Driverless Semi Trucks Are Here, with Little Regulation and Big Promises
nytimes.com
nytimes.com
Exactly.
Even if a robot vehicle could "just" match the performance of a human, we would still vastly prefer them because we want people to do higher complexity tasks than operate motor vehicles.
Same challenge Uber at al face with autonomous vehicles - when you can underpay a driver 50 cents a mile to drive their own car, it doesn't make much sense to have an autonomous vehicle that costs you closer to 70 cents a mile to operate.
Why? And who is the "we" here?
I really could care less what kind of job someone else enjoys doing. They don't owe me anything and I don't want to tell them what kind of job to have any more than I want someone telling me what job to have.
I don’t want to actively “make” people lose jobs but I also don’t want to limit innovation. The relatively free movement of labor is a benefit for all including workers. I guess if someone really wants that Amazon warehouse picker job great but I am not sure that’s true.
i, too, want free money without having to do any work.
my point is, there's no good way to help those who is economically worthless from technological progression. Any way to slice and dice it is just another word for welfare.
There is no universal direction for progress to move towards, its ultimately just changing things. Regulation removes options, that must limit progress by cutting off certain paths people may otherwise choose to progress towards.
Instead of "we want people to do higher complexity tasks than operate motor vehicles," I would say:
Moving humans from lower to higher complexity tasks generally correlates to increased productivity, higher efficiency, or "getting more with less." Generally, this leads to a wealthier society and other positive things.
The above is not true about everything but holds up in general.
It is likely that part of the reason they are truckers is because they aren't really suited for higher complexity tasks
I'm not trying to be an asshole. I respect the hard work and long hours that truckers do. I don't think they are stupid, and I greatly value the service they provide to society
But the reality is trucking is still a relatively low skill job, mostly filled by people who are not really equipped for high skill jobs
At some point if we automate all of the low skill jobs, we're going to have to deal with the fact that a ton of people who aren't cut out for complex work are now unemployed (and probably poor, and pissed off)
I think the answer is and has always been education and retraining for different occupations.
My general experience is that people tend to fall backwards into either lower-quality work or permanent semi-unemployment, but I could be mistaken about that.
Ultimately, after large-scale technological innovation that leads to societal change like worker displacement (like happened after agricultural innovations displaced farm workers in the American South, or manufacturing innovations displaced American factory workers, or unemployment during the Great Depression) you end up with societal unrest (like organized protest and political action) that leads to increased societal safety nets and welfare benefits… temporarily and if the unrest is big enough.
But eventually, as markets adjust and other types of employment opportunities pop up or as the capital class gets annoyed at rising labor costs, the welfare roles contract to encourage people back into the workforce (in American society, I’m talking about. I’m not sure how other countries handle this.)
There’s a thorough and well-researched (though very dense) book called Regulating the Poor if you want to see more along those lines.
But job retraining tends not to have much of an effect.
Basically, the labor theory of Planck's principle: '"A great scientific truth does not triumph by convincing its opponents and making them see the light, but rather because its opponents eventually die, and a new generation grows up that is familiar with it", Planck once wrote.'
I also wonder how many “new employment opportunities” are generally worse off for the displaced (gig economy etc) and what that means for society.
How well did this work for the previous globalization of manufacturing?
Next time you run through that line of logic, think about that.
We can educate truck drivers probably, but not all. But the concept as a whole has issues. Can we re-educate and engineer to be a 10x engineer? How many of them? So, what happens when you must be a 10x engineer?
People have limits. It varies person to person, but we all have limits and as we go higher "up" the rate of hitting limits increases. We can cut out low hanging fruit like, say, literacy. But I can't be Einstein, and neither can you. So, let us both pray we won't need to become Einstein.
Point being, it was not accidental what happened to farmers during industrialisation of food production. they had a path. they were not left with nothing. so back to you. what next for out of work drivers or their children?
But many of them were.
