As every dad who insists on making good time by getting the family started on a road trip by leaving the house at 4am knows, the obvious potential huge gain for logistics is getting around the bottleneck of bad road congestion and vastly improving the infrastructural-capital utilization efficiency by using those roads at off-peak times of minimum demand, usually in the middle of the night, when most human-transport competition is off the road. And sure, there's a bit of this happening. However, to date, this strategy has proven infeasible or uneconomical enough that it still not nearly as widely employed as it could be. Mostly this seems to be due to human workday coordination (even the Soviets found they couldn't do very much about that to spread labor activity and and availability around) and the risks and other inefficiencies caused by impairment of vision by lack of solar illumination, and it being really expensive (and perhaps illegal) to equip humans with the best low-illumination visual augmentation tech. Limited congestion pricing also doesn't encourage nearly as efficient a utilization of road infrastructure as could be achieved with better tolling.
But now, with robot drivers with high resolution maps and all kinds of sensors across multiple spectrums that can operate nearly as well at night, the former human and tech roadblocks to shifting logistics to night may be significantly mitigated or, maybe soon, simply neutralized. That opens up tons of cheap, spare capacity on existing infrastructure and without the need for substantial additional investments in new roads beyond some extra maintenance and repair to handle the new load.
So, the business case seems solid, and it could be a huge deal.
While I think Vernon is somewhat on track with efficiency opportunities, he is wrong with cost assumptions. I'd like to see him show is work. For one, even though cost components are similar, the cost structures are quite different for LTL (less than truckload) versus TL (truckload). Also, neither uses ton-miles for costing, which is normally used for bulk truck or rail freight. The lack of operational understanding is apparent throughout his article. Again, I applaud the enthusiasm but a dose of reality needs to be applied to correct some of the very optimistic assumptions.
"The first-order answer for cost reductions is clear. Remove the driver with software and reduce fuel costs with batteries."
Austin Vernon apparently knows how an entire industry should be fixed. While he's at it maybe he can fix how we buy and store groceries and prepare food and how we build and clean our homes and toilets and, well, you get the idea. Everything in life has inefficiencies and there are experts willing to tell us how things really should be.
sigh
I remember having conversations with technology providers in the early 2000s about their frustration with trucking companies not buying their driver management systems. The complaint being that the companies were too cheap and undercapitalized and didn't want to spend the money.
Well yeah. Trucking is a highly competitive, low margin business. Is it inefficient? Maybe. But what if that inefficiency is because it is so relatively cheap to put a truck in service? Increasing the cost of truck operation might actually improve efficiency, but this would require shifting the industry from a low cap to high cap business (ie make it a technology business instead of a human services business).
So if the trucking industry was wiped out - both trucks and drivers - and replaced with million dollar autonomous vehicles and ecosystem to support those vehicles then VOILA! the inefficiency of the industry could be fixed. Except inefficiencies would still exist because there is no perfect order in the human ecosystem. Which means the ultimate answer is to get rid of the humans everywhere.
"Inefficiency" isn't the right word. In general it's just more "automation" with innovation pushing out the production possibilities frontier of what is both possible to do at all, and also rapidly and significantly lowering the cost of adequate capital substitutes for labor such that it is now economically competitive with labor.
If you keep using expensive labor to do the exact same thing you could be doing with cheaper capital -then- one is talking about a problem with the 'inefficient' use of available resources. There are a number of professions for which this is probably already the case, and without sustained rent-seeking (and I expect a whole lot of that), it seems inevitable in due time that these services will come to be provided mostly by machines and not by humans.
So, applied to this case, until very recently, it just wasn't possible at any price to automate trucking. Now it's possible, though perhaps still more expensive and not quite yet a perfectly adequate substitute. But it seems that very soon it will shoot past adequacy to be genuinely superior to any human driver, and also become considerably cheaper than the total cost involved in employing a driver.
One special factor to keep in mind with regard to these new capabilities is that they are based more on bits than atoms, and, unlike skilled and experienced humans, bits can be scaled indefinitely, instantly, and at zero marginal cost. If it takes a million dollars to raise and train a good professional, then even if it costs a billion to develop a substitute machine, you get the 1001st to the billionth additional substitute for peanuts. If you can make your money back at all, you can and will simply go all the way and eat the whole sector.
