As the tech sector gets increasingly productive compared to the rest, people will be more and more interested in plundering it and less and less interested in doing something productive themselves.
"Keynes just looked at the productivity development of the day and concluded that after some decades it would take much less time to produce the necessities of life. He somehow assumed that as soon as people had reached a certain level of material comfort, they would just live and let live.
Why did he assume that? Such a thing hasn't happened at any time, at any place ever. "
Really? How long ago was it that most people worked 6 or 7 days a week? How many currently work less than "full time"? How much has average hours worked changed in the last century? The last decade?
How much has average hours of unpaid housework dropped in the last century? The last decade?
"Andreessen argues against a Universal Basic Income on the grounds that humans are made to be productive." Yes, sumans ARE "made to be productive". Trouble is, Andreessen's understanding (or preaching) of "productive" is terribly flawed. See https://mfioretti.substack.com/p/ubi-ai-and-reality-always-in-the
That baumol's cost disease exists because people need to feel productive/purposeful/useful/valuable (etc) is a great insight.
I'll look into MA, who I've presumed was a pointy headed american blow hard.
If he's arguing "against a Universal Basic Income on the grounds that humans are made to be productive" I'm with him.
However, while State charity is destructive of one's soul, I'm not sure having baumol consume most of the economy is a better option. Maybe Keynes was correct: we should all be 'working' waaaaay less by relying on ai to facilitate our productive efforts and allocate the great majority of our time to forms of self-expression (art and craft, etc) with our fellow humans. Not dissimilar to the apparent approach of that Norwegian chess champion.
I don't have time to write a more detailed/thoughtful version of this, but I'll start: I am immediately annoyed by and suspicious of any analysis that partitions the work that various people do into two groups, "those who do productive work" and "parasites". Certainly Andreesen isn't the first at this; the US city of Trenton has the slogan "Trenton makes, the world takes" in lights on its main bridge. That dates back to when Trenton was a major manufacturing city.
It's true that there is parasitism, but for most authors, the division turns out to be between "those who do work I understand / directly benefits me / of the type I do" and "everybody else".
Marc Andreessen (and, in this case, you) have sidestepped the central question of AI causing joblessness. In his case, it seems almost deliberate -- it is for certain solidly in the middle of the lane for a tech-optimist's response. The actual argument (not mine -- my concerns are elsewhere) is that any discussion about jobs and automation that draws from historical data (including e.g. last week) is on the wrong topic. The printing press makes obsolete the job of copyist, letting those who would have been copyists do something else. Sure there's some disruption at first (an abbot who wrote a tract, not long after gutenberg, explaining that monks would always still copy manuscripts had the tract printed to increase its circulation; oops), but the workers go elsewhere. Same with farmers to tractors or buggy-whip-manufacturers to car assembly line workers. Actual AI wouldn't do anything similar. AI would render obsolete not any one particular kind of intellectual work (leaving other kinds), but would rather make all intellectual workers themselves completely and permanently obsolete. Yes in theory a human plumber will still be needed to fix your physical drain, but true AI would likely lead to advances in robotics as well. Even if somehow it doesn't, and people still need human plumbers and framers and suchwhat, eventually it'll be cheaper to give any random biped a pair of goggles that show him how to assemble the pipes or sticks of lumber that come marked from the plant. Plumbers aren't paid well because of the courage it takes to go under a house where the sewer is leaking, but because they have to solve problems when they're down there. Even if somehow (in an AGI world) the goggles are never cheap enough to let an illegal immigrant do the work for peanuts, whose homes will plumbers then be fixing up, and for what money?
Now of course I don't think we'll get there, because I don't think we're on track for true AGI. But I do expect automated tools to shrink the size of the white-collar jobs market enough to cause real squeezes further down (which will coincide with us getting poorer for the same other reasons we've been getting poorer lately).
I kept reading "Kling" as "killing" (in the article and in the comments) while debating with myself about whether or not to point out the only thing that could be done with bureaucrats and regulators that would improve things. Freud would nod knowingly at me.
This resonated with me a lot, since my country is currently in turmoil over many issues, one of which is that there is a class of ultra-religious people who do not work and do not participate in society in any constructive way, but still receive considerable subsidies from the government, as they are in the ruling coalition.