It's easy, at this distance, to look back and see that a large number of former farmers were able to find work in the new industrial factories and the like, and to simply ignore the fact that many of them did not.
Many of them saw their income dry up, and failed to pay back the loans they had taken out in the spring to be able to plant, and had their farms repossessed.
Many died.
Without some kind of comprehensive, structured program to provide jobs and/or retraining to people being put out of work by a major technological advancement, yes, some of those people are still going to do OK, but many if not most will struggle, suffer, and die. And right now, the government is actively slashing and burning structured programs to help people.
So I'd say the burden of proof lies on those who posit that "something will happen" to ensure that truckers being put out of work by self-driving trucks aren't just left destitute.
At the time farmers were being made destitute by industrialization we also invented the modern education system. This was a massive societal investment to compensate for the reduced number of farmers needed.
There needs to be a similar level of commitment to establishing a viable future for those displaced by AI. So see no indication that this investment will happen.
But yes the reality is that net jobs will probably be lost. Some will transition into the new role of the truck driver, some go into different industries and some will get lost through the transition. The real truth is that it is incredibly hard to make everyone whole.
This is the idealistic take. What's your practical take for moving hundreds of thousands if not millions of people from truck driving jobs to higher complexity tasks? Do those jobs exist today and truckers just refuse to take the step up? Are we creating new jobs that are just the right amount of "more complex" to be realistic for a truck driver?
In a world where AI is doing away exactly with the lower complexity jobs, how are we bridging the gap between the average skill level and expertise of a trucker compared to high enough complexity job that keeps them safe from being replaced by AI in 5 years? What's the reskilling or upskilling plan for these people's skills and career to advance at a faster pace than AI is catching up?
Some will take on more and more of the much more complex logistics jobs like delicate loads, large loads, etc.
Some will switch professions.
Some will go to school to switch professions.
Some new professions will open up exactly because it will become cheaper to move stuff by truck. For example, if a business can now afford to move X to their location, now they can do more business. Perhaps that business hires former truck drivers.
3 million truckers in the US alone. Retire early? To live on what, the wealth accumulated in not even a full lifetime of trucking? Go to school? Who pays for it, who supports the person/family while they go to school, and what are the odds that the skill set increases so drastically that AI/automation won't just catch up with them in a couple of years? "Just" switch professions? If there was a better one they would have done it already, so they can only switch for the worse.
Given the massive number of new drivers and that half+ of them are voluntarily switching to new jobs your whole thesis here is total bunk that they can only switch for the worse.
Trucking is one of those industries where even worse than most others recruiters lie, cheat, and steal their way into getting people into the industry and as soon as they're in they mostly find there really are greener pastures elsewhere.
Are they more complex task jobs? I rejected OP's theory because no evidence was put forward that drivers will just move to "higher complexity tasks". Are they? Are we creating a constant stream of higher complexity task jobs that truckers can routinely reach with moderate upskilling and which aren't at risk of being automated in a couple of years?
We have millions of jobs which haven't changed in years skills wise and are somewhere on the bottom rungs of job quality. Yes trucking is bad, but even "just" 1.5 million people doing this for years means they have no realistically better career option. There are 50% more truckers now than 20 years ago. So we haven't really created all that many "higher complexity task jobs" so far, we just created more of the same.
When the bottom falls out on this one, it falls all at once. The moment automation comes for one trucking job, it comes for all. There will be no bargaining power, there will be just "accept even worse conditions or take a hike". Automation doesn't fix their problem, it fixes yours: stuff's cheaper. Is your theory also that all those people will just magically get better jobs like stated here:
https://news.ycombinator.com/item?id=44114878
https://news.ycombinator.com/item?id=44115072
Is the theory bunk because you jumped straight in the middle of the thread? Or because it doesn't give you the good feelies? I'm not saying it's all doom and gloom. But let's not pretend it's all roses either. We'll get cheaper stuff, and maybe the next generation of people will have better than just trucking jobs, and all it will cost is that most of the current generation of truckers will get screwed worse.
