Analyses should focus on the digital economy, which is doing most of the automation. Its main impact can be see in the stats: inequality. And it will keep going up.
Perhaps in the rise of the USA populists and violence we are already seeing the beginning of the implosion of society due to the rising inequality and the song of angry men.
I'm 68 and a geek grad from 3 letter school. On the DARPANET in 72. Seen all the tech evolution from mainframes to DEC to PCs to laptops to phones. Saw how PC based solid modeling like Solid Works replaced drafters in automotive as engineers did more and most of the whole design and analysis of parts and subsystems.
I love Noah Smith.
My observation as an active participant over 50s years in this topic follows.
A. Yes, automation eliminates jobs. Drafters, admin/secretaries, travel agents.
B. Totally agree that the variable cost of IT amd automation is incredibly low
C. Physical products still require component manufacturing, tooling, assembly and test.
D. Unfortunately the Apple et al determined that China wae an easier route than US labor (it's competitive here in Arizona, plus automation) would probably raise an iPhone variable cost by one dollar. Or 2. Meaningless actually
E. The long arc of allnof this is interesting to me:
Conclusion: Not enough work is very very bad. Coupled with today's lower work ethic in under 35 yr olds, it's gonna be a huge problem.
Meanwhile labor has 2 core issues.
1. Struggle forna living wage
2. Struggle for Healthcare payments
Solutions:
A. National Healthcare. Saves $2.5 Trillion annually of non value. Move from 20% GDP back to 10% GDP.
B. This allows worker mobility to optimize jobs and equally allows startups to hire a BigCorp talent afraid to not have Healthcare.
C. Mandatory lower working hours. Overtime is illegal. Minimum wage $20/hr.
Productivity to infinity is 1 person making 500 million cell phones in a Foxconn factory. With millions unable to be supported.
On the manufacturing cost of an iPhone topic. https://www.cnet.com/culture/iphone-manufacturing-costs-revealed is 2012 data. This would agree with you on higher costs to make in the US. I mean yes it would be higher cost, but I believe today the wage gap has closed between China and say Arizona labor. I led 2 technical manufacturing businesses here in Arizona 2007 to 2018 and our direct labor costs were pretty competitive. Albeit low technical volumes and not automotive/Apple scale.
Except Foxconn has some other advantages most don't know. I visited Foxconn with some major customers in the 2005 6 period. Terry Gau, the owner, has his own personal magnesium mine. Used for die casting ultra light laptop cases. Back then, this factory made 300 million circa 2006 cell phones a year. They leap ahead of western companies because they buy state of the art machine to make and design tools and tooling. In days weeks. The LCD display line was making 100s of millions.
This looks at worker engagement by age group. It's doesn't support my personal observations which of course are small sample and biased. The suevey looks at my Boomer cohort today though, not how we were in the 70s and 80s
But briefly for your consideration. Think of M4A as releasing and optimizing the talents of workers across the US. No longer tied to BigCorp for its Healthcare, now freedom to move away from management and culture that doesn't work, to pursue your dream job, to go to any startup. Startups don't often have Healthcare so they have a suboptimal hiring situation. Mobility to start your own company. To move near family or friends or the location that fits.
I think M4A would release and reorganize talent with a huge net win for business and workers.
The work ethic comes from my own employee observations. Especially bachelor vs PhD. At least in hard products. Second is echoing Noah's comments about TSMC is finding or not finding here in Arizona.
Agree on M4A, but as I mentioned, it's independent from Automation. It would help it, but not more than any other standard industry. It was true 5 years ago and 10 years ago.
I believe your hypothesis re young ppl, but the massive confounder there is you've been aging! And anecdotal evidence. Time and again, I've seen these reasonable-seeming hypotheses fall under the weight of proper data
A. National Healthcare will destroy millions of jobs in the private health industry, isn’t your view and the authors is that saving any middle skill job the most important thing?
B. You may have heard of ObamaCare, right? Anyone that wants to go from big corp to a small startup can buy ACA coverage so that is not a barrier anymore.
C. Mandatory restriction on how many hours someone can work is a great threat to individual rights. This harms the ambitious low income worker’s ability to get ahead. I’m also a geek but graduated from a for-profit college that XIT grads would call a joke, but I worked many 90 hour weeks (salary, no overtime) at startups without health insurance or ACA and this made me successful and wealthy so I don’t think any government should be preventing others from working for a good future.