So here the situation is much worse than “The simple and boring answer is that no one knows which rules, inspectors and bureaucracy are actually necessary”. Every reasonable person (and even economists ;) ) know that having a large part of the population that doesn't participate in the work-force and is not educated in basic STEM subjects but still receives a large part of the pie is untenable and unreasonable. But political power enables this new "warrior class" to take its unjust share of the spoils in any case.
So I'm less optimistic about UBI. I think that when most people get UBI but are still the powerful majority, they will want more than the Basic. I think they'll want an increasing share of the pie, to the point where it will seriously hurt production.
That's what happened with many of the warrior classes, did it not? The peasants were squeezed to the last drop and when the first drought or flood arrived, the whole economy collapsed and many died. This has happened many times.
So I suppose just "paying them off" would not be enough to stave off the collapse for long. Either we need to find a way to make more of the society productive, or we need to somehow ensure that the majority of the population doesn't enact self-serving laws (which means minority-rule and is probably very problematic from a human-rights angle).
Let me start by echoing Arnold Kling that "Wood from Eden is in general a great substack", and it's particularly valuable because it doesn't come out of the conventional "intellectual" class.
But let me disagree with Andreesen in a number of ways.
First, I suspect his numbers for "Food and Beverages" are "incorrect"; the information I've heard is that the fraction of people's incomes they spend on food has been decreasing steadily. It may be that that fraction hasn't decreased in the US in the last 20 years, but over the long run it does. Witness that the most destitute people are now "the homeless" rather than "the starving".
And you have to be careful what you count. According to the USDA "Food prices and spending", spending on "food at home" was 3.5 times spending on "food away from home" in 1962, but now "food away from home" is slightly higher. I suspect that just over half of Andreesen's "Food and Beverages" is actually restaurants, where most of the cost is labor rather than the actual food.
Looking at the red categories, most of them are services. You would naturally expect childcare to rise exactly with wages, because childcare is essentially the labor of average people. (This will change suddenly when we figure out a workable childcare robot.) Medical and educational services are done by highly-skilled labor and it's difficult to change that, which means that its costs are going to rise with wages. In principle, you can replace the teachers with the Internet, but in practice it's hard:
"The current interest in 'competency-based learning'--liberating students to earn degrees not by amassing credit hours but by preparing for assessments of particular skills at whatever pace and by whichever route they choose--is part of the same trend. Some reformers see the seeds of a revolution in college education, promising ultraconvenient, self-guided, low-cost courses of study for everyone. The 'beginning of the unbundling of the American university' is how one observer has described the transformation. All it will take for students to avail themselves of this emerging opportunity is a clear sense of where they're headed, lots of self-motivation, and good access to information about what mix of skills is likely to lead to a promising career. And therein, of course, lies the problem." -- "How to Escape the Community-College Trap" by Ann Hulbert
Housing cost in major metro areas, of course, is dominated by the fact that constructing housing is highly regulated, partly in its "quality" but in most places very strictly in its *quantity*.
The blue categories are not so much regulated as manufactured things, where technology easily reduces the amount of labor needed to make one unit (as long as the number of units produced is very large). Clothing is the only obvious exception; clothing is labor-intensive, but people in Bangladesh can be hired to do the work.
"You can have a piece of the economic pie, as long as you do a make-work job neither you nor the person receiving it like too much."
Of course in some way having to go to the make-work jobs gatekeeps things. What's "basic" anyway. Why not basic + 1, etc to infinity. As long as all it takes is I have to complain until a large # gets deposited into my account.
“The simple and boring answer is that no one knows which rules, inspectors and bureaucracy are actually necessary.”
While that is a simple answer, I don’t think it bears much scrutiny. In fact, most of your article seems to argue against it. Most people could easily identify most counterproductive regulations. The victims of regulation do not push back because they are weak and the regulators are strong. Why don’t you ignore the bureaucrats supervising the construction of your house? Because you know that they could crush you and you also (likely rightly) surmise that they would delight in doing so.
College textbooks are starting to be replaced by open source versions. My wife always teaches her logic class out of forall x from the open logic project.
The textbook vendors are desperately working to create lock-in and exploit principal-agent misalignment but long term it's a losing battle.