I appreciate your thoughts. What I see with "Big Tech" are a blend of strategies all with the intent of grabbing market share and claiming consumer surplus for itself.
There is the classic software approach and we can agree that software makers deserve a reward for making customer work easier. The problem here is the software makers resort to anti-competitive methods to deny competition and to grab more consumer surplus for themselves - the "cloud" w/ subscription fees and data lock-in and the DOJ not enforcing antitrust laws is a gift to Big Tech.
Then there is Big Tech using information technology to take control of a market. Uber, Instacart and AirBnb are examples. The actual product being sold is rentals or human services and software simply facilitates the transaction. Of course the goal is to use the software to gain a competitive advantage and then extract greater consumer surplus. But recognize that the actual service offered is not technical at all - Uber is literally some guy driving passengers around in a Ford Focus!
What Big Tech seeks is the Holy Grail of employing software and machines to own and control an industry. They want an Uber where the car is fully autonomous. They want an Instacart where food is automagically delivered to people's homes No "dirty work" of dealing with human employees. But getting to that point logistically is non-trivial. The main issue being the "last-mile".
It makes perfect sense for tractor trailers to drive autonomously and this is being done on Interstates. But the last mile, when the truck has to get off the interstate and deal with the uncertainty of road obstructions remains difficult. The machine logistics are also a great challenge and hand waving that electrification is the answer is, well, hand waving.
What amazes me about the Big Tech visionaries is they are no different than all the Social Planners that have come before. If only society did things "our way" then society would be better. Of course "our way" makes us money or gives the person prestige so maybe we ought to be skeptical of what the visionary is selling.
We ought to welcome life improvements that organically come about because the solution improves lives! But this sale pitch that government must intrude and facilitate massive restructuring of the economy and laws so that life can be improved is dangerous and it doesn't matter if Big Tech is selling it or if a Marxist advocate is selling it.
"What amazes me about the Big Tech visionaries is they are no different than all the Social Planners that have come before."
No. There are a few exceptions, but in general they are very different. The important distinction is that unlike those old social planners, most of the the time the contemporary tech visionaries are calling for private investment and for the government to get or stay out of the way.
As an example, ride-hailing services really have transformed the urban transit experience in the way a tech visionary might have hoped two or three decades ago, but they did it with private money and, if not quite technically 'repeal', then the widespread de facto suspension of enforcement of local laws regulating taxi services.
In contrast, the Social Planners of the past would have insisted that the government enforce its own monopoly of these services, raise taxes to buy all the cabs and employ all the cabbies with group preference quotas and exorbitant pensions, charge different subsidized amounts to a hundred different categories of politically-favors consumers, and also develop the government-uber (Goober?) App, which would be exactly as terrible as every sane person would expect it to be. Uber has been accused of abusing the privileged-accesses of its app to more-or-less spy on its users. That's probably nothing compared to how much Goober would be spying on people.
Yes, private companies will try to do everything they can get away with to capture and maintain as much market share as possible, and yes they will also try everything they can get away with to lobby the government for subsidies and protections and arrangements which allow them to privatize gains but socialize losses. These tactics are worthy of condemnation, but the most intellectually mature approach is to put them in the "don't hate the player, hate the game" category and focus on restraining these corruptions of state action so as to make them unprofitable to attempt in the first place.
Fair rebuttal, except "electrification" is a demand on government to impose a solution and have consumers bear the cost and inconvenience.
And as for autonomous machines, Big Tech will do anything and everything to offload costs onto others. Consider self-driving cars. If I have one, then shouldn't the manufacturer of the car assume liability coverage for any harm and damage the vehicle causes to others when it is in "self-driving" mode? Until the "autonomous vehicle" companies step up and assume that responsibility I am skeptical of their technology.
I don't think it's a fair statement of the current law on the subject to say that autonomous vehicle manufacturers and their insurers aren't held liable for those damages, especially if one can demonstrate malfunction or defect as the proximate cause. Certainly some manufacturers will try to contract away as much of this liability as the law will allow them to and shift the burden to customers (who, however, would receive substantial discounts for doing so) but - and please correct me if I'm wrong - my impression is that their legal capacity to do this is very limited, which is what one might expect in very-highly regulated industries like automobiles.
