WEBVTT - What Victorian Bootmakers Can Teach Us About The Future of Work | Diving In 0:00:00.240 --> 0:00:04.320 Dario Emode runs anthropic and if he recently warned that 0:00:04.480 --> 0:00:08.480 AI could wipe out half of all entry level white 0:00:08.560 --> 0:00:12.320 collar jobs within five years, he even reckons it could 0:00:12.320 --> 0:00:16.680 push unemployment to ten percent, maybe even twenty percent. And 0:00:16.760 --> 0:00:21.759 yet today I want to talk about bootmakers in Victorian England. 0:00:28.800 --> 0:00:32.760 Hey bear with me. You see, I reckon there's useful 0:00:32.840 --> 0:00:36.640 lessons buried in our economic history that might help each 0:00:36.680 --> 0:00:39.720 of us think a bit more clearly about our AI 0:00:39.920 --> 0:00:43.519 fueled future. I'm just back from a conference where I 0:00:43.600 --> 0:00:47.239 saw a terrific news paper by Hilly River Pond, a 0:00:47.360 --> 0:00:50.559 brilliant young economic historian, And what I want to do 0:00:50.600 --> 0:00:54.279 today is share the story she told. Say. Here's how 0:00:54.280 --> 0:00:56.400 we're going to do it. Five chapters. First, we're going 0:00:56.440 --> 0:00:59.000 to go back to Victorian England. Join me on a journey. 0:00:59.000 --> 0:01:01.400 We're going to watch a new technology hit in the 0:01:01.480 --> 0:01:05.200 established industry. Second, we're going to look at how better 0:01:05.240 --> 0:01:09.680 technology meant cheaper boots, and cheaper boots meant more boots. Third, 0:01:09.920 --> 0:01:14.000 I'll show you how that translated into more boots keeping 0:01:14.200 --> 0:01:17.600 the same number of jobs. That sounds calm, but beneath 0:01:17.600 --> 0:01:22.520 the surface, the tasks that people were doing were completely remade. Fourth, 0:01:22.800 --> 0:01:25.400 we're going to investigate how the geography of this new 0:01:25.400 --> 0:01:29.080 work shifted. And finally, we're going to ask whether any 0:01:29.120 --> 0:01:33.800 of this tells us anything useful about AI. Okay, let's 0:01:33.800 --> 0:01:37.560 get into it. Let's begin back in the middle of 0:01:37.600 --> 0:01:40.560 the nineteenth century. Come back in time with me, and 0:01:40.600 --> 0:01:43.120 if you want a literary reference for the mood, think 0:01:43.200 --> 0:01:48.120 Charles Dickens, not Downton Abbey. Boot making was huge, had 0:01:48.160 --> 0:01:51.360 about two hundred and twenty thousand workers in eighteen fifty one. 0:01:51.920 --> 0:01:55.400 It was the fifth largest occupation in England, and roughly 0:01:55.480 --> 0:02:00.919 one in thirty Englishmen made boots. Back then, boot making 0:02:01.000 --> 0:02:03.920 was craft work. A bootmaker cut the leather shape, that 0:02:04.040 --> 0:02:07.120 stitched the upper together, attached the sole, and finish the product. 0:02:07.400 --> 0:02:10.679 This was skilled, hands on work. A boot maker might 0:02:10.800 --> 0:02:15.040 know the whole trade, from leather to lacers. Then that 0:02:15.160 --> 0:02:19.720 world got hit by a radical new labor saving technology. 0:02:20.040 --> 0:02:24.880 The breakthrough was the adaptation of the sewing machine to 0:02:25.040 --> 0:02:27.760 heavy boot leather, and that occurred in the late eighteen fifties. 