Latest / What's Your Problem? / What the Industrial Revolution Teaches Us About the AI Revolution

Transcript

What the Industrial Revolution Teaches Us About the AI Revolution · What's Your Problem?

0:15
Speaker 1: Pushkin.
0:20
Speaker 2: If you look back at the history of technology, there
0:22
Speaker 2: is this very very very long period of time, like
0:27
Speaker 2: thousands of years, where not much happened for humanity, at
0:32
Speaker 2: least not by modern standards. There were certainly some advances,
0:36
Speaker 2: there were periods of technological ferment here and there, Song
0:41
Speaker 2: Dynasty China as one famous example, But then those moments
0:44
Speaker 2: would pass, and things in terms of technology at least
0:48
Speaker 2: would go on largely as before people traveled. When they
0:52
Speaker 2: traveled at all, by foot or by animal or by
0:55
Speaker 2: sailing ship, and century after century, most of the people
0:59
Speaker 2: on earth worked as subsistence farmers, trying to grow enough
1:02
Speaker 2: food to survive. And then everything changed. It started in
1:08
Speaker 2: England about two hundred and fifty years ago. The steam engine,
1:12
Speaker 2: a modern factories all emerged in this moment, this moment
1:17
Speaker 2: that came to be called the Industrial Revolution, that was
1:21
Speaker 2: in many ways the beginning of the world we're living
1:23
Speaker 2: in today, This world where continuous technological breakthroughs, not just
1:29
Speaker 2: year after year, but generation after generation make workers more
1:34
Speaker 2: efficient and more productive. And this world where generation after
1:38
Speaker 2: generation the material conditions of humanity keep improving, the economic
1:43
Speaker 2: pie gets bigger. The changes that began with the industry
1:46
Speaker 2: of revolution are the reason that most people alive today
1:49
Speaker 2: are profoundly richer than their ancestors in the centuries since
1:53
Speaker 2: the industry of revolution. That link between new technologies and
1:57
Speaker 2: increasing broad based prosperity has been true for most people
2:02
Speaker 2: over the long run. But and this is a very
2:06
Speaker 2: important but this has not been true forever everybody. And
2:10
Speaker 2: sometimes that long run takes a very very long time.
2:15
Speaker 2: In other words, just because the pie gets bigger, it
2:18
Speaker 2: doesn't mean that everybody gets more pie. I'm Jacob Goldstein,
2:26
Speaker 2: and this is what's your problem. My guest today is
2:28
Speaker 2: Simon Johnson. He's an MIT economist and the co author
2:32
Speaker 2: of a new book called Power and Progress, Our thousand
2:36
Speaker 2: year Struggle over Technology and Prosperity. Simon's problem is this,
2:41
Speaker 2: how do you create the conditions for technological change to
2:44
Speaker 2: benefit everybody, or at least almost everybody, instead of only
2:49
Speaker 2: a powerful few.
2:52
Speaker 3: The basic argument is that it is wrong historically and
2:57
Speaker 3: also not sensible economics to assume that technological change always
3:03
Speaker 3: brings immediately and broadly shared prosperity. In fact, sometimes it
3:09
Speaker 3: does the opposite. Sometimes it helps just a few people,
3:13
Speaker 3: and sometimes it's really good for a lot of people.
3:15
Speaker 3: But the conditions under which you get a lot of
3:18
Speaker 3: benefits for a lot of people from technological change require
3:22
Speaker 3: some thought, and particularly thinking about today in America. Do
3:26
Speaker 3: we have the right conditions, for example, for an artificial
3:30
Speaker 3: intelligence revolution, if that's what we're facing, for that to
3:32
Speaker 3: really deliver lots of benefits lots of people, or is
3:35
Speaker 3: it going to be like some of those previous historical
3:37
Speaker 3: episodes where only a few people gain or hardly anybody gains,
3:40
Speaker 3: and a lot of people lose.
3:41
Speaker 2: Simon and I talked a lot about the development of
3:44
Speaker 2: AI and how to maximize the chances that AI delivers
3:47
Speaker 2: lots of benefits to lots of people. That's basically the
3:50
Speaker 2: second half of the interview. But I also really wanted
3:53
Speaker 2: to talk with him about the Industrial Revolution because I
3:56
Speaker 2: truly think it has deep lessons that are really useful
4:00
Speaker 2: in thinking about technology and the world today. And the
4:03
Speaker 2: Industrial Revolution, Simon said, really started with textiles, in particular,
4:09
Speaker 2: one kind of fabric.
4:11
Speaker 3: The real breakthrough, Jacob, is about cotton. So cotton is
4:14
Speaker 3: a material that the British did not invent. It's been
4:17
Speaker 3: in long use in many parts of the world. The
4:19
Speaker 3: Indian spinners were the leading spinners of cotton, and the
4:23
Speaker 3: export of spun cotton from India was a big deal.
4:27
Speaker 3: But the British figure that they can spin cotton and
4:31
Speaker 3: then subsequently weave cotton more efficiently than Indian artisans by
4:38
Speaker 3: applying machinery to this problem, and that became this cotton
4:42
Speaker 3: industry became the powerhouse of the English.
4:44
Speaker 2: So you have this the birth of technological innovation as
4:48
Speaker 2: we know it is happening right by the early eighteen hundreds.
4:51
Speaker 2: You're having these incredible productivity gains. This new technological era
4:56
Speaker 2: is just charging forward. Productivity is increasing. What's happening with
5:03
Speaker 2: wages for most workers in England at this point.
