The automation charade(logicmag.io)
logicmag.io
The automation charade
https://logicmag.io/05-the-automation-charade/
5 comments
IP is the new "means of production". Nobody wants to sell you a tractor outright anymore. They want you to buy the expensive to maintain hardware while they own the IP that makes it work. This ensures they can keep getting a recurring slice of the revenue you create with that machine - without the maintenance and other costs associated with a rental business.
Kitchen robots still haven't happened much. Except for computer-controlled woks.[1] Those are big in China.
This seemed closer in 1960.[2]
[1] https://www.youtube.com/watch?v=jJUMAWL3LXg
[2] https://www.youtube.com/watch?v=1Xop9py8zBY
This seemed closer in 1960.[2]
[1] https://www.youtube.com/watch?v=jJUMAWL3LXg
[2] https://www.youtube.com/watch?v=1Xop9py8zBY
Conveyor ovens toast a lot of bread, cook a lot of burgers, and bake a lot of pizza. It's not sexy, but it's automation and it reduces the labor needed.
Apparently the bots can cook burgers better: https://www.forbes.com/sites/christinatroitino/2018/06/21/me...
This article reads a lot like one of the best frustrating books I have read: David F. Noble's Forces of Production: A Social History of Industrial Automation. Noble's book is about the introduction of computer numerically controlled machine tools in American manufacturing after WW II, covering the technology but focusing on CNC's effect on the balance of power between skilled workers and management. The technology was very expensive and fragile when first introduced, and was more effective at cowing workers than at quickly making high precision parts. American makers of CNC machines started with defense industry applications and couldn't manage to make machines competitively priced for civilian use even after they had stopped making wholly defective ones. The chronology ends in the late 1970s; the book was published in 1984.
Management forces crazy process changes into a business, and is forced to backtrack after wasting years and huge sums of money -- a familiar story, right? Except that's not how it ended. There was no backtracking. The optimistic early claims about CNC became, after much time and money, realistic claims. German and Japanese machine tool manufacturers eventually produced the priced-for-civilian-work machines that American tool manufacturers couldn't or wouldn't. Nowadays, no American manufacturer would pay high wages for someone to precisely guide the motions of a lathe by coordinated hand-eye movements. They wouldn't even pay low wages for that particular task; human performance is too inconsistent.
What of this vignette from the article, that also takes place between the immediate postwar and the 1970s?
In 1946, some 550,000 auto workers had produced a little more than three million vehicles, but in 1970 some 750,000 auto workers had produced a little more than eight million vehicles.” The rapid pace took a devastating toll of people’s health and lives, leaving millions of workers disabled or dead; one study from the period estimated “sixty-five on-the-job deaths per day among autoworkers.” Auto industry executives credited the industry’s productivity boom to advances in machinery, but the predominantly black workforce knew it was in fact due to old-fashioned exploitation, not automation: heavier workloads and unsafe, unhealthy conditions.
Today, General Motors employs 180,000 people and sells 9.6 million vehicles per year. Simultaneously, injury rates in manufacturing are down a lot since the 1970s.
I think that both Noble's book and this article would be more persuasive if they committed to one major rhetorical thrust. Either "management is bluffing about how labor can be substituted by machines, so workers shouldn't be intimidated" or "machines can substitute for labor, so workers shouldn't expect the market economy to always provide paid work", but not both.
Management forces crazy process changes into a business, and is forced to backtrack after wasting years and huge sums of money -- a familiar story, right? Except that's not how it ended. There was no backtracking. The optimistic early claims about CNC became, after much time and money, realistic claims. German and Japanese machine tool manufacturers eventually produced the priced-for-civilian-work machines that American tool manufacturers couldn't or wouldn't. Nowadays, no American manufacturer would pay high wages for someone to precisely guide the motions of a lathe by coordinated hand-eye movements. They wouldn't even pay low wages for that particular task; human performance is too inconsistent.