0:02:28.280 --> 0:02:31.920 A machine could stitch the leather up together much faster 0:02:32.000 --> 0:02:35.639 than the person could by hand. Workers. Using this new technology, 0:02:35.919 --> 0:02:40.119 they could produce four times as much as workers doing 0:02:40.160 --> 0:02:43.440 the same task the old way. And here's the think 0:02:43.480 --> 0:02:46.799 that first breakthrough, it didn't stay contained to that one 0:02:46.840 --> 0:02:50.240 stage of production. Once the stitching got much faster, the 0:02:50.280 --> 0:02:53.000 rest of the process became the bottleneck. So there was 0:02:53.040 --> 0:02:55.840 pressure to split up the cutting and the sole attaching, 0:02:55.919 --> 0:02:59.200 the finishing and the packing the whole production line, and 0:02:59.280 --> 0:03:03.079 the old draft systems started giving way to something more mechanized, 0:03:03.720 --> 0:03:08.000 more factory based. For the boot makers, this was not 0:03:08.160 --> 0:03:11.639 an abstract future of work seminar, and believe me, I 0:03:11.800 --> 0:03:15.560 sat through that seminar. This was their trade, their skills, 0:03:16.200 --> 0:03:19.880 their way of life, their livelihoods. So they fought back. 0:03:20.600 --> 0:03:23.320 In the main boot making districts, thousands of workers went 0:03:23.360 --> 0:03:25.960 on strike for over a year trying to stop this 0:03:26.120 --> 0:03:31.120 technology from remaking their industry. So when new technology arrived, 0:03:31.400 --> 0:03:34.920 it made a central task dramatically faster, and an entire 0:03:35.000 --> 0:03:40.440 industry finds itself having to work out what to do next. Now, 0:03:40.440 --> 0:03:42.320 I don't want to jump straight to what happened in 0:03:42.320 --> 0:03:44.680 the labor market. I want us to pause first and 0:03:44.720 --> 0:03:49.000 sip with the market for boots first. If a new 0:03:49.040 --> 0:03:52.320 technology lets each work and make more boots per hour, 0:03:53.080 --> 0:03:55.680 then the cost of making boots falls. If the cost 0:03:55.680 --> 0:03:59.000 of making boots falls, then boots tend to get cheaper. 0:03:59.400 --> 0:04:02.920 And when boots get cheap at more people buy boots. 0:04:03.440 --> 0:04:08.840 That last step it really matters because if you listen closely, 0:04:08.960 --> 0:04:11.680 a lot of the discussions of automation they skip from 0:04:12.080 --> 0:04:14.400 this saves labor straight to it. Therefore, there'll be fewer 0:04:14.440 --> 0:04:18.400 jobs that leaves out the work done by markets and prices. 0:04:18.920 --> 0:04:22.359 You see, prices adjust and when they do, they stimulate 0:04:22.520 --> 0:04:25.440 new demand. This is the point at which I want 0:04:25.480 --> 0:04:29.880 to introduce you to a charming character named William Stanley 0:04:29.960 --> 0:04:33.640 Jevons is an English economist, a bit of a polymath really. 0:04:34.160 --> 0:04:36.560 He actually spent a few years working in Australia at 0:04:36.560 --> 0:04:39.000 the Sydney Mint before we went back to Britain to 0:04:39.000 --> 0:04:44.039 complete his education. Jevons never wrote about boots, but I 0:04:44.080 --> 0:04:47.200 reckon he probably wore them. I do know, though, that 0:04:47.279 --> 0:04:50.440 he cared a lot about coal, and the puzzle he 0:04:50.520 --> 0:04:55.200 noticed was this, when coal powered steam engines became more efficient, 0:04:56.000 --> 0:05:02.200 Britain didn't necessarily use less coal. Sometimes you more efficiency 0:05:02.320 --> 0:05:05.760 made coal powered activity cheaper, which made more of it 0:05:05.800 --> 0:05:10.039 worth doing. That idea use a resource like coal or 0:05:10.080 --> 0:05:13.520 our case, we're thinking about labor more efficiently, then the 0:05:13.520 --> 0:05:16.000 stuff it makes gets cheaper and you might end up 0:05:16.080 --> 0:05:19.280 using more coal or in our case, hiring more people. 