5:07
Speaker 3: In terms of what's happening in England's and what seemed
5:10
Speaker 3: to happen what seems to happen in terms of industrial
5:12
Speaker 3: wages as a whole, is they really don't progress. More
5:16
Speaker 3: so speaking quite broadly here anywhay near like what you
5:18
Speaker 3: would imagine given the pro tivy advances, and there is
5:21
Speaker 3: some of them, suggests broadly, not without its exceptions, but
5:25
Speaker 3: broadly there was wage stagnation in that early period, which
5:27
Speaker 3: is incredible and weird and pretty unsettling. Given how much
5:31
Speaker 3: productivity transformation, given how much technological change was underway, Well.
5:34
Speaker 1: That seems like.
5:37
Speaker 2: Sort of the key point from this period for the
5:41
Speaker 2: argument of your book to be a little.
5:43
Speaker 1: Reductive about it.
5:44
Speaker 2: And yes they're asterisks and caveats, but the big idea
5:49
Speaker 2: is you have this call it fifty this year period,
5:52
Speaker 2: the first half of the eighteen hundreds, where there's incredible
5:55
Speaker 2: productivity gains in England and overall workers don't seem to
6:00
Speaker 2: be capturing any of those gains, right, So what's going on?
6:03
Speaker 1: Why aren't workers' wages going up in this period?
6:07
Speaker 3: So the core issue Jacob in this moment, and it
6:12
Speaker 3: exists across all of technological transformation, but here we do
6:15
Speaker 3: see a nasty version here is that when you automate things,
6:18
Speaker 3: you obviously make some people more productive than the people
6:20
Speaker 3: who are run the machines, but you're also replacing other workers. Now,
6:24
Speaker 3: once we start to automate weaving, that's two hundred and
6:28
Speaker 3: fifty thousand weavers who are going to lose their jobs.
6:31
Speaker 3: About another fifty thousand people who are auxiliary service workers
6:34
Speaker 3: around that hand loom sector. So three to thousand people
6:37
Speaker 3: losing their jobs. Where are they going to work.
6:40
Speaker 2: Two stories that I feel like people have as kind
6:42
Speaker 2: of heuristics for this, right, as sort of toy models
6:44
Speaker 2: in their head. The sort of popular version is, oh,
6:48
Speaker 2: machines throw people out of work and we have technological unemployment. Right,
6:52
Speaker 2: everybody always says that, and yet here we are in
6:54
Speaker 2: twenty twenty three with an incredible amount of technology and
6:58
Speaker 2: unemployment at historic cloths right, So clearly that heuristic is
7:01
Speaker 2: not right. And then the other, the kind of economics
7:04
Speaker 2: heuristic is no, no, technological innovation makes things cheaper and
7:08
Speaker 2: so well, people can buy more stuff, and in buying
7:11
Speaker 2: more stuff goods and services, we create new jobs. But
7:14
Speaker 2: plainly that second one was not happening in the eighteen twenties.
7:18
Speaker 1: Why not.
7:19
Speaker 3: So there's a third possibility, actually, Jacob, which is people
7:22
Speaker 3: lose their jobs. They don't become unemployed because they can't
7:25
Speaker 3: afford to be unemployed. There's no uneployment assurance. They have
7:27
Speaker 3: to go to work at a very very low wage,
7:30
Speaker 3: and that wage may be you know, essentially at or
7:33
Speaker 3: slightly below subsistence. And I think that's also what we've
7:36
Speaker 3: experienced in the United States. To flash forward, which is
7:38
Speaker 3: since the digital transformation of the nineteen eighties, we've not
7:42
Speaker 3: had the technological unemployment that people were afraid of, but
7:45
Speaker 3: we have had a polarization of the job market. So
7:47
Speaker 3: some people have done well and a lot of people
7:50
Speaker 3: previous middle class, middle skill jobs have disappeared and people
7:53
Speaker 3: have got pushed down to low wage jobs. And that
7:55
Speaker 3: parallel I think we see in the early Industrial Revolution, Jacob,
7:59
Speaker 3: That's what I think absolutely happened to the weavers, for example.
8:03
Speaker 2: So, okay, that's the first half of the eighteen hundreds
8:06
Speaker 2: in England. In England, that changes in the second half. Right,
8:10
Speaker 2: in the latter part of the eighteen hundreds, we just
8:12
Speaker 2: do go up for ordinary people. These productivity gains that
8:15
Speaker 2: have been happening for decades and decades, they finally sort
8:18
Speaker 2: of pay off in a broader way. Right, Why what changes?
8:22
Speaker 2: Why does that happen?
8:23
Speaker 3: Then? So many of the frustrations of the first half
8:27
Speaker 3: of the nineteenth century, including the lack of representation for
8:30
Speaker 3: people living in these newly industrial areas that had bubbled
8:33
Speaker 3: up after the eighteen thirties, and there was much more
8:36
Speaker 3: awareness of anger and also very difficult living standards in
8:41
Speaker 3: these big industrial cities, and that leads to approcess of
8:44
Speaker 3: partial but ongoing democratization. So there's better rights for workers,
8:49
Speaker 3: including ultimately the right to form trade unions, and those
8:52
Speaker 3: trade unions are demanding higher wages by the eighteen eighties,
8:55
Speaker 3: and high wages in return to match the higher productivity.
8:58
Speaker 3: That's what starts to build much more broadly shared prosperity.
9:02
Speaker 2: So there's the political piece, trade unions being a notable
9:06
Speaker 2: piece of it that you have by the second half.
9:09
Speaker 2: You also in the book talk about a technological piece,
9:13
Speaker 2: how the nature of the technologies that are emerging and
9:17
Speaker 2: spreading more than emerging spreading by the second half of
9:19
Speaker 2: the eighteen hundreds, are also driving broader wage gains.
9:23
Speaker 1: Talk a little bit about that, right.