I think he means mainly as a cost reduction. This is true. A free single payer option would give a free benchmark for many healthcare problems, bringing the private price down. Also, there would be things like single buyer negotiation, which is illegal today in the US AFAIK. Just comparing costs and life expectancy in the US vs national healthcare countries strongly suggests this would happen
I don't think at all that saving jobs is the most important thing. I'm concerned about the social, political, and humanitarian ramifications of doing it too fast.
A. Won’t automation destroy the same jobs. Ultimately only 2 or 3 healthcare systems will emerge just as all the other industries Then they might decide to specialize as in any other industry Only one company builds the B2 air craft When I was in practice 20 - 25 percent of patients had to switch insurance due to change in jobs or contracts quarterly with lots of unnecessary paperwork
With mandatory sharing of medical records and transparent billing change will occur Same as government care but only with a few more middle elite autocrats and less competition
You seem to be framing the issue as if inequality in income earned before taxes is intrinsically a bad thing. An income or wage is a price, and prices within a market system serve the roles of a signal and an incentive. In this case, the market is broadcasting for all who are willing to listen that the big opportunities for creating value for fellow humans is in higher skilled jobs in technology and capital investment in these industries. This isn’t just a good thing, it is a necessary thing for a functioning complex adaptive learning system such as decentralized markets.
I don't think at all that inequality of income earned before taxes is intrinsically bad. I think it's crucial to capitalism, which is the biggest engine of growth humanity has ever seen.
I do think the inequality of disposable income has probably an optimal equilibrium. Too little, and there's no incentive to work. Too much, and there's also less of an incentive, plus inequality brings unhappiness and social conflict.
I am concerned about how much control we will have on the inequality of disposable income.
It's the inequality in society that's a bad thing. One thing you are ignoring in your stylized world that the author does not is power. Inequality in society isn't a problem unless rich people try to entrench their power in an oligarchy (keeping everyone else down), which they are incented by self-interest to do.
And the poor and middle class, who vastly outnumber and outvote them, are incentivized to take from the rich. Just saying…
I am not arguing for rent seeking or keeping others down, but there is a big leap from arguing against exploitation and arguing against inequality. Not the same things.
Side note, if we look at trends in the US, the tax/redistribution system is getting increasingly progressive, and technology employees and the rich are overwhelmingly aligned with the left which is fighting for higher taxes and progressivity, not lower.
You are avoiding or missing the point I made in my initial comment above on market signals and incentives.
It is a necessary feature of markets to signal better returns to education and investment in more productive fields, not a bug. Inequality in market outcomes is not necessarily a reflection of a problem in a properly functioning economy, in fact it is absolutely necessary.
Your "and yet" comment was aimed at my argument that redistribution and tax progressivity have increased, not decreased, but you missed the point and switched back to pre-transfer levels. And no, after taxes and transfers, inequality has not increased in the US.
Finally, you are concerned that communities which depend upon skills no longer in demand have been devastated. Yes, that is how markets work. When the gold dries up, mining towns turn to ghost towns.
That’s for income inequality. Wealth inequality is much higher.
I understood your point but might not have expressed myself correctly, or I might have been interested in another aspect of your point.
Inequality is crucial for productivity. But inequality doesn’t need to be linear because dollar utility lowers logarithmically
If you make $10k a year, an additional $1k is a lot, and you’re going to work very hard to get it.
If you make $1M a year, an additional $1k is nothing.
But also, I believe an additional $100k means much less to this person than the $1k to the poorer one.
This is why progressive taxation works.
When we’re deciding income tax brackets, we’re implicitly guessing the average income elasticity at different levels. In other words, maybe 40% is the optimal top income marginal tax rate. Maybe it’s 60%. Maybe it’s 80%. The right level depends on how much less incentive there is for these people to work vs the social benefit of the redistribution.
AI will usher in an era with a rate or pace of disruption that will be difficult to comprehend. On another note, Tomas, are you going to invest in Substack? I will.
So where do we go now? If we want a business that will grow, but we can't (don't want to) join the digital/automation revolution, who will be our customers?
