I don't know what the reserach used for the book was, but it sounds like it might be rather USA-centric.
But onto particular nits about both of your opinions:
1. Entrepreneurs cannot know in advance whether a technology is labor-substituting or labor complementing. That may well be true of major inventions, but the vast majority of innovations/inventions are pretty fully understood by entrepreneurs. BTW, while Jobs wanted the personal computer to be the "bicycle of the mind" Apple's business model hampered software development for teh platform compared to the far more open DOS and later Windows that eventually nearly destroyed Apple.
2. Noah thought a committee contemplating new technologies was teh stupidest idea he had ever heard. yet this is exactly what the Amish do. Kevin Kelly wrote a chapter on this in his book "What Technology Wants" and he seemed more favorable than not.
3. Unions. If unions do nothing, why did US auto manufacturers move to non-union states in teh South? Surely the argument that unions do not benefit workers would suggest that they would not do this. If the argument is that unions do not benefit the worforce as a whole, then that is another discussion. But then they will be arguing against Krugam and Stiglitz, to name just 2.
Interestingly, Mancur Olson would probably argue in favor of the problem of unions based on his thesis that established and rigid institutions (like unions) stifle economic growth. [We certainly saw that in the UK, which eventually resulted in over a decade of Thatcherism and nearly 2 decades of Tory rule.
Most minor" investions are to reduce mnufacturing costs and are therefore predicated on labor reduction which used to be the major cost element. Even if the invention was wildely successful, while an individual firm might increase employment, other firms would shrink. This is very much your initial, and I think correct, argument about demand elasticity. But at heart it is a short-sighted "lump of labor" argument. But any technology that reduces costs will increase its demand if the elasticity is high enough, and that will create new opportunities for labor. Cloth making went from weaving and bulk dying to printing patterns (designers needed) to new clothes designs, and specialty stores and skills to cater for this greater demand for cheaper and therefore more variety of clothes. The Romans were quite boring in dress compared to average citizens of the industrial revolution., let alone the wealthy. Whether Christensen's "Innovators Dilemma" or Brian Arthur's contention that there is a combinatorial explosion with technology as new technologies are created, This creates a greater space for people to get involved in technologies, which enhances employment. Whether that increases wages or just hobby artisanal work depends on the value the work offers.
I have one current hobbyhorse, which is why I wrote a whole book about it (https://patrickmarren.com/about/summary). Entrepreneurs indeed cannot know in advance which technologies will be labor-substituting or labor-complementing. But this is just a single example of our utter inability to predict things that are not trivial. Conversely, ability to predict things renders them strategically trivial. If you can predict it, others can too, and you can eliminate any associated risk by selling it to someone else. But there's no deep liquid market for "societal risk of labor substitution," because it simply cannot be predicted. Which ironically makes it strategically important, at least from a societal perspective. Inventors in my experience feel no particular duty to make their inventions labor-complementing. Nor should they, not because that would not be an admirable thing, but because they can't possibly do it. Reality is a machine for producing unintended consequences, and the ultimate effect on unemployment of any new technology is unknowable by any human, because reality is so infernally complex and interdependent. Even the best possible council for determining complementary vs. substituting technologies would be hard pressed to even establish the full range of plausible future outcomes for any one technology. I know this because I've spent 30+ years in rooms with smart people trying to come to grips with such questions. And Apple's business model is a great example of the unknowability of the ramifications of technologies. At first it seemed really cool and successful and profitable; then it was "eventually nearly destroyed" by MS/DOS and open architecture; then it became a monster of success; in ten years, where will it be? We don't know. And we can't. This applies to the vast majority of strategic business/organizational issues, far more than is generally admitted by businesspeople, inventors, tech-bros, economists - you name it. It's a huge issue. Which is why I wrote a book on it. The only cure for this kind of strategic uncertainty (Frank Knight's "higher uncertainty") is to fully flesh out all plausible large-scale future outcomes, using imagination rigorously. Entrepreneurs (and I would say inventors), as Knight noted, are not engaged in rational prediction. They are engaged in the societally necessary task of taking shots in near-total darkness. If entrepreneurs were restricted to what was rationally predictable, economic progress would largely grind to a halt.