You are correct. There is a term in the industry about trucking cases where the outcome is called "nuclear verdicts" where there are huge sums awarded. In these cases, the plaintiff attorneys go after everyone connected to the accident, including manufacturers. This will continue.
That battery comment set off my radar. If he thinks EV long haul trucks are the answer to automation and that diesel trucks cannot be automated, I'm reluctant to trust much else that follows.
What happened after 1964 is Civil Rights moved on from "stop government discrimination" to "the government has to actively force people to integrate". The Fair Housing Act passed in 1968. Affirmative Action slowly evolved over this time. School busing also got big during this time period.
MLK was going on about things like Reparations and basically going socialist in the late 60s (everyone forgets he was immensely unpopular at the time of his assassination).
It also expanded beyond black people. Title IX passed in 1972.
But nobody on the center left today would say that all of this stuff was bad.
While it's not necessarily dispositive to the significance of Reagan's elections, it's good to remember that except for six years roughly coinciding with Reagan's two terms when the GOP controlled the Senate by slim margins, the Democrats controlled both the House and Senate from about 1960 to 2000.
On your review: I have been wondering whether our economic problems might be due to not being able to measure what contributes to wealth. The book and your review seems to confirm that. I like your idea that:
"That suggests that we ought to be skeptical of the project of coming up with a single measure of progress. Instead, we can pay attention to a variety of anecdotes and indicators. And then we can expect people to argue over what these observations imply."
What are the efforts in that direction? What should I read (as a non-economist) and where can participate in a debate on this? What are the relevant indicators?
This seems more like an exercise in finding excuses to keep looking to government for answers, and no one will like the answers that result. I realize that trying to get government to butt out is a fool's dream, but it ought to at least be mentioned.
"And in the digital world, where much is provided for free, we do not observe prices. And for many items, prices are artificial: hospital charges, college tuition, and the imputed rent of housing services."
I get the difficulty with digital services but could some estimates be made based on what public companies spend on the basket of services they provide? ... Maybe I didn't get the difficulty. Can quantity be determined for free digital services? What about unlimited subscriptions?
I don't get the problem with hospital and college. They still have prices, right?
Why does rent need to be imputed? Just look at house prices, utilities, property taxes, etc.?
That makes me wonder, how are taxes figured into cost of living and inflation?
GDP seems more accurate than any other single number for comparing last year with prior years, tho requiring some inflation (/value of money) adjustments. Similarly between countries, tho the Big Mac index values are also usable simple for FOREX.
The employment rate is a better metric than the unemployment rate.
The median wage is also good, and changes in this wage level are likely more predictive of electoral support or opposition to the current Party than GDP changes, tho there is a lot of correlation.
(My support for more taxes on the rich is limited such that the 90% and 99% income levels, after taxes & transfers, should go up more slowly than the 50% median. If not, raise more from the rich until the avg worker’s increase, the poor getting richer, is faster than the tax slowed rich getting richer.)
Normal folk want a simple, single, number, and the
In comparing people, there are various metrics, like height, weight, sex, race, age, education, party. Income bracket. Seldom are all these characteristics compared at once, and only for college admissions is SAT score important, but much of high school is oriented at attempting to maximize it.
No one number is a great guide, much less perfect, for rating people or films or companies or countries, but people demand a simple answer, and most all of us want a higher GDP, and a spouse who is a 10.
OMG, I would hate a spouse who is a ten. She's probably got an attitude, and will probably spend the rest of her life desperately trying to keep her looks--a task that is doomed to fail.
I'd start with trying to get more out of rail logistics before fooling with robot drivers. See also Richard Hammond's wreck of a Rimac Concept one from Top Gear.
They say ( whoever they is ) that rail transport is constrained by regulation. I'm sure that's overly simplistic.