Instinct tells me that the future lies in providing for and caring for people. That means creating a desirable environment, and enlisting carers who are focussed on the human experience (this includes cooks, nurses, masseurs,... OMG, even sex workers?). The market is the section of humanity that is excessively stressed by modern life, those who need to regain control (this includes diabetics, the obese, drug users...) but who also have enough money to pay for a "reset".
Theorists say that if AI takes over then there are 2 options: (a) It wipes us out because we are a virus and unnecessary, (b) It creates a utopia for its creators, but keeps us like pets.
For option (a) I should follow my instincts in the sure knowledge that one day nothing will matter.
For option (b) I should follow my instincts in the sure knowledge that I will be one of the providers of Utopia.
I agree with you. The professions that require physical and human interaction and are generalist and deal with idiosyncratic situations (such as care professions like nursing but also bar tending), but especially those ones that require an education and license (nursing and many other allied health professions) will be the last to be replaced by AI and robotics.
It would seem that AI is much like an iceberg to the Titanic. What we can currently perceive "above the surface" falls short of appreciating it's full impact to a system/structure designed to withstand single compartment damage but not injury across several at once. The implementation of "trickle down economics" in the 80s and its continuance to date despite its severe negative consequences to the majority of workers in this nation was possible because it was and continues to be successfully "sold" to those it damages because tax policies and Supreme Court decisions have allowed the undermining of one person one vote by allowing the windfalls accumulated by the rich to be used to market the continuance of that agenda. It seems the assumption by the author is that AI, like an iceberg, is a force/entity that cannot be managed or controlled in its implementation and therefore its impact on our society, and that, as a consequence, and like the Titanic, we can only hope for survivors. Given the inertia and indifference of Congress, populated as it is by the wealthy, such an assumption is probably only realistic. One should not ignore, however, that the government could play a role in throttling the pace of change since it would be in its interest and that of the majority of people it purportedly represents to reasonably apply controls to the implementation of this technology, allowing for the survival of the nation as a democracy and permitting a greater percentage of its citizens a means to support themselves in the face of such anticipated change. The Titanic lies on the bottom of the Atlantic in part because it was operated under the premise that it was designed not to sink. Post mortem analysis demonstrated that it did not have to happen that way. I would like to think that the significant and anticipated disruptions to society AI is expected to bring does not mean that we have to be dragged behind it helplessly but might instead temper its pace to better accommodate our needs for purpose and meaning and allow more time for transition.
Is it possible that men are dropping out of salaried jobs just because they can? In decades past, the man had to work. But if social benefits and incomes are now sufficient to allow an "adequate" lifestyle without working, then why not choose an easier option?
I live in Vietnam, and as an investor I am effectively off-grid. I have disappeared from the economy where I worked for 40 years.
This covers lots of the details of the automation issue, but I don’t agree that the main worry should be preserving existing jobs. Labor is not a thing to be cherished, but a means to an end which should be eliminated as much as possible. I believe LLMs like ChatGPT will reduce the time knowledge workers spend working on the products they make, so they should get more pay and/or the products will get less expensive.
Very cool Tomas. The detail of the data does give a different picture that seems to be more coherent with general public opinion about automation.
Would be curious if we can assign different shares of the inequality growth in advanced economies to automation vs globalisation, lower top marginal tax rate, near-end of organised labour... There are several contenders there that seem to all act starting in the mid-70s. And they would require different policy response.
I saw this study in the 2000s that suggested until then it was either 75% automation and 25% globalization, or the other way around. Can't remember. My guess is the globalization effect has been mostly tapped out by now (and it was an internal inequality driver, but a global equality driver), and most of the remaining inequality generation will come from automation (especially AI)
Certainly tapped out when it comes to blue collar jobs, at least until Africa can start competing for those.
White collar jobs are now being outsourced to India, though I expect wages there to rise rapidly enough so that within 10-20 years, shipping white collar jobs to India doesn't make much sense any more.
AI will definitely be a big inequality driver that countries will have to deal with.
I find the statistics on males dropping out of the workforce very interesting.
Have you cross-checked this with the numbers of stay-at-home fathers? It may be nothing given declining fertility rates.
With more women entering the workforce, we'd obviously expect an increase in child care workers (low productivity, low pay) though most seem to be female.
Another possibility is increase in incarceration rates. This would disproportionately affect young males. Plus criminals on the outside don't earn wages.