"Keynes just looked at the productivity development of the day and concluded that after some decades it would take much less time to produce the necessities of life. He somehow assumed that as soon as people had reached a certain level of material comfort, they would just live and let live.
Why did he assume that? Such a thing hasn't happened at any time, at any place ever. "
Really? How long ago was it that most people worked 6 or 7 days a week? How many currently work less than "full time"? How much has average hours worked changed in the last century? The last decade?
How much has average hours of unpaid housework dropped in the last century? The last decade?
"Andreessen argues against a Universal Basic Income on the grounds that humans are made to be productive." Yes, sumans ARE "made to be productive". Trouble is, Andreessen's understanding (or preaching) of "productive" is terribly flawed. See https://mfioretti.substack.com/p/ubi-ai-and-reality-always-in-the
That baumol's cost disease exists because people need to feel productive/purposeful/useful/valuable (etc) is a great insight.
I'll look into MA, who I've presumed was a pointy headed american blow hard.
If he's arguing "against a Universal Basic Income on the grounds that humans are made to be productive" I'm with him.
However, while State charity is destructive of one's soul, I'm not sure having baumol consume most of the economy is a better option. Maybe Keynes was correct: we should all be 'working' waaaaay less by relying on ai to facilitate our productive efforts and allocate the great majority of our time to forms of self-expression (art and craft, etc) with our fellow humans. Not dissimilar to the apparent approach of that Norwegian chess champion.
The things that are more affordable can be provided by Chinese slave labor; the things that are less affordable cannot.
I don't have time to write a more detailed/thoughtful version of this, but I'll start: I am immediately annoyed by and suspicious of any analysis that partitions the work that various people do into two groups, "those who do productive work" and "parasites". Certainly Andreesen isn't the first at this; the US city of Trenton has the slogan "Trenton makes, the world takes" in lights on its main bridge. That dates back to when Trenton was a major manufacturing city.
It's true that there is parasitism, but for most authors, the division turns out to be between "those who do work I understand / directly benefits me / of the type I do" and "everybody else".
Marc Andreessen (and, in this case, you) have sidestepped the central question of AI causing joblessness. In his case, it seems almost deliberate -- it is for certain solidly in the middle of the lane for a tech-optimist's response. The actual argument (not mine -- my concerns are elsewhere) is that any discussion about jobs and automation that draws from historical data (including e.g. last week) is on the wrong topic. The printing press makes obsolete the job of copyist, letting those who would have been copyists do something else. Sure there's some disruption at first (an abbot who wrote a tract, not long after gutenberg, explaining that monks would always still copy manuscripts had the tract printed to increase its circulation; oops), but the workers go elsewhere. Same with farmers to tractors or buggy-whip-manufacturers to car assembly line workers. Actual AI wouldn't do anything similar. AI would render obsolete not any one particular kind of intellectual work (leaving other kinds), but would rather make all intellectual workers themselves completely and permanently obsolete. Yes in theory a human plumber will still be needed to fix your physical drain, but true AI would likely lead to advances in robotics as well. Even if somehow it doesn't, and people still need human plumbers and framers and suchwhat, eventually it'll be cheaper to give any random biped a pair of goggles that show him how to assemble the pipes or sticks of lumber that come marked from the plant. Plumbers aren't paid well because of the courage it takes to go under a house where the sewer is leaking, but because they have to solve problems when they're down there. Even if somehow (in an AGI world) the goggles are never cheap enough to let an illegal immigrant do the work for peanuts, whose homes will plumbers then be fixing up, and for what money?
Now of course I don't think we'll get there, because I don't think we're on track for true AGI. But I do expect automated tools to shrink the size of the white-collar jobs market enough to cause real squeezes further down (which will coincide with us getting poorer for the same other reasons we've been getting poorer lately).
I kept reading "Kling" as "killing" (in the article and in the comments) while debating with myself about whether or not to point out the only thing that could be done with bureaucrats and regulators that would improve things. Freud would nod knowingly at me.
This resonated with me a lot, since my country is currently in turmoil over many issues, one of which is that there is a class of ultra-religious people who do not work and do not participate in society in any constructive way, but still receive considerable subsidies from the government, as they are in the ruling coalition.
