The problem in BCs analysis comes from lumping "technology" into a single category. Medical, communication, transport, physical-labor replacement, human computation/tabulation replacement certainty eliminates SOME jobs, but those past tech waves have generally reduced, but not eliminated human effort in production generally. What if instead we have ZERO humans needed, for agriculture, mining, transport, storage, processing, medical treatment & diagnosis ?
Some problems are ... How do we distribute goods without jobs ? Why should any producer make product for "customers" who do not/cannot pay ?
I anticipate that 'lights-out factories" will expand, but in balance with transfer of employment to "service" jobs. Can we have a society based only on service jobs ?
It's not impossible AI will be different. But I doubt it'd lead to nation wide layoffs over months, it'd probably be over a few years at absolute worst. And there'd be plenty of time to respond with redistributive welfare then as it's needed instead of trying to preemptively set it up when we don't need it.
It would take many decades, but the incremental changes to production are limited by the number of ppl able to pay for the product. Who gets the profits ?
My job is literally to automate away work, mostly accounting. The issue is not the overall number of jobs. The issue is whether we can retrain accountants to data scientists (that looks like a popular, growing sector now), or will run into some obstacle. Remember when journos were told to learn to code. The thing is, not everybody can.
So it is possible we run into the situation where there are plenty of jobs, but the existing people are not a good match for those.
Shouldn’t it be “nest of eggheads”? Also, since the unemployment rate excludes those who are not looking for work, the argument would be that technological unemployment would be invisible on such a chart, since the workers would disappear once UC/UI benefits ran out. The technologically unemployed wouldn’t show up on the chart, so the theory would go, until those workers were either ready to do something else, or were desperate to do anything. And some of those would just surf from wave to wave of unemployment as the rising tide of technology swept along. The 2020 unemployment rate, which was off the chart, both literally and figuratively, kind of refutes that thinking, since it was at least partly technological innovations that brought the rate down so fast.
Even with high tech growth, the "unemployment rate" is low. It's just most of the jobs are crappy, low-wage jobs that don't support a person or family.
Bryan did your read 'the rise of bullshit jobs' my David Graeber? I found it eye opening. I'd love your take on it. According to the book technology didn't replace jobs because there was a big raise in jobs with no real purpose. Jobs for complying with ridiculous administrative rules, jobs that could be replaced by technology but they are not precisely because they would imply lying people off, etc.
Not Bryan, but I found it preposterous. If there is one thing the entire econo-political spectrum from Marxists to Rothbardists agree on, it is that that corporations exist to make profits, not to give rich people bragging rights about how many dependents they have!
All those jobs make sense if you dig deeper. Ridiculous administrative rules are well there was one screwup, heads were rolling, so now the higher ups are protecting their butts. And so on.
Unemployment doesn’t account for people switching careers, due to redundancy. Automated factories may not have reduced total jobs but there sure are a lot less cabinet makers enjoying their craft.
Not only is this stupid, it’s smug and callous - this is why economists are so disliked (I’m one myself).
Do people work in “the economy”, or “the labor force?” No, they work specific jobs in specific industries, using specific skills they’ve taken time to acquire (“human capital”, if you will). Your claim - that individuals have nothing to worry about because overall employment rates are not determined by technological progress - is an invalidation of fears that have been and will continue to come true.
Of course, economists blithely claim that “creative destruction” - which you conflate with technological unemployment - will ensure that people still have jobs. But even if that were true, there’s nothing guaranteeing that the obsolete people will re-train into more modern jobs. There’s plenty of literature on the failures (defined by you as anything <100% success, if people have nothing to worry about) of job retraining programs.
Are you so sure that individuals have nothing to worry about?
I don't seriously doubt that this will hold true up to a point. Sooner or later, though, there won't be anything humans can do that AIs and their robot bodies can't do better, faster, and cheaper. You can quibble about comparative advantage, but at the end of the day I find it hard to believe that human labor will be economically meaningful in such a world. I think Bryan's point here is mostly a good one. As long as AI remains a "normal" technology, it will have normal effects. Once it becomes, you know, another sapient species on Earth...
People dread the prospect of _change_. They don’t like to lose their jobs, even if they get another, better job reasonably quickly.
People who claim to like change don’t want it to happen to _them_. They want to choose changes for themselves. Most actually like variety; temporary changes they choose for themselves.
The problem in BCs analysis comes from lumping "technology" into a single category. Medical, communication, transport, physical-labor replacement, human computation/tabulation replacement certainty eliminates SOME jobs, but those past tech waves have generally reduced, but not eliminated human effort in production generally. What if instead we have ZERO humans needed, for agriculture, mining, transport, storage, processing, medical treatment & diagnosis ?