Other thoughts I had are (1) cash economy and (2) crime. They have to be getting money somehow if they're not incarcerated or on benefits. A very intriguing phenomenon.
Blimey. Possibly the record for the longest (and thought-provoking) comment thread/most comments on your articles Tomas?!
And that sentence where you linked to a dozen different articles in each word (bravo) - it overwhelmed me and thus cured my weakness for link rabbit holes. :D
My take aways - inequality, speed of change (related) and universal healthcare... (thinking emoticon)
And I'm sure you and most others have seen the Boston Dynamics robots - be afraid, be very afraid (I thought of it when looking at the graph in your previous articles (cognitive, non-routine, physical etc).
STOP - DO NOT INVEST IN SUBSTACK'S BROKEN MODEL!!!
Seriously if they are begging for money and VC bailed it means its a broken model. Think about it, how many regular people are paying per writer $5 to $10 per month? Not enough apparently and the next thing they will introduce is advertising which is 1million times worse. This monetization model is for elites, elite writers and elite readers who can afford to pay to "benefit" from their writing. Its not Twitter but its just as elitist as Twitter and will devolve into the same mess and control mechanism.
Go check out web3. Go check out crypto. Go check out the MVP of my solo hobbiest project "dplatform.me" The next platforms will be web3, crypto, micro-transactions, and governed by decentralized autonomous organizations (DAOs) or no one
If there is only one optimal solution for each problem and A I can find it. Like the traveling salesman The what does that mean for the end of competition for a better product and capitalism.
I doubt there is only 1 optimal solution for each problem. For one, many people disagree about what is optimal. What is an optimal/ideal society to a conservative patriarchal Christian very likely is not at all the optimal society to a left-wing feminist.
An excellent economics article without needing to use the word wealth. Attention is given not just to the creation of valuable things, but also to the loss of value that is caused in other areas. So instead of mentioning the somewhat magical concept of “wealth creation” that techno-fools and capitalist-fools like to highlight, there is also the implied mention of “wealth destruction” and the clear LINK between the two is made. I would argue that this link is consistent with a flow model of wealth where the total amount of wealth is in fact conserved. Using the concept of wealth inequality as you do in the comments section is still useful and consistent with the flow model.
You also highlight one of the biggest problems with GDP statistics: they only measure things by the amount that people PAY for them (legally!). This means that many valuable activities are underweighted or just ignored entirely. The generally accepted economic model of reality is bedevilled by definitional problems, measurement issues and an elastic measuring stick. If societies make decisions based solely on this model the margin for error is very large.
Tech is indeed highly deflationary as is offshoring production to low income countries. So why haven’t we been experiencing a deflationary boom? The answer to that question goes to the heart of any discussion about reshaping our economic system.
Perhaps in the rise of the USA populists and violence we are already seeing the beginning of the implosion of society due to the rising inequality and the song of angry men.
When I gave this as a presentation a few years ago, I led with that hypothesis
I'm 68 and a geek grad from 3 letter school. On the DARPANET in 72. Seen all the tech evolution from mainframes to DEC to PCs to laptops to phones. Saw how PC based solid modeling like Solid Works replaced drafters in automotive as engineers did more and most of the whole design and analysis of parts and subsystems.
I love Noah Smith.
My observation as an active participant over 50s years in this topic follows.
A. Yes, automation eliminates jobs. Drafters, admin/secretaries, travel agents.
B. Totally agree that the variable cost of IT amd automation is incredibly low
C. Physical products still require component manufacturing, tooling, assembly and test.
D. Unfortunately the Apple et al determined that China wae an easier route than US labor (it's competitive here in Arizona, plus automation) would probably raise an iPhone variable cost by one dollar. Or 2. Meaningless actually
E. The long arc of allnof this is interesting to me:
Conclusion: Not enough work is very very bad. Coupled with today's lower work ethic in under 35 yr olds, it's gonna be a huge problem.
Meanwhile labor has 2 core issues.
1. Struggle forna living wage
2. Struggle for Healthcare payments
Solutions:
A. National Healthcare. Saves $2.5 Trillion annually of non value. Move from 20% GDP back to 10% GDP.
B. This allows worker mobility to optimize jobs and equally allows startups to hire a BigCorp talent afraid to not have Healthcare.
