 Speaker 1: the movie was time pressure. They really needed to get 617 00:33:48,400 --> 00:33:52,180 Speaker 1: the movie shot before The tourists all showed up and 618 00:33:52,220 --> 00:33:56,260 Speaker 1: then you can't shoot anything because every scene has kind 619 00:33:56,280 --> 00:34:00,460 Speaker 1: of boats in it. Fortunately, the problems with the shark 620 00:34:00,640 --> 00:34:02,410 Speaker 1: kind of gave them more time because they would just 621 00:34:02,470 --> 00:34:06,190 Speaker 1: sit around thinking about the script, thinking about problems in 622 00:34:06,210 --> 00:34:09,510 Speaker 1: the script. And there was this forced idleness. So the 623 00:34:09,590 --> 00:34:14,490 Speaker 1: time pressure was somewhat relieved. So you've got one constraint 624 00:34:14,530 --> 00:34:17,230 Speaker 1: that's great. The shark doesn't work. You've got one constraint 625 00:34:17,250 --> 00:34:19,680 Speaker 1: that is potentially very dangerous, which is you need to hurry. 626 00:34:19,700 --> 00:34:23,050 Speaker 1: You need to hurry. how do you know what's working 627 00:34:23,090 --> 00:34:25,250 Speaker 1: and what's not? What's a helpful constraint and what isn't? 628 00:34:25,980 --> 00:34:29,080 Speaker 2: Yeah, it's really difficult because even thinking of just time 629 00:34:29,160 --> 00:34:34,400 Speaker 2: pressure alone, there's research showing that it can either destroy 630 00:34:34,460 --> 00:34:39,920 Speaker 2: or enhance creative problem solving. And it basically depends on 631 00:34:39,960 --> 00:34:43,710 Speaker 2: how people use it. If they use it to multitask 632 00:34:43,770 --> 00:34:45,450 Speaker 2: and try to get everything done at once, it destroys. 633 00:34:46,030 --> 00:34:46,910 Speaker 3: Creative problem solving. 634 00:34:47,010 --> 00:34:49,210 Speaker 2: If they use it to monotask and become focused on 635 00:34:49,230 --> 00:34:51,500 Speaker 2: the most important parts of a task, it enhances creative 636 00:34:51,540 --> 00:34:55,940 Speaker 2: problem solving. So even that particular constraint is contingent on 637 00:34:56,020 --> 00:34:58,440 Speaker 2: how one uses it, or I think in many cases, 638 00:34:58,480 --> 00:35:03,830 Speaker 2: probably whether one's boss protects them, their focus during that period. 639 00:35:04,830 --> 00:35:06,610 Speaker 2: And so I think broadly speaking, if you have a 640 00:35:06,630 --> 00:35:10,770 Speaker 2: constraint and to say, can I still surprise myself? Then 641 00:35:10,790 --> 00:35:14,470 Speaker 2: there's still value there. If not, then it's definitely bad. 642 00:35:14,490 --> 00:35:16,730 Speaker 2: And you can see this in sort of studies of 643 00:35:16,790 --> 00:35:21,620 Speaker 2: inventions where people will get more creative. Say there's like 644 00:35:21,620 --> 00:35:24,100 Speaker 2: 100 pieces that you can use to make an invention. 645 00:35:24,360 --> 00:35:26,240 Speaker 2: And if people are told, here's 100 pieces, use whatever 646 00:35:26,260 --> 00:35:29,890 Speaker 2: you want, make something, they are less creative than the 647 00:35:29,930 --> 00:35:32,210 Speaker 2: people who are told, here are 20 pieces from the 648 00:35:32,230 --> 00:35:33,450 Speaker 2: set you have to use and you have to make 649 00:35:33,470 --> 00:35:34,150 Speaker 2: a piece of furniture. 650 00:35:34,950 --> 00:35:35,590 Speaker 3: Creativity goes up. 651 00:35:35,850 --> 00:35:38,830 Speaker 2: But if those people are told, here are 20 pieces 652 00:35:38,850 --> 00:35:40,620 Speaker 2: you have to use and you have to make a chair... 653 00:35:41,330 --> 00:35:44,470 Speaker 2: that's when creativity starts going down again, right? They're getting 654 00:35:44,530 --> 00:35:47,740 Speaker 2: both the sort of what you have to do and 655 00:35:47,820 --> 00:35:50,320 Speaker 2: the how you have to do it. And so there's 656 00:35:50,400 --> 00:35:54,180 Speaker 2: not really a lot of room for innovation there. But 657 00:35:54,260 --> 00:35:56,500 Speaker 2: I think it's a balance, right? Like one of the 658 00:35:56,540 --> 00:35:59,750 Speaker 2: first readers of this book was a guy named Ed Hoffman, 659 00:35:59,770 --> 00:36:02,210 Speaker 2: who was the first chief knowledge officer at NASA, kind 660 00:36:02,230 --> 00:36:04,570 Speaker 2: of like the head psychologist in charge of organizational learning. 