9:27
Speaker 3: So it's interesting that in the eighteen fifty one Great
9:30
Speaker 3: Exhibition that was held in London, there was almost no
9:33
Speaker 3: American technology on exhibit. The exceptions were some stuffed wild
9:37
Speaker 3: animals and some guns. Actually seems very American even today,
9:43
Speaker 3: but that was eighteen fifty one. By eighteen ninety, the
9:46
Speaker 3: US is the leading industrial power in the world. And
9:48
Speaker 3: that transformation was a lot about movement west. It was
9:51
Speaker 3: a lot about developing technology using agriculture. People were leaving
9:55
Speaker 3: farms in the Midwest to go to live in Chicago
9:58
Speaker 3: and work for McCormick Reaper's company to make equipment that
10:01
Speaker 3: would let more people leave the farms. And as these
10:05
Speaker 3: companies expanded, Singer Sewing Machine Company is another one, they
10:09
Speaker 3: look to European markets and they brought over their factories
10:12
Speaker 3: and their business models which were very or into making
10:15
Speaker 3: effective productive use of relatively unskilled labor. And when they
10:19
Speaker 3: brought that into Europe, that helped boost the demand for
10:22
Speaker 3: unskilled workers. And it was unskilled workers, make them highly productive,
10:26
Speaker 3: pay them a decent wage. That put the industrial development
10:31
Speaker 3: onto a different and we would argue much better try
10:34
Speaker 3: and they.
10:34
Speaker 1: Pay them a decent wage.
10:35
Speaker 3: Piece.
10:35
Speaker 1: I mean, it's that just a market dynamic.
10:38
Speaker 2: Because there are these new technologies like sewing machines, there
10:44
Speaker 2: is more demand for unskilled labor across the board, and
10:48
Speaker 2: therefore it's just a competitive equilibrium. Well, I got to
10:51
Speaker 2: hire somebody, and the guy at the factory nextdoor is
10:53
Speaker 2: going to hire this unskilled worker if I don't, so
10:55
Speaker 2: I have to offer them more money. Is it just that?
10:59
Speaker 1: Yeah?
10:59
Speaker 3: And I think it's it's part of that. It's the
11:01
Speaker 3: demand for labor. It's also the arrival of trade union.
11:04
Speaker 3: So the people I'm employing, I don't have a union
11:06
Speaker 3: in my factory, but two factories down they have a UNI,
11:09
Speaker 3: and my guys can and move down there and get
11:11
Speaker 3: the union wage if I don't pay them enough money.
11:13
Speaker 3: And I do think personally that railways were really important
11:16
Speaker 3: in this entire process, Jacob, because prior to railways it
11:19
Speaker 3: was very hard to move between places. It could take
11:22
Speaker 3: you two or three days on a stagecoach from London
11:24
Speaker 3: to Manchester, for example, very uncomfortable, quite expensive. Labor mobility
11:28
Speaker 3: was not very high in the sense of you know,
11:31
Speaker 3: are you going how far would you billing to move
11:32
Speaker 3: to get a better job. Once railways arrive, and once
11:35
Speaker 3: you have this unified market, there's a lot more possibility
11:38
Speaker 3: of moving to boom town's and a lot more understanding
11:41
Speaker 3: what's going on in you know, fifty miles away. And
11:44
Speaker 3: I think that was a very important part of critian
11:47
Speaker 3: national market and spreading ideas and boosting the demand for labor.
11:51
Speaker 2: I'm glad you mentioned railroads because in the book you
11:54
Speaker 2: write at some length about the rise of railroads, which
11:57
Speaker 2: is a little bit earlier than the period we're talking
12:00
Speaker 2: about here, right it's the first half of the eighteen hundreds,
12:03
Speaker 2: and in particular you write about this one guy named
12:06
Speaker 2: George Stephenson, who is sort of representative of this new
12:11
Speaker 2: kind of entrepreneur that's emerging around this time. Talk a
12:14
Speaker 2: little bit about George Stephenson.
12:16
Speaker 1: Who was he?
12:17
Speaker 3: So, George Stephens largely self educated. He was a mining engineer,
12:21
Speaker 3: but he didn't have any formal qualifications. He was just
12:24
Speaker 3: a chap who had worked in minds and helped solve
12:27
Speaker 3: problems with the design of the minds underground.
12:29
Speaker 2: Yourew not poor, right, He was not some like gentleman engineer.
12:33
Speaker 3: Absolutely, and in fact that's a common element to a
12:36
Speaker 3: lot of these people who become prominent innovators in the
12:39
Speaker 3: industrial ages. They come from quite modest backgrounds. Now there
12:42
Speaker 3: are gradations of poverty, of course, before the divosy. He's
12:45
Speaker 3: not in the poorest of the poor, but he's not
12:47
Speaker 3: middle class. He doesn't grow up in a nice house.
12:50
Speaker 3: He doesn't He can't read and write actually probably until
12:53
Speaker 3: he's an adult. So he figures stuff out by himself.
12:56
Speaker 3: And he's a tinkerer with machines, and his big breakthrough
12:59
Speaker 3: came when he went up to the coal mine one
13:01
Speaker 3: day and they were having trouble with a new fangled
13:03
Speaker 3: seam engine was pumping water out of the mine and
13:06
Speaker 3: he said, you know what, you know, give me a
13:07
Speaker 3: couple of days and some of my mates and we
13:08
Speaker 3: can fix this. And he did. And the people who
13:11
Speaker 3: owned the coal mine were you know, they were rich people.
13:14
Speaker 3: They were quite smart for people also, and they say, right,
13:18
Speaker 3: this Stevenson Chap is obviously you know a talent that
13:20
Speaker 3: we should and we should you know, back him and
13:22
Speaker 3: help him solve other problems.