C. Mandatory lower working hours. Overtime is illegal. Minimum wage $20/hr.
Productivity to infinity is 1 person making 500 million cell phones in a Foxconn factory. With millions unable to be supported.
Sustainability requires deliberate UN-Productivity.
Thank you, thought-provoking!
Here are some provoked thoughts:
• The cost of manufacturing the iPhone in the US would be way, way higher
• I am not convinced about the work ethic pbm of under 35. Would love to see proper data on this, and neither perception nor correlation
• The healthcare option is a good one, albeit broadly disconnected from automation
• Not sure I favor limiting freedom. I'd rather do redistribution
On the manufacturing cost of an iPhone topic. https://www.cnet.com/culture/iphone-manufacturing-costs-revealed is 2012 data. This would agree with you on higher costs to make in the US. I mean yes it would be higher cost, but I believe today the wage gap has closed between China and say Arizona labor. I led 2 technical manufacturing businesses here in Arizona 2007 to 2018 and our direct labor costs were pretty competitive. Albeit low technical volumes and not automotive/Apple scale.
I'll look for more current data.
https://www.investopedia.com/financial-edge/0912/the-cost-of-making-an-iphone.aspx a more recent analysis shows Foxconn labor costs over $12/hr and up. So there is a reasonable case to make these kn the US.
Except Foxconn has some other advantages most don't know. I visited Foxconn with some major customers in the 2005 6 period. Terry Gau, the owner, has his own personal magnesium mine. Used for die casting ultra light laptop cases. Back then, this factory made 300 million circa 2006 cell phones a year. They leap ahead of western companies because they buy state of the art machine to make and design tools and tooling. In days weeks. The LCD display line was making 100s of millions.
https://www.gallup.com/workplace/404693/generation-disconnected-data-gen-workplace.aspx
This looks at worker engagement by age group. It's doesn't support my personal observations which of course are small sample and biased. The suevey looks at my Boomer cohort today though, not how we were in the 70s and 80s
Correct, not only is this not supporting the evidence, it does so with the massive age confounder.
I think the hypothesis is refuted
Agree!
My pleasure Tomas!
I have some answers for these later this evening.
But briefly for your consideration. Think of M4A as releasing and optimizing the talents of workers across the US. No longer tied to BigCorp for its Healthcare, now freedom to move away from management and culture that doesn't work, to pursue your dream job, to go to any startup. Startups don't often have Healthcare so they have a suboptimal hiring situation. Mobility to start your own company. To move near family or friends or the location that fits.
I think M4A would release and reorganize talent with a huge net win for business and workers.
The work ethic comes from my own employee observations. Especially bachelor vs PhD. At least in hard products. Second is echoing Noah's comments about TSMC is finding or not finding here in Arizona.
Agree on M4A, but as I mentioned, it's independent from Automation. It would help it, but not more than any other standard industry. It was true 5 years ago and 10 years ago.
I believe your hypothesis re young ppl, but the massive confounder there is you've been aging! And anecdotal evidence. Time and again, I've seen these reasonable-seeming hypotheses fall under the weight of proper data
We just need a future where robots can produce food, shelter, and insurance at minimal cost so that basic needs can be met easily.
Then UBI for everyone to ensure that basic needs are at least met.
And we need to provide enough education for people to use their free time
Your 3 solutions are questionable.
A. National Healthcare will destroy millions of jobs in the private health industry, isn’t your view and the authors is that saving any middle skill job the most important thing?
B. You may have heard of ObamaCare, right? Anyone that wants to go from big corp to a small startup can buy ACA coverage so that is not a barrier anymore.
C. Mandatory restriction on how many hours someone can work is a great threat to individual rights. This harms the ambitious low income worker’s ability to get ahead. I’m also a geek but graduated from a for-profit college that XIT grads would call a joke, but I worked many 90 hour weeks (salary, no overtime) at startups without health insurance or ACA and this made me successful and wealthy so I don’t think any government should be preventing others from working for a good future.
I think he means mainly as a cost reduction. This is true. A free single payer option would give a free benchmark for many healthcare problems, bringing the private price down. Also, there would be things like single buyer negotiation, which is illegal today in the US AFAIK. Just comparing costs and life expectancy in the US vs national healthcare countries strongly suggests this would happen
I don't think at all that saving jobs is the most important thing. I'm concerned about the social, political, and humanitarian ramifications of doing it too fast.