661 00:36:05,450 --> 00:36:06,710 Speaker 2: And he said, I have to tell you about this 662 00:36:06,810 --> 00:36:11,510 Speaker 2: mission called LCROSS, where... the engineers got about half the 663 00:36:11,530 --> 00:36:13,670 Speaker 2: time and budget they were expecting. 664 00:36:13,910 --> 00:36:15,610 Speaker 3: And so what did they do? Well, they start complaining. 665 00:36:16,610 --> 00:36:19,900 Speaker 2: And then one of their leaders says, well, if we 666 00:36:19,920 --> 00:36:21,540 Speaker 2: were going to get this done anyway, how would we 667 00:36:21,580 --> 00:36:24,980 Speaker 2: do it? And that leads them to say we'd have 668 00:36:25,020 --> 00:36:27,560 Speaker 2: to borrow technology. So they take imaging equipment straight from 669 00:36:27,620 --> 00:36:30,980 Speaker 2: army tanks and engine temperature sensors straight out of NASCAR racing. 670 00:36:31,740 --> 00:36:33,720 Speaker 2: And they built the probe that confirms water on the moon. 671 00:36:34,660 --> 00:36:36,420 Speaker 2: Now they never would have, and that became part of 672 00:36:36,440 --> 00:36:38,270 Speaker 2: their process. They never would have thought that way if 673 00:36:38,290 --> 00:36:39,630 Speaker 2: they hadn't had half the time and half the budget. 674 00:36:40,210 --> 00:36:42,230 Speaker 2: That said, if they had a quarter of the time 675 00:36:42,250 --> 00:36:44,090 Speaker 2: and a quarter of the budget, could they have still 676 00:36:44,110 --> 00:36:44,630 Speaker 2: pulled this off? 677 00:36:45,050 --> 00:36:45,590 Speaker 3: I don't think so. 678 00:36:46,450 --> 00:36:52,170 Speaker 2: So that titration of what becomes sort of an empowering 679 00:36:52,210 --> 00:36:56,590 Speaker 2: constraint that drives you into productive experimentation versus one that 680 00:36:56,610 --> 00:36:59,950 Speaker 2: just makes the task impossible, I think that's where a 681 00:36:59,970 --> 00:37:02,710 Speaker 2: lot of the art is. And if the deadline is 682 00:37:02,750 --> 00:37:06,390 Speaker 2: so short or the resource is so low, that there's 683 00:37:06,450 --> 00:37:09,370 Speaker 2: no flexibility in how you get to the strategic goal, 684 00:37:09,480 --> 00:37:10,880 Speaker 2: then I think clearly it's gone too far. 685 00:37:11,580 --> 00:37:17,100 Speaker 1: For our Valentine's Day episode of Cautionary Tales, I studied 686 00:37:17,320 --> 00:37:22,690 Speaker 1: online dating. And one of the conclusions I came to 687 00:37:22,950 --> 00:37:28,790 Speaker 1: was that the apparently limitless availability of partners was probably 688 00:37:28,870 --> 00:37:33,900 Speaker 1: not very helpful. Because people were endlessly second guessing. Well, 689 00:37:33,920 --> 00:37:35,820 Speaker 1: is this the right person for me? I can always 690 00:37:35,860 --> 00:37:40,000 Speaker 1: go back and find somebody else. And because it seemed 691 00:37:40,040 --> 00:37:42,540 Speaker 1: as though you were unconstrained, in the end, you just 692 00:37:42,800 --> 00:37:49,040 Speaker 1: end up lonely and never finding the partner for you. 693 00:37:49,140 --> 00:37:53,870 Speaker 1: Whereas actually in the pre-internet dating era where you were 694 00:37:53,930 --> 00:37:57,210 Speaker 1: far more constrained, you would meet somebody and you'd kind 695 00:37:57,230 --> 00:37:57,830 Speaker 1: of make it work. 696 00:37:58,550 --> 00:38:00,900 Speaker 2: Yeah. I think this is a problem. This is an 697 00:38:00,940 --> 00:38:04,270 Speaker 2: analogy to everything now where, because our brains are comparison 698 00:38:04,310 --> 00:38:07,750 Speaker 2: engines to some degree, and we can now compare everything 699 00:38:07,790 --> 00:38:10,950 Speaker 2: about our life to the entire internet, and it's kind 700 00:38:10,970 --> 00:38:14,980 Speaker 2: of a recipe for dissatisfaction. There's this short story I 701 00:38:16,100 --> 00:38:19,890 Speaker 2: love by the Japanese writer Haruki Murakami, who, by the way, 702 00:38:19,920 --> 00:38:22,360 Speaker 2: developed his writing style by writing in his limited English 703 00:38:22,400 --> 00:38:24,080 Speaker 2: and reverse translating into Japanese. 