0:05:19.760 --> 0:05:23.680 That set of ideas, that's called Jevons paradox. Okay, so 0:05:23.800 --> 0:05:27.239 let's apply that to boot making. More productive boot making 0:05:27.320 --> 0:05:30.320 meant cheaper boots. Cheaper boots meant more boots got bought, 0:05:30.360 --> 0:05:33.960 ready to wear footwear expanded, the market got bigger. There's 0:05:34.279 --> 0:05:36.640 a habit I want you to notice here. The question 0:05:37.360 --> 0:05:39.880 isn't just how many workers do you need per pair 0:05:39.880 --> 0:05:42.200 of boots? You also want to think about how many 0:05:42.279 --> 0:05:46.320 more pairs of boots will people now buy? That second 0:05:46.440 --> 0:05:51.240 question helps see things in starker relief. That's why economists 0:05:51.240 --> 0:05:55.520 talk about creative destruction. And the most important reminder here 0:05:55.560 --> 0:06:01.240 is there's two words in creative destruction creative and destruct Too. 0:06:01.320 --> 0:06:03.640 Many of us are misled because we only stare at 0:06:03.640 --> 0:06:06.040 the destruction. But the creative bit matters too, because new 0:06:06.040 --> 0:06:10.080 technologies they create new demand, new tasks, new firms, new 0:06:10.120 --> 0:06:12.719 forms of work, new ways are doing things, And in 0:06:12.760 --> 0:06:16.120 this story, that creative side turns out to matter a lot. 0:06:17.640 --> 0:06:21.800 So now for the astonishing headline. English boot making went 0:06:21.839 --> 0:06:25.920 through a full technological revolution, and yet the total number 0:06:25.920 --> 0:06:28.720 of jobs barely moved. It was about two hundred and 0:06:28.760 --> 0:06:31.559 twenty thousand workers in eighteen fifty one and two hundred 0:06:31.600 --> 0:06:35.800 and thirteen nineteen eleven, across sixty years of upheaval. Those 0:06:35.880 --> 0:06:40.679 numbers are so close it's like nothing changed except beneath 0:06:40.720 --> 0:06:46.920 that calm surface. Nearly everything changed. The work was fundamentally remade. 0:06:47.839 --> 0:06:49.920 You see the old craft world. It was built around 0:06:49.920 --> 0:06:55.040 occupations like cordwainers. I've always wanted to say that word binders, closers, 0:06:55.080 --> 0:07:00.400 and cloggers. As mechanizations spread, those occupations shranked dramatically, and 0:07:00.440 --> 0:07:02.960 in their place came a much more factory shaped world 0:07:02.960 --> 0:07:07.359 and occupations that we might recognize today machinists and riveters 0:07:07.600 --> 0:07:11.160 and operators. And along came a bunch more jobs to 0:07:11.240 --> 0:07:14.520 coordinate this new, more complex organization of work. Some of 0:07:14.560 --> 0:07:18.120 the obvious ones like managers and foremen and supervisors, but 0:07:18.200 --> 0:07:22.960 also accountants and clerks and wholesalers. So roughly the same 0:07:23.040 --> 0:07:26.559 number of folks were making birds, but they were doing 0:07:26.560 --> 0:07:29.960 completely different things. That are in new occupations, doing new tasks, 0:07:30.480 --> 0:07:33.679 a completely different organization of work, different division of labor, 0:07:33.760 --> 0:07:37.120 different kinds of film. Just about everything changed except the 0:07:37.200 --> 0:07:42.720 number of jobs. Here's how that happened. Roughly two thirds 0:07:43.200 --> 0:07:47.520 of the jobs in the old artisanal occupations they disappeared, 0:07:48.240 --> 0:07:50.840 and they were replaced by an almost identical number of 0:07:50.880 --> 0:07:55.880 new jobs in the new factory related occupations. This is 0:07:55.920 --> 0:07:59.240 the two sided nature of technological change. Jobs are destroyed 0:07:59.280 --> 0:08:03.360 and created. I think this helps explain some of the 0:08:03.360 --> 0:08:07.520 