As every dad who insists on making good time by getting the family started on a road trip by leaving the house at 4am knows, the obvious potential huge gain for logistics is getting around the bottleneck of bad road congestion and vastly improving the infrastructural-capital utilization efficiency by using those roads at off-peak times of minimum demand, usually in the middle of the night, when most human-transport competition is off the road. And sure, there's a bit of this happening. However, to date, this strategy has proven infeasible or uneconomical enough that it still not nearly as widely employed as it could be. Mostly this seems to be due to human workday coordination (even the Soviets found they couldn't do very much about that to spread labor activity and and availability around) and the risks and other inefficiencies caused by impairment of vision by lack of solar illumination, and it being really expensive (and perhaps illegal) to equip humans with the best low-illumination visual augmentation tech. Limited congestion pricing also doesn't encourage nearly as efficient a utilization of road infrastructure as could be achieved with better tolling.
But now, with robot drivers with high resolution maps and all kinds of sensors across multiple spectrums that can operate nearly as well at night, the former human and tech roadblocks to shifting logistics to night may be significantly mitigated or, maybe soon, simply neutralized. That opens up tons of cheap, spare capacity on existing infrastructure and without the need for substantial additional investments in new roads beyond some extra maintenance and repair to handle the new load.
So, the business case seems solid, and it could be a huge deal.
While I think Vernon is somewhat on track with efficiency opportunities, he is wrong with cost assumptions. I'd like to see him show is work. For one, even though cost components are similar, the cost structures are quite different for LTL (less than truckload) versus TL (truckload). Also, neither uses ton-miles for costing, which is normally used for bulk truck or rail freight. The lack of operational understanding is apparent throughout his article. Again, I applaud the enthusiasm but a dose of reality needs to be applied to correct some of the very optimistic assumptions.
"The first-order answer for cost reductions is clear. Remove the driver with software and reduce fuel costs with batteries."
Austin Vernon apparently knows how an entire industry should be fixed. While he's at it maybe he can fix how we buy and store groceries and prepare food and how we build and clean our homes and toilets and, well, you get the idea. Everything in life has inefficiencies and there are experts willing to tell us how things really should be.
sigh
I remember having conversations with technology providers in the early 2000s about their frustration with trucking companies not buying their driver management systems. The complaint being that the companies were too cheap and undercapitalized and didn't want to spend the money.
Well yeah. Trucking is a highly competitive, low margin business. Is it inefficient? Maybe. But what if that inefficiency is because it is so relatively cheap to put a truck in service? Increasing the cost of truck operation might actually improve efficiency, but this would require shifting the industry from a low cap to high cap business (ie make it a technology business instead of a human services business).
So if the trucking industry was wiped out - both trucks and drivers - and replaced with million dollar autonomous vehicles and ecosystem to support those vehicles then VOILA! the inefficiency of the industry could be fixed. Except inefficiencies would still exist because there is no perfect order in the human ecosystem. Which means the ultimate answer is to get rid of the humans everywhere.
"Inefficiency" isn't the right word. In general it's just more "automation" with innovation pushing out the production possibilities frontier of what is both possible to do at all, and also rapidly and significantly lowering the cost of adequate capital substitutes for labor such that it is now economically competitive with labor.
If you keep using expensive labor to do the exact same thing you could be doing with cheaper capital -then- one is talking about a problem with the 'inefficient' use of available resources. There are a number of professions for which this is probably already the case, and without sustained rent-seeking (and I expect a whole lot of that), it seems inevitable in due time that these services will come to be provided mostly by machines and not by humans.
So, applied to this case, until very recently, it just wasn't possible at any price to automate trucking. Now it's possible, though perhaps still more expensive and not quite yet a perfectly adequate substitute. But it seems that very soon it will shoot past adequacy to be genuinely superior to any human driver, and also become considerably cheaper than the total cost involved in employing a driver.
One special factor to keep in mind with regard to these new capabilities is that they are based more on bits than atoms, and, unlike skilled and experienced humans, bits can be scaled indefinitely, instantly, and at zero marginal cost. If it takes a million dollars to raise and train a good professional, then even if it costs a billion to develop a substitute machine, you get the 1001st to the billionth additional substitute for peanuts. If you can make your money back at all, you can and will simply go all the way and eat the whole sector.
I appreciate your thoughts. What I see with "Big Tech" are a blend of strategies all with the intent of grabbing market share and claiming consumer surplus for itself.