I was thinking about this recently while at my father's bedside, listening with him to the Hexapodia podcast (and others). I have mentioned him previously, Brad, as striding around his house, reciting Macaulay's Horatius, thanks to your dredging it up recently. He enjoyed "Slouching," and you and Noah on Hexapodia. He was, in a certain sense, exactly the sort of "entrepreneur" Frank Knight cited as necessary to a smoothly functioning economy. Against fearful odds, he produced nine kids and raised them in cramped quarters, and all of them went to college. He died a few weeks ago. "To every man upon this earth /Death cometh soon or late. / And how can man die better / Than facing fearful odds, / For the ashes of his fathers, And the temples of his Gods." Thanks for diverting him (and me) in his last illness, boys.
I think the latest opinion concerning the Luddites (at least according to progressives) is that the Luddites weren't opposed to progress or making cloth in factories. Rather they were opposed to the owners not sharing the wealth with them that was being created by the greater use of machinery. Of course this is the same argument labor unions are still making today.
In a Malthusian economy, extraction ratios to an extent. More tax, fewer people, same subsistence incomes.
But even with infinite state capacity, I suspect at some point further extraction is not optimal because pop density is useful for a ruler.
Actually, pet theory, but I think it depends on what military technology complements. If you want to field heavy cataphracts, squeeze the coloni. If you want massed infantry, maybe you need population density.
In the UK the Ditto Machine-like duplicator was called a Mimeogragh. Noah should be put in front of one to see if he can work out how to use it. Video like the kids being given cassett tape machines or dial phones. [To be fair, I couldn't recall how to do a carriage return on a manual typewriter in the mid-1990s until my memory clicked in. The CR or return key on a keyboard had largely erased that memory with a new one.]
It would be fantastic if you got Devereaux on your podcast. Did you listen to him interviewed by the Partial Historians? It had the side effect of getting me to listen to more Partial Historians podcasts (the from the founding of Rome ones.)
But regardless of whether A&J are right or not, is it not good to remove taxes on employment such as a) employer “provision” of health insurance, b) finance of SS/Medicare with a wage tax instead of a VAT, c) replace minimum wages with wage subsidy/reformed EITC?
Well, _I_ remember how the Ditto machine was a huge improvement over the “?” which transferred ink onto a plate of Jello and then the Jello transfers ink to ~ 20 sheet of paper. Working in such an office financed my BS in economics in 1964.
Can we (you, not me!) write out a nice little homothetic production function such that with a parameter at one value instead of another or the ratio of certain parameters greater than or less than so-and-so, some trajectory of K, L1, L2 … Ln, T leads to rising; falling MP of L1, L2, Ln and use it to show “complementarity/substitution” comparative/absolute advantage.”
[Hasn’t this been done already?]
So Acemoglu’s model is T1, T2, Tn? But that implies TT = f(T1, T2 ...Tn) and how is TT modeled?
To Eastern Europe first, IIRC. Radiology is interesting as I suspect that the machines, especially MRIs restrict demand for reading the output which should imply that AI will reduce the number of radiologists in that service. X-rays, OTOH are cheap, quick, and could well increase demand. The UK is ensuring that every output of a detected issue is reviewsd by a radiologist that will ensure employment. The AIs are just to do the boring work to dismiss moss outputs as not of a problem nature. [I believe the testing labs use AIs to look at biopsy samples to flag the small percentage that have abnormalities like cancers.]
I don't know what the reserach used for the book was, but it sounds like it might be rather USA-centric.