anxiety around AI right now. You see, when folks try 0:08:07.520 --> 0:08:10.440 to imagine AI, it's very easy to picture the destruction 0:08:10.600 --> 0:08:14.040 side of things. You can picture the occupations disappearing as 0:08:14.120 --> 0:08:17.160 large language models take over tasks that people used to do, 0:08:17.280 --> 0:08:21.040 like translating or copy editing or coding. It's easy to 0:08:21.080 --> 0:08:23.360 picture the first rung of the career lad of getting 0:08:23.440 --> 0:08:26.840 kicked away. What's much harder to picture of the occupations 0:08:26.840 --> 0:08:29.920 that don't yet exist. New tasks, new firms, new ways 0:08:29.920 --> 0:08:33.400 of organizing works. You can't picture them because, in a 0:08:33.480 --> 0:08:35.920 very real sense, nobody even knows what they're going to 0:08:35.960 --> 0:08:39.480 look like yet. Imagine going back to eighteen fifty one 0:08:40.200 --> 0:08:43.319 and asking a boot maker to imagine a prompt engineer. 0:08:44.520 --> 0:08:48.960 The result is anxiety. The anxiety is real and it's human, 0:08:49.920 --> 0:08:52.920 but history can help. It reminds us that a big 0:08:52.960 --> 0:08:55.720 technology shock does not always mean a collapse in the 0:08:55.760 --> 0:08:59.080 total number of jobs. Sometimes the big change is actually 0:08:59.080 --> 0:09:03.880 a deep, real organization of the work itself. That brings 0:09:03.960 --> 0:09:06.680 us to the next part of the story about geography. 0:09:07.280 --> 0:09:10.840 Old artisanal boot makers used to be spread pretty much 0:09:10.880 --> 0:09:14.840 everywhere across England. Every town needed a boot maker, so 0:09:15.520 --> 0:09:18.200 the end of that way of making boots cut jobs 0:09:18.240 --> 0:09:20.800 in nearly every part of the country. But when the 0:09:20.840 --> 0:09:24.040 new work using the new methods came forward, the jobs 0:09:24.040 --> 0:09:27.319 were not so evenly distributed. It's because the new jobs 0:09:27.320 --> 0:09:31.360 they were tied to factories and larger enterprises, and those 0:09:31.480 --> 0:09:34.280 larger boot making companies were clustered heavily in just a 0:09:34.280 --> 0:09:39.640 few places, above all, Northamptonshire and Leicestership. Cracky, I hope 0:09:39.640 --> 0:09:42.560 I said them right. Those two towns, however, you say, 0:09:42.600 --> 0:09:46.640 am together they captured nearly half of all the jobs 0:09:46.679 --> 0:09:51.760 in the newly created occupations. So yes, boot making survived 0:09:51.800 --> 0:09:54.960 in the aggregate, but it increasingly moved and lived at 0:09:54.960 --> 0:09:59.000 different addresses. One more thing. This all played out over 0:09:59.120 --> 0:10:03.440 many decades. Most of the incumbent artisans did not, in 0:10:03.480 --> 0:10:06.920 fact lose their jobs. Instead, the next generation just stopped 0:10:07.080 --> 0:10:11.640 entering the dying occupations. Instead they entered new ones. That 0:10:11.720 --> 0:10:16.200 means the old bootmakers got to retire as bootmakers. It's 0:10:16.240 --> 0:10:18.760 just their kids who never signed up. This was a 0:10:18.840 --> 0:10:22.439 revolution slow enough to let a whole generation off the hook. 0:10:23.400 --> 0:10:25.080 I'm not saying it was painless, but it could have 0:10:25.080 --> 0:10:27.719 been a whole lot worse, and that's important. As we 0:10:27.840 --> 0:10:31.120 now ask what does this tell us about AI? One 0:10:31.160 --> 0:10:33.480 answer is maybe less than some people would like. This 0:10:33.640 --> 0:10:37.760 history doesn't predict what AI will do. It doesn't give 0:10:37.840 --> 0:10:40.440 us a roadmap, but doesn't tell us everything will be fine, 0:10:40.559 --> 0:10:43.480 and it doesn't tell us that everything will fall apart either. 