There isn't one, and there won't be one in the USA for the foreseeable future.
This country's people and majority governments have no issue in throwing people away, with no backup or contingency plans. Sector falls over? Too bad, go back to school (and spend MORE money), or live under a bridge.
But when a company sector falls over, they get government subsidies to tide them over till they create the next fallout.
And this isn't just a 'hate on republicans' or 'hate on democrats' thing. All elected parties share this attribute. R's just remove programs, and D's just means-test them to hell and back... But only for the individuals. Companies get free money (PPP), and have it forgiven.
Its unabashed capitalism for the populace, and socialism for the companies. And that absolutely means that people removed from trucking jobs will have absolutely nowhere to turn, unless they make their exit plan themselves.
But also don't live under a bridge, because that's illegal.
In fact, just don't be poor. Being poor may not be illegal yet, but it damn well oughtta be! And we're certainly going to make it as hard and grinding and cruel as we possibly can to be poor, because being poor is proof of low moral character. After all, if you were a good person, God would have made you rich, like Elon Musk!
Basically, if you can't get and keep a job, just go and die like a good peasant.
(/s, because sadly that always needs saying)
Again, it was a massive example of 'Socialism for the rich and unabashed capitalism for the poor'. BOTH PARTIES DO THIS.
The only major positive change I've seen in this government in the last 2 decades that had a strong positive impact on the people, was the ACA. And in of itself was a federalized version of Romney's healthcare. And that was the Heritage Foundation's response for healthcare when Hillary Clinton put forth a single payer in 1995 as first lady.
But even with the ACA, it ended more up to being a corporate freebie to medical insurance companies. And only secondarily, did it start helping people.
But back to robo-semis: those people are already leaving due to tarrifs destroying shipping. And when workers are completely displaced to robotrucks being better than human drivers, there will be no relief for those displaced humans.
This country relies on sacrificing people for a narrower group to get ahead. And those who are behind are then summarily blamed for being behind, as 'poor individual choices', even if the root is completely systemic.
More people will suffer. More people will become homeless. But hey, some truck AI company will make money hand over fist.
The highest priority is being employed.
Otherwise by friend, a driver with spotless driving record and very poor English, is unemployable with two kids to feed.
AI cant savage the laptop class fast enough.
I know a few long haul drivers that did it for decades, one only stopped because a car accident then motorcycle accident damaged his back too much to sit in a cab all day.
In my experience, all bets are off when an industry downturn hits. People jump ship or are thrown overboard. Usually those that handle it best keep a steadier hand and don't react as emotionally, it wouldn't surprise me if that group is more often those with the most experience.
If the attrition of newcomers is high enough, and oldtimers fill newcomer positions (or just their wages) in rough times, then if the high attrition of newcomers is high enough to overcome the phase-in rate of automation then the old timers will more or less stay employed just likely at the shit wages that newcomers get.
As to the attrition, this is mostly to the collapse of the protections of working conditions.
I have never seen economics or wealth as a goal in and of itself. Economics is a view of past performance and only really interesting to compare against other periods in the past or other countries/economies.
In my opinion people should do exactly the job they want to do and the economics of that will shake out however they shake out. People should be enjoying what they do, but more importantly they should be making whatever decision is best for them at the time.
This may just circle back to the question of central planning - I don't think we are ever better off in the long rum trying to plan the entire system rather than trusting every individual to do what's best for them (and trusting that decision to often include what is best for their community).
I was responding to the idea that we intentionally want to put drivers out of work because we want them in higher complexity roles.
> trusting every individual to do what's best for them (and trusting that decision to often include what is best for their community)
which hasn't, in my view, panned out, because of the reasons I outlined in my comment. Between trusting individuals to be altruistic and centralized planning of everything, it seems there's a lot of room for some central authority to decide we need more eg housing and to provide incentives for that, without becoming eg Soviet Russia.