Some problems are ... How do we distribute goods without jobs ? Why should any producer make product for "customers" who do not/cannot pay ?
I anticipate that 'lights-out factories" will expand, but in balance with transfer of employment to "service" jobs. Can we have a society based only on service jobs ?
most who claim about Technological unemployment will say you will find it in the Labor force participation rate and not the unemployment rate
An economist worries about unemployment trends. A normal person worries about his particular job.
This is also why normal people worry about being replaced by foreigners.
This is, further, why normal people don't worry about what economists think.
It's not impossible AI will be different. But I doubt it'd lead to nation wide layoffs over months, it'd probably be over a few years at absolute worst. And there'd be plenty of time to respond with redistributive welfare then as it's needed instead of trying to preemptively set it up when we don't need it.
It would take many decades, but the incremental changes to production are limited by the number of ppl able to pay for the product. Who gets the profits ?
My job is literally to automate away work, mostly accounting. The issue is not the overall number of jobs. The issue is whether we can retrain accountants to data scientists (that looks like a popular, growing sector now), or will run into some obstacle. Remember when journos were told to learn to code. The thing is, not everybody can.
So it is possible we run into the situation where there are plenty of jobs, but the existing people are not a good match for those.
Shouldn’t it be “nest of eggheads”? Also, since the unemployment rate excludes those who are not looking for work, the argument would be that technological unemployment would be invisible on such a chart, since the workers would disappear once UC/UI benefits ran out. The technologically unemployed wouldn’t show up on the chart, so the theory would go, until those workers were either ready to do something else, or were desperate to do anything. And some of those would just surf from wave to wave of unemployment as the rising tide of technology swept along. The 2020 unemployment rate, which was off the chart, both literally and figuratively, kind of refutes that thinking, since it was at least partly technological innovations that brought the rate down so fast.
Even with high tech growth, the "unemployment rate" is low. It's just most of the jobs are crappy, low-wage jobs that don't support a person or family.
Bryan did your read 'the rise of bullshit jobs' my David Graeber? I found it eye opening. I'd love your take on it. According to the book technology didn't replace jobs because there was a big raise in jobs with no real purpose. Jobs for complying with ridiculous administrative rules, jobs that could be replaced by technology but they are not precisely because they would imply lying people off, etc.
Not Bryan, but I found it preposterous. If there is one thing the entire econo-political spectrum from Marxists to Rothbardists agree on, it is that that corporations exist to make profits, not to give rich people bragging rights about how many dependents they have!
All those jobs make sense if you dig deeper. Ridiculous administrative rules are well there was one screwup, heads were rolling, so now the higher ups are protecting their butts. And so on.
It might surprise you to read this book
Unemployment doesn’t account for people switching careers, due to redundancy. Automated factories may not have reduced total jobs but there sure are a lot less cabinet makers enjoying their craft.
Not only is this stupid, it’s smug and callous - this is why economists are so disliked (I’m one myself).
Do people work in “the economy”, or “the labor force?” No, they work specific jobs in specific industries, using specific skills they’ve taken time to acquire (“human capital”, if you will). Your claim - that individuals have nothing to worry about because overall employment rates are not determined by technological progress - is an invalidation of fears that have been and will continue to come true.
Of course, economists blithely claim that “creative destruction” - which you conflate with technological unemployment - will ensure that people still have jobs. But even if that were true, there’s nothing guaranteeing that the obsolete people will re-train into more modern jobs. There’s plenty of literature on the failures (defined by you as anything <100% success, if people have nothing to worry about) of job retraining programs.
Are you so sure that individuals have nothing to worry about?
The smarter take is that technology increases real per hour labor costs.
I don't seriously doubt that this will hold true up to a point. Sooner or later, though, there won't be anything humans can do that AIs and their robot bodies can't do better, faster, and cheaper. You can quibble about comparative advantage, but at the end of the day I find it hard to believe that human labor will be economically meaningful in such a world. I think Bryan's point here is mostly a good one. As long as AI remains a "normal" technology, it will have normal effects. Once it becomes, you know, another sapient species on Earth...
People dread the prospect of _change_. They don’t like to lose their jobs, even if they get another, better job reasonably quickly.
People who claim to like change don’t want it to happen to _them_. They want to choose changes for themselves. Most actually like variety; temporary changes they choose for themselves.
People prefer incremental change.