So here the situation is much worse than “The simple and boring answer is that no one knows which rules, inspectors and bureaucracy are actually necessary”. Every reasonable person (and even economists ;) ) know that having a large part of the population that doesn't participate in the work-force and is not educated in basic STEM subjects but still receives a large part of the pie is untenable and unreasonable. But political power enables this new "warrior class" to take its unjust share of the spoils in any case.
So I'm less optimistic about UBI. I think that when most people get UBI but are still the powerful majority, they will want more than the Basic. I think they'll want an increasing share of the pie, to the point where it will seriously hurt production.
That's what happened with many of the warrior classes, did it not? The peasants were squeezed to the last drop and when the first drought or flood arrived, the whole economy collapsed and many died. This has happened many times.
So I suppose just "paying them off" would not be enough to stave off the collapse for long. Either we need to find a way to make more of the society productive, or we need to somehow ensure that the majority of the population doesn't enact self-serving laws (which means minority-rule and is probably very problematic from a human-rights angle).
I sure hope that I'm mistaken, though :(
Let me start by echoing Arnold Kling that "Wood from Eden is in general a great substack", and it's particularly valuable because it doesn't come out of the conventional "intellectual" class.
But let me disagree with Andreesen in a number of ways.
First, I suspect his numbers for "Food and Beverages" are "incorrect"; the information I've heard is that the fraction of people's incomes they spend on food has been decreasing steadily. It may be that that fraction hasn't decreased in the US in the last 20 years, but over the long run it does. Witness that the most destitute people are now "the homeless" rather than "the starving".
And you have to be careful what you count. According to the USDA "Food prices and spending", spending on "food at home" was 3.5 times spending on "food away from home" in 1962, but now "food away from home" is slightly higher. I suspect that just over half of Andreesen's "Food and Beverages" is actually restaurants, where most of the cost is labor rather than the actual food.
Looking at the red categories, most of them are services. You would naturally expect childcare to rise exactly with wages, because childcare is essentially the labor of average people. (This will change suddenly when we figure out a workable childcare robot.) Medical and educational services are done by highly-skilled labor and it's difficult to change that, which means that its costs are going to rise with wages. In principle, you can replace the teachers with the Internet, but in practice it's hard:
"The current interest in 'competency-based learning'--liberating students to earn degrees not by amassing credit hours but by preparing for assessments of particular skills at whatever pace and by whichever route they choose--is part of the same trend. Some reformers see the seeds of a revolution in college education, promising ultraconvenient, self-guided, low-cost courses of study for everyone. The 'beginning of the unbundling of the American university' is how one observer has described the transformation. All it will take for students to avail themselves of this emerging opportunity is a clear sense of where they're headed, lots of self-motivation, and good access to information about what mix of skills is likely to lead to a promising career. And therein, of course, lies the problem." -- "How to Escape the Community-College Trap" by Ann Hulbert
Housing cost in major metro areas, of course, is dominated by the fact that constructing housing is highly regulated, partly in its "quality" but in most places very strictly in its *quantity*.
The blue categories are not so much regulated as manufactured things, where technology easily reduces the amount of labor needed to make one unit (as long as the number of units produced is very large). Clothing is the only obvious exception; clothing is labor-intensive, but people in Bangladesh can be hired to do the work.
"You can have a piece of the economic pie, as long as you do a make-work job neither you nor the person receiving it like too much."
Of course in some way having to go to the make-work jobs gatekeeps things. What's "basic" anyway. Why not basic + 1, etc to infinity. As long as all it takes is I have to complain until a large # gets deposited into my account.
https://www.youtube.com/watch?v=nszh_JYpl_E
https://www.youtube.com/watch?v=V8U3gf-im8E
“The simple and boring answer is that no one knows which rules, inspectors and bureaucracy are actually necessary.”
While that is a simple answer, I don’t think it bears much scrutiny. In fact, most of your article seems to argue against it. Most people could easily identify most counterproductive regulations. The victims of regulation do not push back because they are weak and the regulators are strong. Why don’t you ignore the bureaucrats supervising the construction of your house? Because you know that they could crush you and you also (likely rightly) surmise that they would delight in doing so.
College textbooks are starting to be replaced by open source versions. My wife always teaches her logic class out of forall x from the open logic project.
The textbook vendors are desperately working to create lock-in and exploit principal-agent misalignment but long term it's a losing battle.
This is an interesting post and Wood from Eden is in general a great substack.