What of this vignette from the article, that also takes place between the immediate postwar and the 1970s?
In 1946, some 550,000 auto workers had produced a little more than three million vehicles, but in 1970 some 750,000 auto workers had produced a little more than eight million vehicles.” The rapid pace took a devastating toll of people’s health and lives, leaving millions of workers disabled or dead; one study from the period estimated “sixty-five on-the-job deaths per day among autoworkers.” Auto industry executives credited the industry’s productivity boom to advances in machinery, but the predominantly black workforce knew it was in fact due to old-fashioned exploitation, not automation: heavier workloads and unsafe, unhealthy conditions.
Today, General Motors employs 180,000 people and sells 9.6 million vehicles per year. Simultaneously, injury rates in manufacturing are down a lot since the 1970s.
I think that both Noble's book and this article would be more persuasive if they committed to one major rhetorical thrust. Either "management is bluffing about how labor can be substituted by machines, so workers shouldn't be intimidated" or "machines can substitute for labor, so workers shouldn't expect the market economy to always provide paid work", but not both.
one study from the period estimated “sixty-five on-the-job deaths per day among autoworkers.”
That is absolutely insane. It makes me think of Upton Sinclair’s 1904 novel The Jungle, about the meat-packing industry.
Do you know of any books that address the safety/exploitation issue specifically?
That is absolutely insane. It makes me think of Upton Sinclair’s 1904 novel The Jungle, about the meat-packing industry.
Do you know of any books that address the safety/exploitation issue specifically?
No, I haven't read any books specifically about the history of workplace safety regulations. I learned bits and pieces second-hand from my father, an engineer who spent the second half of his career doing environmental, health, and safety oversight work at a large manufacturer.
The 180,000 people at GM today for 9.6 million vehicles per year is a bit like 40,000 at Apple producing 100 million iPhones an year. GM is now mostly an assembly plant company for parts made elsewhere plus marketing. Ford ran iron ore mines and rubber plantations. Total verticle integration. They even started Kingsford charcoal to use the hard wood scraps left over from car manufacturing that they grew on their own timerland.
That's a fair point about differing degrees of integration over time. GM spun off Delphi in 1999. I don't know enough about the current auto supply chain to make fractional assignments of supplier companies' employees toward "GM vehicles."
Sure, robots cannot eliminate all jobs, but they can significantly decrease the number of them.
In McDonalds example, cashiers disappeared, but new servers did not magically appear, so the total number of jobs has decreased.
The same way, in "online order" example, author has not realized that while the human still helped with the order, the electronic system likely allowed them to fire one of the cashiers.
I am not sure how elastic the labor supply is -- but if it is not very flexible, then even relatively small changes can affect wages a lot.
In McDonalds example, cashiers disappeared, but new servers did not magically appear, so the total number of jobs has decreased.
The same way, in "online order" example, author has not realized that while the human still helped with the order, the electronic system likely allowed them to fire one of the cashiers.
I am not sure how elastic the labor supply is -- but if it is not very flexible, then even relatively small changes can affect wages a lot.
The article appears to be approaching the issue from a leftist perspective and only telling some of the story.
Here's some contrast: US manufacturing output has risen while manufacturing jobs have fallen.
https://www.mercatus.org/sites/default/files/manufacturing-f... (data is from the Federal Reserve)
Here's some contrast: US manufacturing output has risen while manufacturing jobs have fallen.
https://www.mercatus.org/sites/default/files/manufacturing-f... (data is from the Federal Reserve)
The article mentions that reality of more-output-less-jobs by stating that the workers are required to work harder, because the boss threatens automation if they don't. It is the fauxtomation (the author's word) that is scaring the workers to produce more.
I guess the point of view is leftist because it takes into account the worker's welfare. That doesn't seem like a bad side to be on...
I guess the point of view is leftist because it takes into account the worker's welfare. That doesn't seem like a bad side to be on...