Agreed on freedom. We shouldn't curtail it.
A. Won’t automation destroy the same jobs. Ultimately only 2 or 3 healthcare systems will emerge just as all the other industries Then they might decide to specialize as in any other industry Only one company builds the B2 air craft When I was in practice 20 - 25 percent of patients had to switch insurance due to change in jobs or contracts quarterly with lots of unnecessary paperwork
With mandatory sharing of medical records and transparent billing change will occur Same as government care but only with a few more middle elite autocrats and less competition
You seem to be framing the issue as if inequality in income earned before taxes is intrinsically a bad thing. An income or wage is a price, and prices within a market system serve the roles of a signal and an incentive. In this case, the market is broadcasting for all who are willing to listen that the big opportunities for creating value for fellow humans is in higher skilled jobs in technology and capital investment in these industries. This isn’t just a good thing, it is a necessary thing for a functioning complex adaptive learning system such as decentralized markets.
I don't think at all that inequality of income earned before taxes is intrinsically bad. I think it's crucial to capitalism, which is the biggest engine of growth humanity has ever seen.
I do think the inequality of disposable income has probably an optimal equilibrium. Too little, and there's no incentive to work. Too much, and there's also less of an incentive, plus inequality brings unhappiness and social conflict.
I am concerned about how much control we will have on the inequality of disposable income.
It's the inequality in society that's a bad thing. One thing you are ignoring in your stylized world that the author does not is power. Inequality in society isn't a problem unless rich people try to entrench their power in an oligarchy (keeping everyone else down), which they are incented by self-interest to do.
And the poor and middle class, who vastly outnumber and outvote them, are incentivized to take from the rich. Just saying…
I am not arguing for rent seeking or keeping others down, but there is a big leap from arguing against exploitation and arguing against inequality. Not the same things.
Side note, if we look at trends in the US, the tax/redistribution system is getting increasingly progressive, and technology employees and the rich are overwhelmingly aligned with the left which is fighting for higher taxes and progressivity, not lower.
And yet, inequality has only increased in the US over time while communities that depended on non-high-skill work have been devastated.
There's this interesting book that claims there has never been an inequality reduction outside of wars, plagues, revolutions, or state failure
Counting the Cold War? Because inequality in the US certainly went down between WWII and the '70's.
From a wealth standpoint, AFAIK it didn't.
The drops in inequality were WWI and WWII.
The high taxation levels of the 50s thrugh 70s *kept inequality low*, but didn't reduce it.
Then it explodes in the 80s.
You are avoiding or missing the point I made in my initial comment above on market signals and incentives.
It is a necessary feature of markets to signal better returns to education and investment in more productive fields, not a bug. Inequality in market outcomes is not necessarily a reflection of a problem in a properly functioning economy, in fact it is absolutely necessary.
Your "and yet" comment was aimed at my argument that redistribution and tax progressivity have increased, not decreased, but you missed the point and switched back to pre-transfer levels. And no, after taxes and transfers, inequality has not increased in the US.
Finally, you are concerned that communities which depend upon skills no longer in demand have been devastated. Yes, that is how markets work. When the gold dries up, mining towns turn to ghost towns.
This is on inequality in the US
https://ourworldindata.org/grapher/inequality-before-and-after-taxes-and-transfers-Thewissen-et-al-data?tab=chart
That’s for income inequality. Wealth inequality is much higher.
I understood your point but might not have expressed myself correctly, or I might have been interested in another aspect of your point.
Inequality is crucial for productivity. But inequality doesn’t need to be linear because dollar utility lowers logarithmically
If you make $10k a year, an additional $1k is a lot, and you’re going to work very hard to get it.
If you make $1M a year, an additional $1k is nothing.
But also, I believe an additional $100k means much less to this person than the $1k to the poorer one.
This is why progressive taxation works.
When we’re deciding income tax brackets, we’re implicitly guessing the average income elasticity at different levels. In other words, maybe 40% is the optimal top income marginal tax rate. Maybe it’s 60%. Maybe it’s 80%. The right level depends on how much less incentive there is for these people to work vs the social benefit of the redistribution.
Thanks, excellent response.
AI will usher in an era with a rate or pace of disruption that will be difficult to comprehend. On another note, Tomas, are you going to invest in Substack? I will.