704 00:38:24,100 --> 00:38:25,480 Speaker 3: It was an interesting creative constraint. 705 00:38:26,120 --> 00:38:30,340 Speaker 2: But this story... involves a young boy and girl who 706 00:38:31,180 --> 00:38:35,680 Speaker 2: meet and randomly and realize immediately that they are the, 707 00:38:36,040 --> 00:38:38,660 Speaker 2: as he says, the 100% perfect match for one another. 708 00:38:39,760 --> 00:38:41,140 Speaker 2: And so they go on a park bench and they 709 00:38:41,180 --> 00:38:43,390 Speaker 2: hold hands and they talk for hours and it's incredible. 710 00:38:44,600 --> 00:38:46,410 Speaker 2: And then a little doubt creeps in where they say, 711 00:38:46,450 --> 00:38:49,270 Speaker 2: is it really okay for happiness to come this easily? 712 00:38:50,510 --> 00:38:52,730 Speaker 2: And so they agree on an experiment where they will 713 00:38:52,750 --> 00:38:56,640 Speaker 2: both leave And if they really are 100% perfect for 714 00:38:56,700 --> 00:38:58,780 Speaker 2: one another, they'll meet again and they'll get married on 715 00:38:58,820 --> 00:39:01,420 Speaker 2: the spot. And so one of them walks off East 716 00:39:01,460 --> 00:39:03,880 Speaker 2: and the other West and they never meet again. 717 00:39:03,900 --> 00:39:04,480 Speaker 3: Yeah. 718 00:39:04,900 --> 00:39:07,080 Speaker 2: And that's how the story ends. And I think that's 719 00:39:07,120 --> 00:39:09,840 Speaker 2: sort of a parable of this world of, of, of 720 00:39:09,940 --> 00:39:14,410 Speaker 2: endless choice where there's always that little doubt of if 721 00:39:14,430 --> 00:39:17,350 Speaker 2: I just swipe right one more time. And I think 722 00:39:17,370 --> 00:39:22,170 Speaker 2: that can really undermine that ability to compare, to undermine 723 00:39:22,190 --> 00:39:24,180 Speaker 2: the experience of the moment itself. And, And we see 724 00:39:24,200 --> 00:39:27,520 Speaker 2: this in strange ways, right, where people have been getting 725 00:39:27,540 --> 00:39:30,670 Speaker 2: progressively more bored since the introduction of infinite scrolling. And 726 00:39:30,730 --> 00:39:32,590 Speaker 2: scientists who are trying to understand how that could be 727 00:39:33,350 --> 00:39:36,000 Speaker 2: ran an experiment where they randomized people to groups where 728 00:39:36,020 --> 00:39:38,160 Speaker 2: they either had 20 videos they could choose from or 729 00:39:38,540 --> 00:39:40,890 Speaker 2: one video. from that set of 20 that they had 730 00:39:40,930 --> 00:39:43,290 Speaker 2: to watch. And the people who had to watch the 731 00:39:43,450 --> 00:39:45,870 Speaker 2: one were less bored than the people who had the 20. 732 00:39:45,870 --> 00:39:49,150 Speaker 2: The idea being that just the notion that you could 733 00:39:49,190 --> 00:39:51,490 Speaker 2: be doing something else that might be better undermines your 734 00:39:51,510 --> 00:39:54,410 Speaker 2: experience of the moment you're actually living. And I think 735 00:39:54,430 --> 00:39:55,810 Speaker 2: that's incredibly nefarious. 