But onto particular nits about both of your opinions:
1. Entrepreneurs cannot know in advance whether a technology is labor-substituting or labor complementing. That may well be true of major inventions, but the vast majority of innovations/inventions are pretty fully understood by entrepreneurs. BTW, while Jobs wanted the personal computer to be the "bicycle of the mind" Apple's business model hampered software development for teh platform compared to the far more open DOS and later Windows that eventually nearly destroyed Apple.
2. Noah thought a committee contemplating new technologies was teh stupidest idea he had ever heard. yet this is exactly what the Amish do. Kevin Kelly wrote a chapter on this in his book "What Technology Wants" and he seemed more favorable than not.
3. Unions. If unions do nothing, why did US auto manufacturers move to non-union states in teh South? Surely the argument that unions do not benefit workers would suggest that they would not do this. If the argument is that unions do not benefit the worforce as a whole, then that is another discussion. But then they will be arguing against Krugam and Stiglitz, to name just 2.
Interestingly, Mancur Olson would probably argue in favor of the problem of unions based on his thesis that established and rigid institutions (like unions) stifle economic growth. [We certainly saw that in the UK, which eventually resulted in over a decade of Thatcherism and nearly 2 decades of Tory rule.
Most minor" investions are to reduce mnufacturing costs and are therefore predicated on labor reduction which used to be the major cost element. Even if the invention was wildely successful, while an individual firm might increase employment, other firms would shrink. This is very much your initial, and I think correct, argument about demand elasticity. But at heart it is a short-sighted "lump of labor" argument. But any technology that reduces costs will increase its demand if the elasticity is high enough, and that will create new opportunities for labor. Cloth making went from weaving and bulk dying to printing patterns (designers needed) to new clothes designs, and specialty stores and skills to cater for this greater demand for cheaper and therefore more variety of clothes. The Romans were quite boring in dress compared to average citizens of the industrial revolution., let alone the wealthy. Whether Christensen's "Innovators Dilemma" or Brian Arthur's contention that there is a combinatorial explosion with technology as new technologies are created, This creates a greater space for people to get involved in technologies, which enhances employment. Whether that increases wages or just hobby artisanal work depends on the value the work offers.
I have one current hobbyhorse, which is why I wrote a whole book about it (https://patrickmarren.com/about/summary). Entrepreneurs indeed cannot know in advance which technologies will be labor-substituting or labor-complementing. But this is just a single example of our utter inability to predict things that are not trivial. Conversely, ability to predict things renders them strategically trivial. If you can predict it, others can too, and you can eliminate any associated risk by selling it to someone else. But there's no deep liquid market for "societal risk of labor substitution," because it simply cannot be predicted. Which ironically makes it strategically important, at least from a societal perspective. Inventors in my experience feel no particular duty to make their inventions labor-complementing. Nor should they, not because that would not be an admirable thing, but because they can't possibly do it. Reality is a machine for producing unintended consequences, and the ultimate effect on unemployment of any new technology is unknowable by any human, because reality is so infernally complex and interdependent. Even the best possible council for determining complementary vs. substituting technologies would be hard pressed to even establish the full range of plausible future outcomes for any one technology. I know this because I've spent 30+ years in rooms with smart people trying to come to grips with such questions. And Apple's business model is a great example of the unknowability of the ramifications of technologies. At first it seemed really cool and successful and profitable; then it was "eventually nearly destroyed" by MS/DOS and open architecture; then it became a monster of success; in ten years, where will it be? We don't know. And we can't. This applies to the vast majority of strategic business/organizational issues, far more than is generally admitted by businesspeople, inventors, tech-bros, economists - you name it. It's a huge issue. Which is why I wrote a book on it. The only cure for this kind of strategic uncertainty (Frank Knight's "higher uncertainty") is to fully flesh out all plausible large-scale future outcomes, using imagination rigorously. Entrepreneurs (and I would say inventors), as Knight noted, are not engaged in rational prediction. They are engaged in the societally necessary task of taking shots in near-total darkness. If entrepreneurs were restricted to what was rationally predictable, economic progress would largely grind to a halt.