Manufacturing output has risen in terms of revenue. That's different than saying it has risen in terms of widgets sold or some other hard measure, just as its different than in terms of per capita employment.
One reason why it's important to note the distinction is because revenue can continue to rise even if manufacturing efficiency and expertise has remained the same or even declined. High labor costs in mature economies means those economies will tend to have a comparative advantage producing high-margin products even if their manufacturing expertise is declining across the board relative to low-wage countries, and even without automation. This is Economics 101, though professional economists don't always appreciate this, especially academics with an agenda like those at the Mercatus Center.
Without knowing more, the fact that US manufacturing revenue has continued to climb (either in absolute or relative terms) does not refute the argument that American manufacturing is dying; not for most vernacular meanings of dying. In fact, it's perfectly consonant with that argument, and increasing revenue isn't particularly probative one way or another.
(Note that I'm not arguing that US manufacturing is dying. But it does that mean that increasing labor costs in China and the shift of low-margin manufacturing elsewhere should make US manufacturers even more uneasy as it means the economics might finally be in favor of increasingly higher margin industrial processes moving to China. Depending on other factors, that could lead to a rapid, precipitous decline in American manufacturing revenue. Long story short, in the cut-throat world of global capitalism, we should be skeptical of anyone who suggests its okay to rest on our laurels. In many ways such views are even more detrimental than those who bash free markets. The surest way to be caught flat footed by free market volatility is by pretending it doesn't exist.)
One reason why it's important to note the distinction is because revenue can continue to rise even if manufacturing efficiency and expertise has remained the same or even declined. High labor costs in mature economies means those economies will tend to have a comparative advantage producing high-margin products even if their manufacturing expertise is declining across the board relative to low-wage countries, and even without automation. This is Economics 101, though professional economists don't always appreciate this, especially academics with an agenda like those at the Mercatus Center.
Without knowing more, the fact that US manufacturing revenue has continued to climb (either in absolute or relative terms) does not refute the argument that American manufacturing is dying; not for most vernacular meanings of dying. In fact, it's perfectly consonant with that argument, and increasing revenue isn't particularly probative one way or another.
(Note that I'm not arguing that US manufacturing is dying. But it does that mean that increasing labor costs in China and the shift of low-margin manufacturing elsewhere should make US manufacturers even more uneasy as it means the economics might finally be in favor of increasingly higher margin industrial processes moving to China. Depending on other factors, that could lead to a rapid, precipitous decline in American manufacturing revenue. Long story short, in the cut-throat world of global capitalism, we should be skeptical of anyone who suggests its okay to rest on our laurels. In many ways such views are even more detrimental than those who bash free markets. The surest way to be caught flat footed by free market volatility is by pretending it doesn't exist.)
The most recent review that controlled for electronics and computers within the overall US manufacturing numbers show very clearly that the rest of the manufacturing sector is shrinking or stagnant, while the margins associated with relatively stable electronics production are increasing and masking the gutting of the remainder of the sector.
> High labor costs in mature economies means those economies will tend to have a comparative advantage producing high-margin products even if their manufacturing expertise is declining across the board relative to low-wage countries, and even without automation. This is Economics 101
Can you explain this in more detail?
Can you explain this in more detail?
One of the most important, foundational concepts in economics is comparative advantage. See https://en.wikipedia.org/wiki/Comparative_advantage and https://www.investopedia.com/ask/answers/041615/how-do-facto...
Comparative advantage says that even if Country X is better (e.g. more efficient, less costly) than Country Y at producing both Product A and Product B, if it's more better at producing A than B it can be more profitable to them (and everybody) to maximize their output of A and leave Country Y to produce B.
In a relatively low wage country, it's possible (contingent on other factors) that the most profitable thing for that country to do is to maximize its output of the types of products that require large numbers of lower wage laborers and leave other countries to producing products that utilize higher wage labor even if the low wage country would be more efficient at producing all types of products, including those requiring highly skilled labor. Given the reality that the former types of product will tend to be low margin and the latter high margin (and, again, contingent on the exact numbers), it's entirely possible for manufacturing revenue in high wage countries to continue increasing even as the low wage country surpasses it in expertise and efficiency in producing all types of products.