0:10:44.320 --> 0:10:47.760 What it does do is widen the range of outcomes 0:10:47.880 --> 0:10:50.960 you should take seriously. It's easy to picture the jobs 0:10:51.000 --> 0:10:53.040 that will be lost. It's harder to picture the ones 0:10:53.080 --> 0:10:57.640 that might yet emerge. The story of Victorian bookmakers puts 0:10:57.720 --> 0:11:01.720 a good news version of technological change into sharper focus. 0:11:02.640 --> 0:11:06.920 A major new technology arrived, a transformed production, and it 0:11:07.040 --> 0:11:11.360 did not show up as mass displacement of incumbent workers. 0:11:11.679 --> 0:11:15.600 Dario Moode, I want you to take note. What happened 0:11:15.640 --> 0:11:19.040 was that slow adjustment and persistent demand helped shield a 0:11:19.040 --> 0:11:22.720 lot of the incumbent workers while the industry transformed around them. 0:11:23.200 --> 0:11:26.320 New jobs were created, new tasks emerged. The creative side 0:11:26.320 --> 0:11:30.480 of creative destruction it also showed up, and the pace 0:11:30.520 --> 0:11:33.920 of change might well be central to why I have 0:11:33.960 --> 0:11:36.800 played out this way. I chatted with the author of 0:11:36.840 --> 0:11:39.560 this research, Hillary River Pond, and she told me she 0:11:39.640 --> 0:11:42.600 thinks the pace of change is a really big deal. 0:11:42.920 --> 0:11:46.680 In her case, Victorian boot making, the technological change was 0:11:46.720 --> 0:11:49.720 moderately slow. It was slow enough to grandfather older workers 0:11:50.080 --> 0:11:53.520 through the transition. And she raised the obvious concern if 0:11:53.559 --> 0:11:56.320 AI moves much faster and most people think it will, 0:11:57.160 --> 0:12:00.239 then that gentler adjustment may be a lot harder to achieve. 0:12:01.360 --> 0:12:04.720 But if the AI transformation slows down a bit, maybe 0:12:04.760 --> 0:12:07.679 we get more of the promised productivity gains without all 0:12:07.720 --> 0:12:11.480 of the promised unemployment. Maybe the creative side of creative 0:12:11.480 --> 0:12:15.000 destruction has time to arrive. Maybe new tasks and new 0:12:15.040 --> 0:12:18.440 firms can emerge quickly enough to absorb the shock, or 0:12:18.480 --> 0:12:22.360 maybe not. I don't know, And that modest is an 0:12:22.400 --> 0:12:27.520 important part of the story too. Economics is really pretty 0:12:27.559 --> 0:12:31.600 good at helping us think through mechanisms. It is much 0:12:31.640 --> 0:12:34.280 worse at giving any of us certainty about the future. 0:12:35.040 --> 0:12:37.720 So the value of this story isn't that it settles 0:12:37.960 --> 0:12:39.800 the AI debate. I want to be clear on that. 0:12:39.880 --> 0:12:42.760 It doesn't, but it teaches us how to think about 0:12:42.800 --> 0:12:46.319 it more clearly. It's tempting to immediately ask how many 0:12:46.360 --> 0:12:48.880 jobs might be destroyed when a new technology arise, but 0:12:48.880 --> 0:12:51.520 we should also ask what might be created and how 0:12:51.600 --> 0:12:55.480 quickly that transition will happen. Will the work show up 0:12:55.559 --> 0:12:57.560 the new work in the same place as inside the 0:12:57.559 --> 0:13:02.320 same firms for the same people. Those change too. One 0:13:02.320 --> 0:13:05.240 thing I know for sure, it's much easier to imagine 0:13:05.280 --> 0:13:09.160 the destruction than the creation. But both of the matter, 0:13:09.440 --> 0:13:11.720 and you should focus on both sides of that equation.