I was thinking about this recently while at my father's bedside, listening with him to the Hexapodia podcast (and others). I have mentioned him previously, Brad, as striding around his house, reciting Macaulay's Horatius, thanks to your dredging it up recently. He enjoyed "Slouching," and you and Noah on Hexapodia. He was, in a certain sense, exactly the sort of "entrepreneur" Frank Knight cited as necessary to a smoothly functioning economy. Against fearful odds, he produced nine kids and raised them in cramped quarters, and all of them went to college. He died a few weeks ago. "To every man upon this earth /Death cometh soon or late. / And how can man die better / Than facing fearful odds, / For the ashes of his fathers, And the temples of his Gods." Thanks for diverting him (and me) in his last illness, boys.
I think the latest opinion concerning the Luddites (at least according to progressives) is that the Luddites weren't opposed to progress or making cloth in factories. Rather they were opposed to the owners not sharing the wealth with them that was being created by the greater use of machinery. Of course this is the same argument labor unions are still making today.
Want to give a take on Goria:
https://www.honest-broker.com/p/they-praised-ai-at-sxswand-the-audience?utm_source=post-email-title&publication_id=296132&post_id=142653042&utm_campaign=email-post-title&isFreemail=true&r=8ylpe&triedRedirect=true&utm_medium=email
How could we quantify Roman-esque bad technology?
In a Malthusian economy, extraction ratios to an extent. More tax, fewer people, same subsistence incomes.
But even with infinite state capacity, I suspect at some point further extraction is not optimal because pop density is useful for a ruler.
Actually, pet theory, but I think it depends on what military technology complements. If you want to field heavy cataphracts, squeeze the coloni. If you want massed infantry, maybe you need population density.
In the UK the Ditto Machine-like duplicator was called a Mimeogragh. Noah should be put in front of one to see if he can work out how to use it. Video like the kids being given cassett tape machines or dial phones. [To be fair, I couldn't recall how to do a carriage return on a manual typewriter in the mid-1990s until my memory clicked in. The CR or return key on a keyboard had largely erased that memory with a new one.]
It would be fantastic if you got Devereaux on your podcast. Did you listen to him interviewed by the Partial Historians? It had the side effect of getting me to listen to more Partial Historians podcasts (the from the founding of Rome ones.)
But regardless of whether A&J are right or not, is it not good to remove taxes on employment such as a) employer “provision” of health insurance, b) finance of SS/Medicare with a wage tax instead of a VAT, c) replace minimum wages with wage subsidy/reformed EITC?
See [https://thomaslhutcheson.substack.com/p/ai-and-labors-share]
Well, _I_ remember how the Ditto machine was a huge improvement over the “?” which transferred ink onto a plate of Jello and then the Jello transfers ink to ~ 20 sheet of paper. Working in such an office financed my BS in economics in 1964.
Can we (you, not me!) write out a nice little homothetic production function such that with a parameter at one value instead of another or the ratio of certain parameters greater than or less than so-and-so, some trajectory of K, L1, L2 … Ln, T leads to rising; falling MP of L1, L2, Ln and use it to show “complementarity/substitution” comparative/absolute advantage.”
[Hasn’t this been done already?]
So Acemoglu’s model is T1, T2, Tn? But that implies TT = f(T1, T2 ...Tn) and how is TT modeled?
To Eastern Europe first, IIRC. Radiology is interesting as I suspect that the machines, especially MRIs restrict demand for reading the output which should imply that AI will reduce the number of radiologists in that service. X-rays, OTOH are cheap, quick, and could well increase demand. The UK is ensuring that every output of a detected issue is reviewsd by a radiologist that will ensure employment. The AIs are just to do the boring work to dismiss moss outputs as not of a problem nature. [I believe the testing labs use AIs to look at biopsy samples to flag the small percentage that have abnormalities like cancers.]