Again, this is all contingent on hard numbers, though FWIW in actuality this is often how things play out at a macro scale. My point is merely that the fact that manufacturing revenue is increasing in Country Y, alone, doesn't tell you much if anything about its efficiency and expertise of manufacturing either products A or B relative to Country X or even in absolute terms; and therefore it doesn't tell you much about the long-term viability and health of those industries. Factors exogenous to the specific skillset related to producing the product, such as labor costs, will often control which products a country makes, regardless of its absolute efficiency in making those specific products.
The upshot is that future manufacturing revenue could be disrupted much faster than you might otherwise expect if it's controlled not by slowly changing, transparent factors (e.g. the generations it takes to build a highly skilled workforce, which may have happened without it being reflected in present market specializations) but by other factors that can change much faster (i.e. whichever factors provide the underlying comparative advantage, such as labor costs, exchange rates, or hidden or poorly understood factors).
Comparative advantage says that even if Country X is better (e.g. more efficient, less costly) than Country Y at producing both Product A and Product B, if it's more better at producing A than B it can be more profitable to them (and everybody) to maximize their output of A and leave Country Y to produce B.
In a relatively low wage country, it's possible (contingent on other factors) that the most profitable thing for that country to do is to maximize its output of the types of products that require large numbers of lower wage laborers and leave other countries to producing products that utilize higher wage labor even if the low wage country would be more efficient at producing all types of products, including those requiring highly skilled labor. Given the reality that the former types of product will tend to be low margin and the latter high margin (and, again, contingent on the exact numbers), it's entirely possible for manufacturing revenue in high wage countries to continue increasing even as the low wage country surpasses it in expertise and efficiency in producing all types of products.
Again, this is all contingent on hard numbers, though FWIW in actuality this is often how things play out at a macro scale. My point is merely that the fact that manufacturing revenue is increasing in Country Y, alone, doesn't tell you much if anything about its efficiency and expertise of manufacturing either products A or B relative to Country X or even in absolute terms; and therefore it doesn't tell you much about the long-term viability and health of those industries. Factors exogenous to the specific skillset related to producing the product, such as labor costs, will often control which products a country makes, regardless of its absolute efficiency in making those specific products.
The upshot is that future manufacturing revenue could be disrupted much faster than you might otherwise expect if it's controlled not by slowly changing, transparent factors (e.g. the generations it takes to build a highly skilled workforce, which may have happened without it being reflected in present market specializations) but by other factors that can change much faster (i.e. whichever factors provide the underlying comparative advantage, such as labor costs, exchange rates, or hidden or poorly understood factors).
That's really interesting, thanks. How would you describe the agenda / bias of the Mercatus centre? I've developed a pretty positive view of them based on what I've seen of Tyler Cowen, Alex Tabarrok and Robin Hanson.
As an alumnus of GMU Law School where the Mercatus Center was once located (maybe still?), I'm hesitant to disparage them. Without a doubt they have very intelligent researchers and students. But the pressure to publish--by both the school and the center's benefactors--is tremendous and, frankly, the research output I saw as a student (often pasted on the walls) never seemed to remotely justify the conclusions. I vividly remember one poster describing MRI scans (with a ridiculously small sample size, not that a large sample would have mattered) used to justify some sketchy hypothesis premised on some neurological/behavioral science that I'm pretty sure has since been debunked. Most of the research I encountered seemed of similar quality.