13:25
Speaker 2: So how does Stevenson get from there to becoming a
13:29
Speaker 2: sort of great entrepreneur slash inventor engineer, Well.
13:35
Speaker 3: By solving problems really And then the main main problem
13:38
Speaker 3: he solved was how to move the coal. So there
13:40
Speaker 3: was a sort of a mechanical issue what's the best
13:42
Speaker 3: way to transport coal along the railways? But there was
13:45
Speaker 3: also an organization issue. So initially they ran their rails.
13:49
Speaker 3: This is in the northeast of England, near Newcastle. They
13:52
Speaker 3: ran them like we run roads today, which is, you know,
13:54
Speaker 3: somebody's in charge of the road, but anybody who's got
13:56
Speaker 3: a licensed car can put it on the road and
13:58
Speaker 3: drive somewhere and so there was an enormous amount of
14:01
Speaker 3: confusion and many hilarious stories, with some of them with
14:04
Speaker 3: quite tragic endings about people not giving way to each
14:07
Speaker 3: other on these limited railways. Stevenson had this vision, if
14:11
Speaker 3: you like, that there was a better way to do this,
14:13
Speaker 3: and that was to have one company owned the rail
14:16
Speaker 3: run the trains. People could provide freight, and of course
14:19
Speaker 3: passengers could step up or not to ride the trains,
14:22
Speaker 3: but you'd have an integrated control over this railway system.
14:25
Speaker 3: And that's what he persuaded some people to adopt for
14:28
Speaker 3: the Liverpool and Manchester Railway, and in what I think
14:32
Speaker 3: is one of the most amazing moments of human history
14:37
Speaker 3: of nginuity, certainly, he ran a combined Nobel Prize slash
14:41
Speaker 3: bakeoff show to determine who had the best railway engine
14:45
Speaker 3: to run on the rails that he designed and that
14:48
Speaker 3: he had planned. And he was also the winner of
14:51
Speaker 3: that competition. So there's some interesting conflicts of interest at
14:55
Speaker 3: the beginning of the Industrial Revolution, but his stuff worked.
14:57
Speaker 3: That was a huge event and it brought massive amount
15:00
Speaker 3: of attention to this industry and kicked off what became
15:03
Speaker 3: known as a railway mania, which was the building of
15:05
Speaker 3: railways and the development of the railway system across first
15:08
Speaker 3: England and then Europe and.
15:09
Speaker 1: Then the world.
15:11
Speaker 2: Mostly on this show I interview founders of tech companies
15:14
Speaker 2: more or less, and when I was reading about Stevenson,
15:17
Speaker 2: I was like, oh, this guy's like a like a
15:19
Speaker 2: tech founder, He's.
15:20
Speaker 1: Like people I interview on the show. I mean, do
15:22
Speaker 1: you think that's fair to some extent? Not fair? Like,
15:26
Speaker 1: am I just dreaming?
15:28
Speaker 3: I think no. I think this parallel. So here's a
15:30
Speaker 3: man with vision. He makes mistakes, he learns the hard way.
15:33
Speaker 3: A lot of these engines blew up, by the way,
15:35
Speaker 3: several of his relatives died in the in the in
15:37
Speaker 3: the shop where they were making engines because it was
15:40
Speaker 3: extremely dangerous business. So he figures stuff out, he learns
15:44
Speaker 3: by doing. He fails fast. I mean these are all
15:47
Speaker 3: buzzwords of today, of course I think there are. And
15:50
Speaker 3: he gets he gets people with money to back him
15:52
Speaker 3: and proves results and then gets more money and so on.
15:54
Speaker 3: I think the social gap between Stevenson and the elite
15:58
Speaker 3: was was enormous, and there was a famous hearing to
16:02
Speaker 3: review whether or not they could build this Liverpool and
16:04
Speaker 3: Manchester railway in which a top barrister who went on
16:08
Speaker 3: to become a prominent lawyer in the UK working for
16:12
Speaker 3: the government. This rarester just ripped into pieces and Stevenson
16:14
Speaker 3: was in articulate and he couldn't explain himself. And I
16:17
Speaker 3: think that social gap, that's the size of it. We
16:21
Speaker 3: don't see that today. I think most of the entrepreneurs
16:22
Speaker 3: we come across are well educated, they've gone to college,
16:25
Speaker 3: they're articulate people, and I don't see people rising up
16:30
Speaker 3: from essentially nowhere like Stevenson did to the same extent
16:35
Speaker 3: today as in the early industrial revolutionship.
16:37
Speaker 2: So, oh, okay, so we've done this story. It's like
16:40
Speaker 2: kind of a little more than one hundred years, right,
16:43
Speaker 2: starting in the seventeen hundreds, going up to the late
16:45
Speaker 2: eighteen hundreds.
16:46
Speaker 1: What are the.
16:47
Speaker 2: Lessons of this story about technological change, political power, and
16:53
Speaker 2: how economic gains are shared or not shared.
16:57
Speaker 3: So I think the main lesson is there's nothing automatic
16:59
Speaker 3: that links technological change, improvements in technology with better outcomes
17:04
Speaker 3: for workers. Certainly that linkage can develop, and it did
17:09
Speaker 3: up in the later nineteenth century. And you know, it
17:12
Speaker 3: was great that that happened, but that was a struggle
17:16
Speaker 3: that had to be one as opposed to something that
17:18
Speaker 3: happened just because technology, just because technology was changing. So
17:21
Speaker 3: I do think the political dimension of the second half
17:24
Speaker 3: of the nineteenth century, the spreader of democratization in Europe,
17:28
Speaker 3: the US becoming more democratic for example, in the progressive era,
17:31
Speaker 3: those were really important elements linking a technological change to
17:34
Speaker 3: better outcomes from people. And if that linkage that political
17:38
Speaker 3: language becomes weaker, which it has in the past forty
17:41
Speaker 3: fifty years, then you should be more concerned about what
17:45
Speaker 3: do you only get from potential future technological transformations.