736 00:39:56,050 --> 00:39:59,480 Speaker 1: Well, I think you've just given us unasked for our 737 00:39:59,560 --> 00:40:04,020 Speaker 1: final piece of self-help advice. David, it's been an absolute 738 00:40:04,080 --> 00:40:06,370 Speaker 1: pleasure talking to you. What's next for you? 739 00:40:06,910 --> 00:40:10,030 Speaker 2: Good question, because for the third time I'm saying, after 740 00:40:10,050 --> 00:40:11,810 Speaker 2: writing a book, I'm saying never again, and I've believed 741 00:40:11,850 --> 00:40:14,450 Speaker 2: it every time. And so it takes me a little 742 00:40:14,470 --> 00:40:17,110 Speaker 2: while to recover from it. Yeah. 743 00:40:17,430 --> 00:40:21,560 Speaker 1: I literally, as I'm speaking to you, David, three hours ago, 744 00:40:22,120 --> 00:40:24,840 Speaker 1: I sent off the manuscript of my latest book. And 745 00:40:25,180 --> 00:40:27,920 Speaker 1: I'm in that moment of like, can I bear to 746 00:40:27,960 --> 00:40:30,620 Speaker 1: do this all over again or not? But yes, so. 747 00:40:31,200 --> 00:40:31,600 Speaker 3: Never again. 748 00:40:31,620 --> 00:40:33,840 Speaker 2: And like I said, I've believed it every time. So 749 00:40:33,860 --> 00:40:35,620 Speaker 2: I'm giving myself, I mean, the book just came out, 750 00:40:35,730 --> 00:40:37,730 Speaker 2: so I'm giving myself a little time. But in the meantime, 751 00:40:38,410 --> 00:40:42,030 Speaker 2: I've been experimenting with making longer form videos on YouTube 752 00:40:42,050 --> 00:40:45,170 Speaker 2: that are sort of reported and journalistic to see if 753 00:40:45,590 --> 00:40:48,040 Speaker 2: that can work. And it's been interesting. It's been more 754 00:40:48,080 --> 00:40:50,200 Speaker 2: work than I expected, but one that I made about 755 00:40:50,820 --> 00:40:53,140 Speaker 2: sports skills and Messi just sort of took off. So 756 00:40:53,180 --> 00:40:55,200 Speaker 2: I'm curious to see if that can become a platform 757 00:40:55,750 --> 00:40:57,390 Speaker 2: where I can do journalistic things. 758 00:40:57,650 --> 00:40:58,250 Speaker 3: And I don't know yet. 759 00:40:59,010 --> 00:41:00,910 Speaker 1: So what's the YouTube channel? Where will people find it? 760 00:41:01,710 --> 00:41:05,050 Speaker 2: It's just my name, David Epstein. And they can find 761 00:41:05,070 --> 00:41:07,260 Speaker 2: it on my website, davidepstein.com. There's a link to it. 762 00:41:08,160 --> 00:41:09,940 Speaker 2: My YouTube handle is David Epstein author. 763 00:41:11,080 --> 00:41:11,380 Speaker 3: Excellent. 764 00:41:11,400 --> 00:41:13,060 Speaker 1: And we will, of course, put the link in the 765 00:41:13,100 --> 00:41:15,820 Speaker 1: show notes. David, it's been an absolute pleasure to have 766 00:41:15,860 --> 00:41:18,160 Speaker 1: you on the podcast. Give us the title of your 767 00:41:18,200 --> 00:41:20,030 Speaker 1: book once again so people can find it. 768 00:41:20,680 --> 00:41:23,410 Speaker 2: Inside the Box, How Constraints Make Us Better. 769 00:41:24,080 --> 00:41:26,820 Speaker 1: And people might also wish to check out The Sports 770 00:41:26,860 --> 00:41:31,480 Speaker 1: Gene and Range, David's two other books. David, thank you. 771 00:41:31,720 --> 00:41:35,080 Speaker 1: And I will be back next week for another Cautionary Tale. 772 00:41:44,400 --> 00:41:48,080 Speaker 1: Cautionary Tales is written by me, Tim Harford, with Andrew Wright, 773 00:41:48,520 --> 00:41:52,610 Speaker 1: Alice Fiennes and Ryan Dilley. It's produced by Georgia Mills 774 00:41:52,690 --> 00:41:57,030 Speaker 1: and Marilyn Rust. The show also wouldn't have been possible 775 00:41:57,070 --> 00:42:01,110 Speaker 1: without the work of Jacob Weisberg, Eric Sandler, Carrie Brody, 776 00:42:01,790 --> 00:42:07,660 Speaker 1: Keira Posey and Owen Miller. Cautionary Tales is a production 777 00:42:07,700 --> 00:42:11,840 Speaker 1: of Pushkin Industries. If you like the show, please remember 778 00:42:12,100 --> 00:42:15,320 Speaker 1: to share, rate and review. It really does make a 779 00:42:15,340 --> 00:42:18,400 Speaker 1: difference to us. And if you want to hear it 780 00:42:18,880 --> 00:42:23,900 Speaker 1: ad-free and receive a bonus audio episode, video episode... and 781 00:42:24,020 --> 00:42:28,210 Speaker 1: members-only newsletter every month, why not join the Cautionary Club? 782 00:42:29,150 --> 00:42:34,820 Speaker 1: To sign up, head to patreon.com slash cautionary club. That's patreon, 783 00:42:35,140 --> 00:42:40,540 Speaker 1: p-a-t-r-e-o-n dot com slash cautionary club.