There is the classic software approach and we can agree that software makers deserve a reward for making customer work easier. The problem here is the software makers resort to anti-competitive methods to deny competition and to grab more consumer surplus for themselves - the "cloud" w/ subscription fees and data lock-in and the DOJ not enforcing antitrust laws is a gift to Big Tech.
Then there is Big Tech using information technology to take control of a market. Uber, Instacart and AirBnb are examples. The actual product being sold is rentals or human services and software simply facilitates the transaction. Of course the goal is to use the software to gain a competitive advantage and then extract greater consumer surplus. But recognize that the actual service offered is not technical at all - Uber is literally some guy driving passengers around in a Ford Focus!
What Big Tech seeks is the Holy Grail of employing software and machines to own and control an industry. They want an Uber where the car is fully autonomous. They want an Instacart where food is automagically delivered to people's homes No "dirty work" of dealing with human employees. But getting to that point logistically is non-trivial. The main issue being the "last-mile".
It makes perfect sense for tractor trailers to drive autonomously and this is being done on Interstates. But the last mile, when the truck has to get off the interstate and deal with the uncertainty of road obstructions remains difficult. The machine logistics are also a great challenge and hand waving that electrification is the answer is, well, hand waving.
What amazes me about the Big Tech visionaries is they are no different than all the Social Planners that have come before. If only society did things "our way" then society would be better. Of course "our way" makes us money or gives the person prestige so maybe we ought to be skeptical of what the visionary is selling.
We ought to welcome life improvements that organically come about because the solution improves lives! But this sale pitch that government must intrude and facilitate massive restructuring of the economy and laws so that life can be improved is dangerous and it doesn't matter if Big Tech is selling it or if a Marxist advocate is selling it.
"What amazes me about the Big Tech visionaries is they are no different than all the Social Planners that have come before."
No. There are a few exceptions, but in general they are very different. The important distinction is that unlike those old social planners, most of the the time the contemporary tech visionaries are calling for private investment and for the government to get or stay out of the way.
As an example, ride-hailing services really have transformed the urban transit experience in the way a tech visionary might have hoped two or three decades ago, but they did it with private money and, if not quite technically 'repeal', then the widespread de facto suspension of enforcement of local laws regulating taxi services.
In contrast, the Social Planners of the past would have insisted that the government enforce its own monopoly of these services, raise taxes to buy all the cabs and employ all the cabbies with group preference quotas and exorbitant pensions, charge different subsidized amounts to a hundred different categories of politically-favors consumers, and also develop the government-uber (Goober?) App, which would be exactly as terrible as every sane person would expect it to be. Uber has been accused of abusing the privileged-accesses of its app to more-or-less spy on its users. That's probably nothing compared to how much Goober would be spying on people.
Yes, private companies will try to do everything they can get away with to capture and maintain as much market share as possible, and yes they will also try everything they can get away with to lobby the government for subsidies and protections and arrangements which allow them to privatize gains but socialize losses. These tactics are worthy of condemnation, but the most intellectually mature approach is to put them in the "don't hate the player, hate the game" category and focus on restraining these corruptions of state action so as to make them unprofitable to attempt in the first place.
Fair rebuttal, except "electrification" is a demand on government to impose a solution and have consumers bear the cost and inconvenience.
And as for autonomous machines, Big Tech will do anything and everything to offload costs onto others. Consider self-driving cars. If I have one, then shouldn't the manufacturer of the car assume liability coverage for any harm and damage the vehicle causes to others when it is in "self-driving" mode? Until the "autonomous vehicle" companies step up and assume that responsibility I am skeptical of their technology.
I don't think it's a fair statement of the current law on the subject to say that autonomous vehicle manufacturers and their insurers aren't held liable for those damages, especially if one can demonstrate malfunction or defect as the proximate cause. Certainly some manufacturers will try to contract away as much of this liability as the law will allow them to and shift the burden to customers (who, however, would receive substantial discounts for doing so) but - and please correct me if I'm wrong - my impression is that their legal capacity to do this is very limited, which is what one might expect in very-highly regulated industries like automobiles.
You are correct. There is a term in the industry about trucking cases where the outcome is called "nuclear verdicts" where there are huge sums awarded. In these cases, the plaintiff attorneys go after everyone connected to the accident, including manufacturers. This will continue.