17:52
Speaker 2: In a minute, Simon and I will talk about what's
17:54
Speaker 2: happening now, what may be coming, and what we can
17:58
Speaker 2: do to maximize the chances that AI will benefit lots
18:01
Speaker 2: of people instead of only a few.
18:15
Speaker 1: Now back to the show.
18:16
Speaker 2: By the end of the eighteen hundreds and into the
18:20
Speaker 2: twentieth century, at least in parts of the world, certainly
18:23
Speaker 2: in the United States and in England and in Western Europe,
18:26
Speaker 2: you do have this period when technological innovation, productivity gains
18:33
Speaker 2: are going along with broad based prosperity gains. Not everywhere,
18:38
Speaker 2: not all the time, but certainly, you know famously in
18:41
Speaker 2: the middle of the twentieth century in the US and
18:44
Speaker 2: in Europe at least, you have this era that is
18:48
Speaker 2: sort of the dream of like things are getting more
18:50
Speaker 2: efficient and workers are getting richer, and it's the version
18:54
Speaker 2: of technological progress that we like because it seems to
18:57
Speaker 2: be good for everybody, or at least very large groups
19:00
Speaker 2: of people. It is broadly shared prosperity. When, in your view,
19:04
Speaker 2: does that break down that link between technological progress and
19:07
Speaker 2: broadly shared you know, prosperity.
19:11
Speaker 3: Well, in retrospect, it came under a lot of pressure
19:14
Speaker 3: in the nineteen seventies, and I think it began to
19:16
Speaker 3: break down really seriously in the nineteen eightiest. Like a
19:20
Speaker 3: lot of these things, I don't think it became clear
19:22
Speaker 3: to people that there was a problem until the nineteen nineties,
19:24
Speaker 3: but by the nineteen nineties there were definitely there was
19:26
Speaker 3: a lot of analysis that said, you know, why approtaivity
19:30
Speaker 3: gain is not becoming higher wages like they used to,
19:34
Speaker 3: and what exactly is broken about this the form of innovation?
19:39
Speaker 3: And I think that's a problem we've grappled with for
19:41
Speaker 3: twenty five years and not yet found.
19:42
Speaker 1: A solution to.
19:44
Speaker 2: And you know, it feels like The answer to that
19:49
Speaker 2: is not entirely clear, but I know you make an
19:53
Speaker 2: argument about it in the book, like why do you
19:54
Speaker 2: think it happened?
19:57
Speaker 3: Well, a combination of factors where automation, the form of
20:01
Speaker 3: automation has become really is really important. Automation is being
20:07
Speaker 3: used primarily to replace people. At the same time, we're
20:11
Speaker 3: not generating a lot of new jobs, new opportunities in
20:14
Speaker 3: the way that we did in the early twentieth century,
20:16
Speaker 3: for example. So we've continued to automate, people have lost
20:19
Speaker 3: those jobs. We've not created a lot of new tasks.
20:21
Speaker 3: And that's partly about added UDEs of them and corporate leadership.
20:25
Speaker 3: It's partly about the way digital technology has been developed
20:28
Speaker 3: and deployed. It's also, unfortunately about globalization and the interaction
20:31
Speaker 3: between automation and how we've outsourced work to lower income
20:35
Speaker 3: country well.
20:36
Speaker 2: So globalization is an interesting piece, right because over that period,
20:41
Speaker 2: global inequality has fallen over the last several decades, right
20:45
Speaker 2: to some extent, globalization has led to great wage gains
20:50
Speaker 2: for people who who live in parts of the world
20:53
Speaker 2: that were formerly very, very very poor.
20:55
Speaker 1: You know, namely China.
20:57
Speaker 2: The media in person in China is now much richer
21:00
Speaker 2: than they were thirty years ago.
21:01
Speaker 1: Like that seems good.
21:04
Speaker 3: Yes, right, So the primary change is about China. There
21:07
Speaker 3: are many poor parts of the world that participated in this,
21:11
Speaker 3: And of course what's also happened in China is productivity
21:13
Speaker 3: gains of outstrip wage gains, so that that wedge has
21:17
Speaker 3: benefited some of the people in the Chinese system, not everybody,
21:20
Speaker 3: but I completely agree that wages have gone up and
21:23
Speaker 3: poverty has declined in China as part of this. So, yes,
21:27
Speaker 3: the Chinese and other countries have benefited from the global
21:31
Speaker 3: trading system, and that's good. But I think there are
21:35
Speaker 3: better ways to arrange that system, and better ways that
21:37
Speaker 3: would have more inclusion for more people, including the US.
21:40
Speaker 1: Let's talk more about the US.
21:42
Speaker 2: Well, what's your worry if things don't change, what do
21:47
Speaker 2: you think the US will be like in five years
21:49
Speaker 2: or ten years? Like, what's the sort of Yeah, what's
21:52
Speaker 2: your prediction if the status quo persists?
21:56
Speaker 3: So Vonnegut wrote, I think it's his first novel, play
21:59
Speaker 3: a Piano, which he, as I understand it, really wrote
22:03
Speaker 3: and developed in the late nineteen forties. Well, and he
22:06
Speaker 3: imagined a society in which there are a few eats
22:09
Speaker 3: people who have high status who run the machinery, and
22:11
Speaker 3: a lot of people who have make work projects at
22:14
Speaker 3: relatively low wages, low stand living, no prospects, And I
22:17
Speaker 3: think that that inequality of status, inequality of opportunity, inequality.