Probably
So where do we go now? If we want a business that will grow, but we can't (don't want to) join the digital/automation revolution, who will be our customers?
Instinct tells me that the future lies in providing for and caring for people. That means creating a desirable environment, and enlisting carers who are focussed on the human experience (this includes cooks, nurses, masseurs,... OMG, even sex workers?). The market is the section of humanity that is excessively stressed by modern life, those who need to regain control (this includes diabetics, the obese, drug users...) but who also have enough money to pay for a "reset".
Theorists say that if AI takes over then there are 2 options: (a) It wipes us out because we are a virus and unnecessary, (b) It creates a utopia for its creators, but keeps us like pets.
For option (a) I should follow my instincts in the sure knowledge that one day nothing will matter.
For option (b) I should follow my instincts in the sure knowledge that I will be one of the providers of Utopia.
My premium article next week covers this!
I agree with you. The professions that require physical and human interaction and are generalist and deal with idiosyncratic situations (such as care professions like nursing but also bar tending), but especially those ones that require an education and license (nursing and many other allied health professions) will be the last to be replaced by AI and robotics.
Would I want a robot to care for me in a hospital bed? Would I want to tells stories with a robot bar tender?
I think the truly human interactions cannot be satisfactory unless you can believe that the robot is human. (See TV series called "Humans".)
You agree with me!
It would seem that AI is much like an iceberg to the Titanic. What we can currently perceive "above the surface" falls short of appreciating it's full impact to a system/structure designed to withstand single compartment damage but not injury across several at once. The implementation of "trickle down economics" in the 80s and its continuance to date despite its severe negative consequences to the majority of workers in this nation was possible because it was and continues to be successfully "sold" to those it damages because tax policies and Supreme Court decisions have allowed the undermining of one person one vote by allowing the windfalls accumulated by the rich to be used to market the continuance of that agenda. It seems the assumption by the author is that AI, like an iceberg, is a force/entity that cannot be managed or controlled in its implementation and therefore its impact on our society, and that, as a consequence, and like the Titanic, we can only hope for survivors. Given the inertia and indifference of Congress, populated as it is by the wealthy, such an assumption is probably only realistic. One should not ignore, however, that the government could play a role in throttling the pace of change since it would be in its interest and that of the majority of people it purportedly represents to reasonably apply controls to the implementation of this technology, allowing for the survival of the nation as a democracy and permitting a greater percentage of its citizens a means to support themselves in the face of such anticipated change. The Titanic lies on the bottom of the Atlantic in part because it was operated under the premise that it was designed not to sink. Post mortem analysis demonstrated that it did not have to happen that way. I would like to think that the significant and anticipated disruptions to society AI is expected to bring does not mean that we have to be dragged behind it helplessly but might instead temper its pace to better accommodate our needs for purpose and meaning and allow more time for transition.
Is it possible that men are dropping out of salaried jobs just because they can? In decades past, the man had to work. But if social benefits and incomes are now sufficient to allow an "adequate" lifestyle without working, then why not choose an easier option?
I live in Vietnam, and as an investor I am effectively off-grid. I have disappeared from the economy where I worked for 40 years.
That's a reasonable hypothesis. Would love to see it confirmed!
This covers lots of the details of the automation issue, but I don’t agree that the main worry should be preserving existing jobs. Labor is not a thing to be cherished, but a means to an end which should be eliminated as much as possible. I believe LLMs like ChatGPT will reduce the time knowledge workers spend working on the products they make, so they should get more pay and/or the products will get less expensive.
I think this is good and unstoppable.
But if it destroys many more jobs than it creates, and does it fast, we're going to have a society-wide issue. That's what Richard is referring to
Oh, and to your point: I agree. My main worry about AI isn't jobs but destroying democracy and society.
Well, tech is deflationary, so it would likely mean more output with some workers getting far more pay and many far less.
Very cool Tomas. The detail of the data does give a different picture that seems to be more coherent with general public opinion about automation.
Would be curious if we can assign different shares of the inequality growth in advanced economies to automation vs globalisation, lower top marginal tax rate, near-end of organised labour... There are several contenders there that seem to all act starting in the mid-70s. And they would require different policy response.