Central planning for housing hasn't worked well, at least in the US. We either over develop poor quality housing or over regulate and drive prices up. Occasionally we try to make affordable housing to help those in need, but the combination of where we build those projects and the bar we try to define for who can use it leads to bad results more often than not.
Does it look like the market is sitting today on a pool of hundreds of thousands of higher complexity jobs that just go unfilled because people choose to "just" operate motor vehicles? Do those jobs look like a realistic fit for today's truck drivers?
There are 3 million truck drivers in the US today. Do "we" understand what's the impact of "our" preference towards what other people should do? Or is what "we" want coming from a place of "I'm safe from all of this so I can afford to have lofty ideals for others"?
https://www.bls.gov/news.release/jolts.nr0.htm
I'm not sure how many of them are more complex than truck driving but it's safe to say that there are plenty of employers looking for workers.
You started with a number you don't understand as the answer to a question I certainly did understand when asking.
You went from "increased skills jobs are so much better for people" to "maybe they'll find some job, I'm not sure". But the average trucker won't find good job in finance, IT, or health. The jobs either aren't better (better work, better pay), or they're not realistically reachable by a normal trucker (easy enough to upskill to make it likely).
Employers aren't usually looking to hire old and inexperienced former truckers when the competition is either people with experience on that job or youngsters willing to take lower pay because they don't have a family to support.
Sadly, this sounds like an anodyne ideal, tuned to rationalize "don't worry, feel happy" indifference in well-to-do people - who are conveniently far away from the affected workers, and the grim realities of how the change will actually play out in their lives and communities.
They'll centralize and monopolize yet another middleman's dream and create the physical Internet, with vast opportunities for rent seeking and wealth extraction.
You will not own the trucks: They'll be rented to you, you won't be able to repair them. If they malfunction, you'll have to talk to some centralized Email or "AI" based customer "service".
Somehow, the monopolists will extract as much money as you previously spent on a driver.
It may not be a monopoly in the beginning, but it will consolidate to at least an oligopoly soon.
https://quoteinvestigator.com/2011/10/10/spoons-shovels/
> If that digging would give work to a hundred men with shovels and picks, why not get a thousand men and give them teaspoons with which to dig up the dirt?”
If a monopoly emerges, by definition control and prices are centralized. And I really don't want broligarchs that follow the agenda 2025 to control the food chain.
It doesn't result in lower prices of items and services necessary for survival in modern society which are also subject to regulatory capture like real estate, healthcare and education.
In the case of the latter, all advance in technology and automation move to the pockets of shareholders but the cost to the consumer stays the same or even goes up once oligopolies form.
Much hailed privatization and automation of the postal services in. e.g., Germany has not led to lower prices. 30 years ago mailmen were literally civil servants with expensive salaries. Privatization and automation has just redirected the money to the stock holders.
There will be no real competition in trucking once a few large corporations divide up the pie and get special permissions to operate.
You can currently get cars for the inflation-adjusted price of a Model T (~$20,000). https://www.cars.com/articles/here-are-the-10-cheapest-new-c...
And they include a lot more standard features.
Cars aren't cheaper because Americans will buy more expensive cars. The average new car purchase price is up around ~$45,000.
(I think buying a new car is dumb, but hey, I'm not the market)
Hopefully so, but it’s not a given. Factors like consolidation, collusion, anti-competitive regulations, etc all can lead to keeping prices high.
Being able to 1.9x equipment utilization rates (or cut transit times over 21 hours) will guarantee adoption.
All transportation alternatives converge into trains over time, it's like carcinization but for transport.
> “I thought for sure I was going to kill those people,” she said.
I hope at least the autonomous truck won't be driving at the very limit of the available stopping distance.
She seems to have been driving too fast around a bend in a signed construction zone.
This driver seems like a perfect example of where the autonomous truck doesn't need to be perfect to improve upon human drivers.
Who?