22:22
Speaker 3: The people on the public works projects, by the way,
22:25
Speaker 3: are not starving, they are provided for, but that they
22:27
Speaker 3: don't have any future. That stark and rather rigid system,
22:36
Speaker 3: I think is a real possibility in this country. And
22:39
Speaker 3: I think we already have some elements of that, and
22:41
Speaker 3: I think it's problematic and maybe even least to worse
22:46
Speaker 3: social outcomes than Vonnegut you imagined.
22:47
Speaker 1: And so how do we change that?
22:51
Speaker 3: So our argument is that artificial intelligence could be beneficial
22:54
Speaker 3: to more people, that it could be empowering, that we
22:57
Speaker 3: could look to waste emphasise more ways to enhance human capabilities.
23:02
Speaker 3: This is not our understanding of the current technological priority
23:07
Speaker 3: from the people who have the dominant visions, people running
23:09
Speaker 3: Google and Microsoft, for example, but we absolutely talk to
23:13
Speaker 3: them and urge them to move in this direction. We'd
23:15
Speaker 3: like to have more competition for ideas in that market
23:19
Speaker 3: for the same reason, and look for ways to use
23:22
Speaker 3: technology to help humans as opposed to continuing down the
23:26
Speaker 3: road where we've become a little bit too obsessed with
23:29
Speaker 3: how many workers can I fire this quarter by hiring
23:33
Speaker 3: new machines on new algorithms. That I think is a
23:36
Speaker 3: bad dynamic for society.
23:38
Speaker 2: I mean, unemployment is that is still below four percent.
23:42
Speaker 2: Wages have been going up. They've been going up more
23:44
Speaker 2: rapidly in the past few years, for workers at the
23:47
Speaker 2: lower end of the income distribution, right, I mean, the
23:52
Speaker 2: facts don't seem exactly like you're describing at this moment.
23:56
Speaker 3: Well, we definitely had a bump up from COVID. Now,
23:58
Speaker 3: COVID was not.
23:59
Speaker 1: Something before COVID.
24:01
Speaker 2: Right by the late twenty teens, we were essentially at
24:04
Speaker 2: full employment. Wages were rising across the income distribution. I mean,
24:08
Speaker 2: it's clearly true that it took too long to get
24:10
Speaker 2: to full employment after the financial crisis, but whatever, that's
24:13
Speaker 2: like a fiscal policy story that we don't need to
24:15
Speaker 2: get into here. In terms of technology and labor demand,
24:18
Speaker 2: it feels like there is robust labor demand now, even
24:21
Speaker 2: for relatively unskilled workers.
24:24
Speaker 3: Labor demand is stronger now than it has been in
24:27
Speaker 3: some recent periods. I absolutely agree. But if you look
24:29
Speaker 3: at that the divergence in incomes of the higher earning
24:32
Speaker 3: and the lower earning. Over the past forty years, twenty
24:36
Speaker 3: percent of that gap has been closed, perhaps a little
24:38
Speaker 3: bit more with the COVID bump, but it's not clear
24:41
Speaker 3: and that look, if the problem solved. Jacob Gray happy
24:44
Speaker 3: to have been a little too worried. But I think
24:47
Speaker 3: the concern is that the underlying dynamics of technological adoption
24:51
Speaker 3: and what we're doing with technology hasn't changed much. But
24:54
Speaker 3: our view is that the dynamic of deployment of AI
24:57
Speaker 3: and what companies are talking about using AI to do
24:59
Speaker 3: replace customer service, replace workers in various customer facing roles,
25:05
Speaker 3: replace workers in back office, that is going to can
25:09
Speaker 3: continue that previous divergence of real incomes.
25:14
Speaker 2: So, I mean, you sort of alluded to a few
25:17
Speaker 2: possible responses before, but maybe you could pick a few
25:20
Speaker 2: to talk about in a little more depth. So there
25:22
Speaker 2: are these potential bad outcomes from AI for sort of
25:26
Speaker 2: workers in the middle of the distribution or for lesser
25:29
Speaker 2: skilled workers. What do we do to help those people?
25:32
Speaker 2: How do we reduce those risks? Like what are a
25:35
Speaker 2: few specific things?
25:37
Speaker 3: So the thing that people talk about all the time
25:39
Speaker 3: and the way you framed it that, Jacob, I think
25:43
Speaker 3: pools that direction. You say, well, people have less skill,
25:45
Speaker 3: let's give them more skill, right, so more education. We're
25:48
Speaker 3: not opposed to that, but we also are trying in
25:50
Speaker 3: this book to talk about the direction of technological change.
25:53
Speaker 3: Who has the visions, who invents things? So this says
25:56
Speaker 3: back to your sort of your your interest in George Stephenson.
25:58
Speaker 3: Who was George Stevenson? Where did he come from? How
26:02
Speaker 3: did he get this opportunity? Right? And I think what
26:05
Speaker 3: we'd like to encourage is more use of all new
26:08
Speaker 3: technolog including AI, to bring out more George Stephenson's to
26:13
Speaker 3: create more new tasks, to try and become more innovative
26:17
Speaker 3: with this technology in a way that creates opportunity and job.
26:20
Speaker 2: Nobody's going to be opposed to that, Like sure, everybody,
26:23
Speaker 2: When you say that, I'll say sure, But how do
26:25
Speaker 2: you do that? Like, what is a concrete thing in
26:27
Speaker 2: the world that could cause that to happen?