I'm not a professional economist, but my undergraduate (International Politics) and graduate school (Law) programs heavily emphasized economics. The thing about economics, such as with the Principle of Comparative Advantage, is you can often draw whatever conclusions you'd prefer if you don't rigorously examine, empirically, the actual state of affairs. The parameters matter, and there are alot of them to pin down. That kind of rigorous examination is extremely difficult and expensive, and even huge, well-funded, well-staffed organizations like the World Bank and IMF have historically sucked at performing all but the simplest empirical studies to justify their policy arguments. It's a good bet that most institutions, such as the Mercatus Center, simply lack the resources to do the kind of empirical research necessary to justify their conclusions. So when I see these centers make concrete policy statements or recommendations, I'm always very skeptical.
I used to work at a national security think tank on K Street. I see the Mercatus Center on the same continuum as K Street think tanks; perhaps more academically focused given where its based, but ultimately its work is both fundamentally limited by the inherent complexity of the research area, comparatively meager resources, and strongly tainted by its institutional biases. Importantly, neither the organization nor its researchers acquire funding or prestige by admitting as much; they get funding and prestige by publishing strong, sometimes radical claims based on methodological approaches preordained to provide the answers they sought.
The value of something like the Mercatus Center is in how it chooses to apply basic economic principles. The arguments they make are often theoretically interesting. But, again, the dominate factors in reality are invariably empirical, not theoretical. So at the very least they provide me food for thought, keep me honest, and help me form my own constructive criticisms on issues; but I heavily discount their policy conclusions, just as I do most such organizations (but especially the Mercatus Center because, frankly, they aggressively pursue funding from a very particular subset of benefactors, such as the Koch brothers, which means they're incentivized more than many other organizations to publish results that align with those benefactors' worldview, if not the immediate financial interests of those benefactors).
I'm not a professional economist, but my undergraduate (International Politics) and graduate school (Law) programs heavily emphasized economics. The thing about economics, such as with the Principle of Comparative Advantage, is you can often draw whatever conclusions you'd prefer if you don't rigorously examine, empirically, the actual state of affairs. The parameters matter, and there are alot of them to pin down. That kind of rigorous examination is extremely difficult and expensive, and even huge, well-funded, well-staffed organizations like the World Bank and IMF have historically sucked at performing all but the simplest empirical studies to justify their policy arguments. It's a good bet that most institutions, such as the Mercatus Center, simply lack the resources to do the kind of empirical research necessary to justify their conclusions. So when I see these centers make concrete policy statements or recommendations, I'm always very skeptical.
I used to work at a national security think tank on K Street. I see the Mercatus Center on the same continuum as K Street think tanks; perhaps more academically focused given where its based, but ultimately its work is both fundamentally limited by the inherent complexity of the research area, comparatively meager resources, and strongly tainted by its institutional biases. Importantly, neither the organization nor its researchers acquire funding or prestige by admitting as much; they get funding and prestige by publishing strong, sometimes radical claims based on methodological approaches preordained to provide the answers they sought.
The value of something like the Mercatus Center is in how it chooses to apply basic economic principles. The arguments they make are often theoretically interesting. But, again, the dominate factors in reality are invariably empirical, not theoretical. So at the very least they provide me food for thought, keep me honest, and help me form my own constructive criticisms on issues; but I heavily discount their policy conclusions, just as I do most such organizations (but especially the Mercatus Center because, frankly, they aggressively pursue funding from a very particular subset of benefactors, such as the Koch brothers, which means they're incentivized more than many other organizations to publish results that align with those benefactors' worldview, if not the immediate financial interests of those benefactors).
Even if a bot can't cook burgers correctly, a worker being paid 50-cents an hour in Timbuktu may be cooking burgers via remote-control. At a fast-food restaurant, there may be one human overseer monitoring several remote-controlled bots. I suspect that will happen before direct AI burgers. (It might even be a combo of AI and RC worker to speed response.)
In general it seems pretty clear that the owners of technology are growing ever rich while the rest are stagnating. GDP is growing, but it's not trickling down much since the 1980's. Nobody is sure, but technology appears to be the main "culprit". We may have to adjust our economic systems to avoid log-jamming the fruits of automation at the top.