Further we learn they are proud of "360-degree sensors that can detect objects 1,000 feet away" and "already logged more than 1,000 driverless miles shuttling goods along Interstate 45 in Texas" - that's what? two days? "We have something like 2.7 million tests that we run the system through" and "two trucks without a driver".
What in the world?
Now, I will grant that some of this nonsense is probably the [checks notes] New York Times writer but still. This is absurd.
The conclusion goes beyond that: "What Aurora’s doing is being much more careful than most,"
Even the trucks will probably start experiencing mechanical failures more quickly than anyone expects (based on the passage of calendar time) due to the 100% duty cycle I would think
Most of them get paid $0.X per mile, I imagine the interstate miles are the easy money and the last mile through Manhattan is the hell money.
Right now, we have drivers for all of the miles the trucks go.
Even if we posit that there may be more trucks on the road with self-driving long-haul routes, it's hard to imagine that it's enough more to make up for all the long-haul truckers who will no longer be needed. And at least for most purposes, it seems highly likely that the number of last-mile deliveries is primarily driven by demand, not the availability of long-haul truck routes.
Even if they could increase slightly, there are limits to how many trucks the highways can fit without congesting pretty badly. (Anecdotally, the NYS Thruway is already approaching that limit at some times and in some areas.)
So how would we end up with more drivers total than now?
Driverless vehicles dont need to be perfect drivers, they need to be as good as the worst driver who can get insured. The 16 year old kid who doesnt know how to drive at all and didnt do drivers ed...
Realistically the driverless vehicle AI is probably far better than the average driver already.
All cars and trucks will eventually be driverless.
"Well, the truck plowed into a schoolbus full of middle schoolers, killing everyone, but at least all the families got a nice chunk of cash!"
The 16 year old kid driving the F250 on bald tires, whose driving experience is grand theft auto 4 could drive into that school bus as well.
Whereas you cant source a single story of any driverless vehicle crashing into a school bus.
Do you think we should hold drivers to a higher standard so that we can hold driverless vehicles to higher standard?
Or that 16 year old kid could take his parents' cache of assault weapons and shoot up a mall. (They're preppers, you know.) Or he could research making chemical weapons online and release a nerve gas through his whole school.
Humans have agency, and the kinds of things they can do and are likely to do are fairly well-known.
Also, even the biggest, most lifted overcompensation pickup truck is unlikely to be a match for a full schoolbus in weight. A semi truck is a different story. Crash a pickup truck into a schoolbus, you might kill a kid who's right there; crash a semi truck into a schoolbus, you might send the schoolbus flying 200m spinning top over teakettle, bashing everyone in it around until they're hamburger.
> Whereas you cant source a single story of any driverless vehicle crashing into a school bus.
Humans have also been around for, well...all of human history.
Driverless vehicles have, thus far, been around for just a few years, and only in very constrained circumstances. They have also, to the best of my knowledge, been exclusively passenger cars. Dangerous if misused, to be sure, but not on the same scale as a semi truck. If I invented a brand-new self-flying airplane, and you were (rightfully) concerned about its safety, and I said "you can't source a single story of my DanarisPlanes crashing into buildings!", I would be 100% correct—and it would still be bullshit, and the safety concerns would still be very real.
Putting driverless semi trucks on public roads without extensive safety testing done to the satisfaction of third-party regulators with no vested interest in anything but the public's safety is a huge difference.
The problem is not that there is a chance a driverless semi truck could cause a tragic accident. It's that I don't see any evidence that we have proper understanding of what the risks are—and with a Republican-dominated Congress willfully abdicating its responsibilities for years now, unfortunately we don't have any sane legislation or regulation to give us a common framework for handling this.
The problem is also that just making sure they are insured is not a solution. First of all, the insurance companies also don't have enough information to properly assess the risk, because that information simply does not yet exist. Second of all, as I implied with my earlier snarky comment, when the negative consequences of the self-driving truck failing to behave properly could very, very easily include many dead people, money just doesn't make that better.