26:31
Speaker 3: Well, for example, almost all of the research and development
26:34
Speaker 3: around AI right now, it takes place in a few
26:36
Speaker 3: big companies. There's a very little takes place in any
26:39
Speaker 3: government institute or actually the universities are losing a lot
26:42
Speaker 3: of talent into those companies. What is the priority of
26:45
Speaker 3: those companies It is to make money. What is an
26:48
Speaker 3: alternative set of priorities is to generate new tasks, new opportunities,
26:52
Speaker 3: breakthroughs in technology that we haven't yet.
26:54
Speaker 1: Money have to happen.
26:54
Speaker 2: For a more public spirited development of AI to happen,
26:58
Speaker 2: is it public funding for AI development?
27:02
Speaker 3: Public funding would be important and it has been important
27:04
Speaker 3: in many previous technologies, including development of computer chips and
27:09
Speaker 3: the Internet and modern pharmaceuticals. Often the government does very
27:13
Speaker 3: well when it provides a market of It says we're
27:15
Speaker 3: going to buy things from you if you develop them right.
27:18
Speaker 3: That's what we do with COVID vaccines, for example. So
27:20
Speaker 3: there are various ways that you can put public resources
27:23
Speaker 3: to work. And in the current conversation in Washington, the
27:26
Speaker 3: good news is those possibilities are not off the table.
27:29
Speaker 3: But I go to some of these conversations take and
27:32
Speaker 3: I wouldn't say it's the top item. It probably struggles
27:35
Speaker 3: to get into the top five. Mostly it's large private
27:38
Speaker 3: sector companies saying, hey, don't get in our way, don't
27:40
Speaker 3: put a lot of rules on us. If you do,
27:42
Speaker 3: then China will take over the world. And that is
27:45
Speaker 3: unfortunately the thrust of that technological conversation right now, whereas
27:50
Speaker 3: we would suggest you start much more. Okay, what are
27:52
Speaker 3: you trying to invent here? What do we not have?
27:54
Speaker 3: What is the private sector not going to come up with?
27:56
Speaker 3: And then what's the system of carents? Various kinds of
27:59
Speaker 3: carrots probably no stakes, various kinds of carrots that's going
28:01
Speaker 3: to pull innovative people into this better direction where I agree,
28:07
Speaker 3: you know, it may be a little cloudy exactly what
28:09
Speaker 3: better means, but generally speaking, creating new tasks for people
28:13
Speaker 3: and creating new tasks that the people were previously not
28:15
Speaker 3: highly skilled can do productively, that's a smart way forward.
28:19
Speaker 2: What is an example of the thing you were just
28:21
Speaker 2: saying of a carrot from the government that would you know,
28:25
Speaker 2: induce people, provide an incentive for people to come up
28:28
Speaker 2: with a direction for AI that would benefit you know,
28:32
Speaker 2: broadly benefit workers, right, this is what we're looking for.
28:34
Speaker 1: What's an example of that?
28:36
Speaker 3: Make workers in electrical power grid companies, electrical disputors, make
28:42
Speaker 3: them more make them safeer, make them more productive, to
28:44
Speaker 3: help healthcare workers become more productive and better able to
28:50
Speaker 3: diagnose problems and so on, So a smarter expert advisory
28:54
Speaker 3: system that a nurse practitioner can have. And the really
28:56
Speaker 3: big one, Jacob, which is this one I think is
28:58
Speaker 3: probably worth a bounty, is to actually really finally use
29:01
Speaker 3: AI and education to help again kids with less advantage
29:09
Speaker 3: and less family resources. And these things have been imagined,
29:12
Speaker 3: but they've not been ever developed at scale so that
29:15
Speaker 3: teachers can actually help students learn better in more tailored
29:20
Speaker 3: fashion using AI type tools, and then get those systems
29:25
Speaker 3: deployed and adopted across public education. None of these are,
29:30
Speaker 3: of course, are easy problems. That's why they're problems. But education,
29:34
Speaker 3: healthcare workers in critical sectors, including around electricity, clean energy,
29:40
Speaker 3: those would seem to be sensible priorities.
29:42
Speaker 2: What else are there things you want to talk about
29:44
Speaker 2: that we didn't talk about?
29:45
Speaker 3: What's very interesting, Jacob, is you dug deeper into the
29:48
Speaker 3: Industrial Revolution pieces of it than many people have, including
29:51
Speaker 3: HA been on lots of good podcasts. But I think
29:54
Speaker 3: that you're absolutely right that it's sort of to understand
29:57
Speaker 3: the sequence of history, to understand these the episodes, and
29:59
Speaker 3: then to think, okay, we're obviously living history people going
30:03
Speaker 3: to look back and say, Okay, this was an episode,
30:05
Speaker 3: but what is this episode? Is this like a continuation
30:07
Speaker 3: of the nineteen eighties? Is it something brand you? I
30:10
Speaker 3: think that we often exaggerate the moment in which we
30:13
Speaker 3: live because that's the moment in which we're living. Yeah, right,
30:15
Speaker 3: So I always tend to think but but sometimes the
30:19
Speaker 3: crisis of two thousand and eight was different. For example,
30:22
Speaker 3: nine to eleven was different. There are these departures, there
30:25
Speaker 3: are these forks.
30:26
Speaker 2: The question right now I feel like, is are we
30:30
Speaker 2: just sort of persisting in the same universe we've been in,
30:33
Speaker 2: or is what's happening with AI creating.
30:39
Speaker 1: An inflection point? To use an overused term.
30:42
Speaker 2: Right like, is something really new happening with AI right now?
30:46
Speaker 1: Or are is it going to be more of the same?
30:49
Speaker 3: So I don't know how big is AI going to be? Unclear?