That battery comment set off my radar. If he thinks EV long haul trucks are the answer to automation and that diesel trucks cannot be automated, I'm reluctant to trust much else that follows.
What happened after 1964 is Civil Rights moved on from "stop government discrimination" to "the government has to actively force people to integrate". The Fair Housing Act passed in 1968. Affirmative Action slowly evolved over this time. School busing also got big during this time period.
MLK was going on about things like Reparations and basically going socialist in the late 60s (everyone forgets he was immensely unpopular at the time of his assassination).
It also expanded beyond black people. Title IX passed in 1972.
But nobody on the center left today would say that all of this stuff was bad.
While it's not necessarily dispositive to the significance of Reagan's elections, it's good to remember that except for six years roughly coinciding with Reagan's two terms when the GOP controlled the Senate by slim margins, the Democrats controlled both the House and Senate from about 1960 to 2000.
https://upload.wikimedia.org/wikipedia/commons/7/72/Combined--Control_of_the_U.S._House_of_Representatives_-_Control_of_the_U.S._Senate.png
On your review: I have been wondering whether our economic problems might be due to not being able to measure what contributes to wealth. The book and your review seems to confirm that. I like your idea that:
"That suggests that we ought to be skeptical of the project of coming up with a single measure of progress. Instead, we can pay attention to a variety of anecdotes and indicators. And then we can expect people to argue over what these observations imply."
What are the efforts in that direction? What should I read (as a non-economist) and where can participate in a debate on this? What are the relevant indicators?
This seems more like an exercise in finding excuses to keep looking to government for answers, and no one will like the answers that result. I realize that trying to get government to butt out is a fool's dream, but it ought to at least be mentioned.
"And in the digital world, where much is provided for free, we do not observe prices. And for many items, prices are artificial: hospital charges, college tuition, and the imputed rent of housing services."
I get the difficulty with digital services but could some estimates be made based on what public companies spend on the basket of services they provide? ... Maybe I didn't get the difficulty. Can quantity be determined for free digital services? What about unlimited subscriptions?
I don't get the problem with hospital and college. They still have prices, right?
Why does rent need to be imputed? Just look at house prices, utilities, property taxes, etc.?
That makes me wonder, how are taxes figured into cost of living and inflation?
And "free" digital content isn't really free but a situation where the consumer of the content is product being sold to advertisers.
Yeah, hospitals and colleges have revenues and customer counts so prices can be easily calculated.
It has been clear to me for some time that people are pretty terrible mental time travelers even within their own lifetime.
GDP seems more accurate than any other single number for comparing last year with prior years, tho requiring some inflation (/value of money) adjustments. Similarly between countries, tho the Big Mac index values are also usable simple for FOREX.
The employment rate is a better metric than the unemployment rate.
The median wage is also good, and changes in this wage level are likely more predictive of electoral support or opposition to the current Party than GDP changes, tho there is a lot of correlation.
(My support for more taxes on the rich is limited such that the 90% and 99% income levels, after taxes & transfers, should go up more slowly than the 50% median. If not, raise more from the rich until the avg worker’s increase, the poor getting richer, is faster than the tax slowed rich getting richer.)
Normal folk want a simple, single, number, and the
In comparing people, there are various metrics, like height, weight, sex, race, age, education, party. Income bracket. Seldom are all these characteristics compared at once, and only for college admissions is SAT score important, but much of high school is oriented at attempting to maximize it.
No one number is a great guide, much less perfect, for rating people or films or companies or countries, but people demand a simple answer, and most all of us want a higher GDP, and a spouse who is a 10.
OMG, I would hate a spouse who is a ten. She's probably got an attitude, and will probably spend the rest of her life desperately trying to keep her looks--a task that is doomed to fail.
I'd start with trying to get more out of rail logistics before fooling with robot drivers. See also Richard Hammond's wreck of a Rimac Concept one from Top Gear.
They say ( whoever they is ) that rail transport is constrained by regulation. I'm sure that's overly simplistic.
And being highly agentic is even easier with YouTube and ChatGPT. Excellent essay. Thank you both.