I saw this study in the 2000s that suggested until then it was either 75% automation and 25% globalization, or the other way around. Can't remember. My guess is the globalization effect has been mostly tapped out by now (and it was an internal inequality driver, but a global equality driver), and most of the remaining inequality generation will come from automation (especially AI)
Certainly tapped out when it comes to blue collar jobs, at least until Africa can start competing for those.
White collar jobs are now being outsourced to India, though I expect wages there to rise rapidly enough so that within 10-20 years, shipping white collar jobs to India doesn't make much sense any more.
AI will definitely be a big inequality driver that countries will have to deal with.
I find the statistics on males dropping out of the workforce very interesting.
Have you cross-checked this with the numbers of stay-at-home fathers? It may be nothing given declining fertility rates.
With more women entering the workforce, we'd obviously expect an increase in child care workers (low productivity, low pay) though most seem to be female.
Another possibility is increase in incarceration rates. This would disproportionately affect young males. Plus criminals on the outside don't earn wages.
Brilliant piece btw.
I haven’t. Off the top of my mind, neither of these numbers are growing dramatically.
Stay-at-home dads have 2xed but there’s still just only 2M of them. I think there are fewer Incarcerated men than there used to
Other thoughts I had are (1) cash economy and (2) crime. They have to be getting money somehow if they're not incarcerated or on benefits. A very intriguing phenomenon.
Blimey. Possibly the record for the longest (and thought-provoking) comment thread/most comments on your articles Tomas?!
And that sentence where you linked to a dozen different articles in each word (bravo) - it overwhelmed me and thus cured my weakness for link rabbit holes. :D
My take aways - inequality, speed of change (related) and universal healthcare... (thinking emoticon)
And I'm sure you and most others have seen the Boston Dynamics robots - be afraid, be very afraid (I thought of it when looking at the graph in your previous articles (cognitive, non-routine, physical etc).
https://youtu.be/-e1_QhJ1EhQ - I promise I'm not a bot posting this (yet)...
You should see the threads on COVID stuff, especially the controversial opinions...
really good one!! worth to read it thoroughly!! :)
it's always a good idea to align on concepts and data before starting the debate on core topics
STOP - DO NOT INVEST IN SUBSTACK'S BROKEN MODEL!!!
Seriously if they are begging for money and VC bailed it means its a broken model. Think about it, how many regular people are paying per writer $5 to $10 per month? Not enough apparently and the next thing they will introduce is advertising which is 1million times worse. This monetization model is for elites, elite writers and elite readers who can afford to pay to "benefit" from their writing. Its not Twitter but its just as elitist as Twitter and will devolve into the same mess and control mechanism.
Go check out web3. Go check out crypto. Go check out the MVP of my solo hobbiest project "dplatform.me" The next platforms will be web3, crypto, micro-transactions, and governed by decentralized autonomous organizations (DAOs) or no one
The future is decentralized!
If there is only one optimal solution for each problem and A I can find it. Like the traveling salesman The what does that mean for the end of competition for a better product and capitalism.
I doubt there is only 1 optimal solution for each problem. For one, many people disagree about what is optimal. What is an optimal/ideal society to a conservative patriarchal Christian very likely is not at all the optimal society to a left-wing feminist.
That makes sense Richard
An excellent economics article without needing to use the word wealth. Attention is given not just to the creation of valuable things, but also to the loss of value that is caused in other areas. So instead of mentioning the somewhat magical concept of “wealth creation” that techno-fools and capitalist-fools like to highlight, there is also the implied mention of “wealth destruction” and the clear LINK between the two is made. I would argue that this link is consistent with a flow model of wealth where the total amount of wealth is in fact conserved. Using the concept of wealth inequality as you do in the comments section is still useful and consistent with the flow model.
You also highlight one of the biggest problems with GDP statistics: they only measure things by the amount that people PAY for them (legally!). This means that many valuable activities are underweighted or just ignored entirely. The generally accepted economic model of reality is bedevilled by definitional problems, measurement issues and an elastic measuring stick. If societies make decisions based solely on this model the margin for error is very large.
Tech is indeed highly deflationary as is offshoring production to low income countries. So why haven’t we been experiencing a deflationary boom? The answer to that question goes to the heart of any discussion about reshaping our economic system.
I would appreciate it if my name were not placed in front of fictitious quotes. :-)
Oops, I had sent you the drafts, I assumed we were aligned.
I deleted all mentions. Sorry for that!