30:54
Speaker 3: But have we is this a process that is likely
30:58
Speaker 3: to develop deliver big shared benefits? I think a bit
31:04
Speaker 3: worried about that.
31:04
Speaker 2: You're a bit worried that all the benefits are going
31:06
Speaker 2: to go to a small number of people, and some
31:08
Speaker 2: large number of people will be not helped and potentially harmed.
31:14
Speaker 3: Right, And you know we have seen episodes in history
31:17
Speaker 3: where the harm can be substantial, it can be widespread,
31:19
Speaker 3: so let's not kid ourselves about that.
31:23
Speaker 2: We'll be back in a minute with the lightning round.
31:27
Speaker 2: M back to the show before we go, I want
31:39
Speaker 2: to do a lightning round. I want to ask you
31:42
Speaker 2: some fun, somewhat silly questions.
31:46
Speaker 3: I use a lightning round in my classes, Jacob. I
31:47
Speaker 3: don't forget if I got it from Planetmoni or someone else,
31:49
Speaker 3: but I think the lightning round is a great thing.
31:51
Speaker 3: Just fire fire away, fire away.
31:53
Speaker 1: Okay.
31:55
Speaker 2: You are a tennis player, a recreational tennis player. As
31:58
Speaker 2: a tennis player, what is your view of pickleball?
32:03
Speaker 3: That is a really sensitive question among tennis players. I
32:06
Speaker 3: enjoy pickaball, the whole family can play its barriers to entry.
32:10
Speaker 3: But I understand that some of my tennis playing friends
32:12
Speaker 3: may not be tennis may not want to be my
32:14
Speaker 3: friends anymore.
32:15
Speaker 2: After I said that, on a slightly less important topic,
32:18
Speaker 2: you used to be the chief economist at the IMF,
32:21
Speaker 2: the International Monetary Fund.
32:23
Speaker 1: It's the IMF overrated or underrated?
32:28
Speaker 3: Oh, this sounds like the pickaball question in a different guys.
32:31
Speaker 2: Equally slightly less controversial than the pickleball question.
32:36
Speaker 3: The IMF is a very important organization that exists. Almost
32:40
Speaker 3: the entire life of the IMF exists just to the
32:43
Speaker 3: side of the front page of the New York Times,
32:44
Speaker 3: by which I mean you very rarely read anything about
32:48
Speaker 3: all the important things that happened at the IMF, because
32:50
Speaker 3: you know, it's just a little too boring. As far
32:52
Speaker 3: as the New York Times editor is a concern. And
32:54
Speaker 3: that's good because they can get a lot of things done.
32:56
Speaker 3: They can be constructive, and it's enabled by that particular
33:00
Speaker 3: by the way in which they were positioned relative to
33:02
Speaker 3: the US political system. But they do exist. That they
33:05
Speaker 3: exist just slightly. You know how there's something in your
33:08
Speaker 3: perple vision you can't ever quite focus on. That's where
33:12
Speaker 3: they are with pickball.
33:15
Speaker 2: Pickleball's right in the middle, man. Pickleball is front and center.
33:23
Speaker 2: What's your favorite novel?
33:27
Speaker 3: Oh, snow Crash, Neil Stevenson. That's very easy. I reread
33:31
Speaker 3: that every every year or so. So this is a
33:33
Speaker 3: book written came out in the early nineteen nineties, and
33:37
Speaker 3: I originally read it because Paul Kruman said, an economist,
33:39
Speaker 3: if you want to understand the future don't read any
33:41
Speaker 3: futurologists read Neil Stevenson, and in terms of globalization, in
33:45
Speaker 3: terms of technology, in terms of social impact. It's including
33:49
Speaker 3: the metaverse. It's remarkable.
33:51
Speaker 1: And Crypto, you know, it's funny.
33:55
Speaker 3: I've thought it was a different book. Actually that was
33:57
Speaker 3: a different series. Crypto was less just as a new author, right,
34:00
Speaker 3: same author. Oh yeah, yeah, just as a Neil Stevenson
34:02
Speaker 3: groupie to be clear that that was the Crypto, the
34:04
Speaker 3: Crypto series, the Money series.
34:05
Speaker 2: This is the second time in this interview I've thought
34:07
Speaker 2: of sal Khan. I didn't mention it last, but I
34:09
Speaker 2: just interviewed Sel Con of the Con Academy, which you
34:11
Speaker 2: might know of, and I thought of him when you
34:13
Speaker 2: were talking about education, because they are developing AI an
34:18
Speaker 2: AI tutor basically that at the moment is not free,
34:20
Speaker 2: but we'll be free, I suspect quite soon. Also, he
34:24
Speaker 2: loves Neil Stevenson. There's some Neil Stevenson book set. I
34:27
Speaker 2: believe in China where there's like a tablet that has
34:30
Speaker 2: education and like the masses get the tablet education. We've
34:34
Speaker 2: been trying to book Neil Stevenson, but he doesn't do interviews.
34:36
Speaker 2: So if you ever see.
34:37
Speaker 1: Him, tell him to call.
34:41
Speaker 3: We'll do.
34:47
Speaker 2: Simon Johnson is a professor at MIT. His new book
34:51
Speaker 2: is Power and Progress. Today's show was produced by Edith Russello.
34:55
Speaker 2: It was edited by Sarah Nix and engineered by Amanda K.
34:59
Speaker 1: Wong.
35:00
Speaker 2: You can email us at problem at Pushkin dot fm.
35:04
Speaker 2: You can find me on Twitter at Jacob Goldstein. I'm
35:07
Speaker 2: Jacob Goldstein and we'll be back next week with another
35:09
Speaker 2: episode of What's Your Problem.
35:18
Speaker 1: M