> These people have local knowledge that is not easily transferable. They know the quirks of the routes, they have relationships with customers, they learn how best to navigate through certain areas, they understand how to optimize by splitting loads or arranging for return loads at their destination, etc. They also learn which customers pay promptly, which ones provide their loads in a way that’s easy to get on the truck, which ones generally have their paperwork in order, etc. Loading docks are not all equal. Some are very ad-hoc and require serious judgement to be able to manoever large trucks around them. Never underestimate the importance of local knowledge.
I would expect that even if trucks became autonomous, a lot of truckers would survive the transition because people would still need to do a lot of that stuff.
These people have local knowledge that is not easily transferable. They know the quirks of the routes, they have relationships with customers, they learn how best to navigate through certain areas
Isn't that what we were told about taxi services when Uber started taking over? But it turns out that people prefer a driver who follows a GPS and you know he's 5 minutes away than a more expensive local driver that might come in 5 minutes or might come in 50 minutes (or not at all), all you know is that the dispatcher said "He's on the way"
optimize by splitting loads or arranging for return loads at their destination, etc.
Most of this seems like something logistics software can do - and it can optimize across a fleet instead of optimizing for a driver.
They also learn which customers pay promptly, which ones provide their loads in a way that’s easy to get on the truck, which ones generally have their paperwork in order, etc.
Also things that the business office would do -- if you want the cheap automated truck delivery, you'll pay on time and have your paperwork ready.
Loading docks are not all equal. Some are very ad-hoc
Again, this is something that shippers would change if they want the automated truck delivery. They can either pay to make their dock automated-truck-accessible, or they can pay a human driver to back it in. And those human drivers will become more scarce and expensive as time goes on.
I would expect that even if trucks became autonomous, a lot of truckers would survive the transition because people would still need to do a lot of that stuff.
Initially, sure, but long term, economics will win out.
Not around here -- most of my drivers have been from well outside of the area -- the people that need to supplement their income by driving for Uber are pretty much the same set of people that can't afford to live in the area.
They have a general working knowledge of the area, but tend to work such a large area (i.e. the entire SF Peninsula) that they don't have detailed knowledge
Trucks autonomously deliver themselves between cities/ports -and there are a bunch of Drivers at the local hub who then pickup the truck and do "last mile" delivery operations.
People shouldnt need to drive from Oakland to Reno, as an example.
Then, truckers could go to work 9-5 and be home for dinner with their families every night.
I believe this will be the first iteration of autonomous truck driving.
> Then, truckers could go to work 9-5 and be home for dinner with their families every night.
You seem to indicate this will be good for the drivers, but it won't be. The bad news is you couldn't take all of the current drivers and consolidate them into "last mile" drivers. There will be A LOT more physical work involved since drivers will no longer be sitting and driving 99% of the time. And there will be fewer hours driven by humans, so it will truly be survival of the fittest.
Old and less able-bodied drivers will be forced to retire or switch careers. It will probably be impossible for someone new to get into the industry and to be honest, it would be a poor career move to try to get into the industry at this point. The drivers that remain will continue to fight tooth and nail to block full autonomous driving, but every hour they spend fighting is an hour less that they will be earning money in a career that will eventually be extinct.
It sucks that so many people will lose their livelihoods, but it really is inevitable. And big retail stores will die, too, unless they are automate the whole "last mile" delivery. Loading docks and warehouses will have to adapt or the retailers/shippers will die, too.
> to be honest, it would be a poor career move to try to get into the industry at this point.
... Isn't that sort of the point? Should we be designing our society around people getting into an industry for its own sake? Bare minimum, for jobs that are going to go away or be depleted, it seems like pretty solid market economics (and I'm not into capitalism) that reduced demand for labor in a job should most affect the people least committed to that job.
I like that idea too(and as people point out several companies are experimenting with this idea).
It would also be interesting to set up "remote drivers" for the last mile. The truck itself can still be unmanned and someone drives the truck remotely for the last mile (or when problems crop up). Since the infrastructure would already be there for autonomous driving, remote driving is a relatively straightforward exercise in engineering I think.
Although the paperwork problem will definitely be an under-appreciated obstacle to making these systems work, autonomous trucks will eliminate a significant portion of it because there will no longer be any such thing as hours of service, or mileage logging for drivers.
Ditto for licenses, drug tests, background checks, and most of the common kinds of moving violations. and scheduling has its own set of paperwork which is a colossal nightmare because you are dealing with the always conflicting priorities of drivers on one side and customers on the other.
The major logistical obstacle outside of paperwork will remain to be resolving last-mile issues, but that has already been resolved long ago for containerized railroad freight: a.k.a. autonomous trucking technology from the 1800s. There's been a century's worth of time to solve this, manifested in the form of short-haul service providers known as intermodal draymen who pick up the container from a designated landing area (a rail yard) and haul it the last mile. This can easily be adapted to autonomous trucking logistics.
Between the Flexport guy who's always on here, and Uber Freight, someone will probably get the paperwork automated. After all, DHL, UPS, and FedEx already have that problem solved.
There is also the situation of a truck breaking down and needing repairs, which can happen relatively frequently, given the amount of miles it runs daily.
Separately, how would a truck handle running out of gas if it spends too much time stuck in slow-moving traffic? I guess they would need electric human-operated charging stations for autonomous trucks?
Tires blowing out are a much bigger (in terms of simple cost and the thornier problem of schedule optimization and ability to guarantee precise delivery times) and much more common of a problem than mechanical breakdowns. They're also a problem that does not go away with autonomous trucks or electric motors or anything conceivably around the corner, outside of some miraculous development in materials science.
As such, a (somewhat) well-organized industry has developed around getting replacement supplies like tires and other accessories to the site of a breakdown as quickly as possible, and this would simply need to be extended to include whatever it is autonomous trucks would require that doesn't touch the road.
Why do tires blow out? Do they have a predictable lifespan? If there is a huge cost savings associated with autonomous driving, could some of those $$$ be put toward more frequent tire changes that prevent blowouts from happening?
(edit) Personal injury lawyers assert that it is largely a preventative maintenance issue:
It is much more than a lack of maintenance. The weight of the cargo, the temperature of the road, the way the cargo is distributed in/on the trailer and manufacture defects can all contribute to blowouts of servicable tires. Those issues also make the lifespan of identical tires vary by thousands of miles traveled.
While replacing tires more frequently would reduce the number of blowouts, it would be very wasteful. In the US, DOT inspections are frequent and they look at the tires to make sure they are in good condition.
It's interesting that you bring up this point because these are all aspects that would have to be taken into consideration before autonomous highway driving is safe for trucks, tire blowouts aside.
Most people think of driving on a highway as the simplest thing there is, but when you have tens of thousands of pounds, sometimes unevenly loaded, sometimes as barrels of liquid that slosh around and change your center of gravity, all that on a stormy, windy, icy day driving in the mountains, it can make for very uncomfortable driving. Is it solvable? I don't see why not, but I think it might be a bit more difficult than it sounds.
For every single liquid load, there are likely 50 non-liquid loads that can be better if automated. Perhaps humans only did the "dangerous" loads such as gasoline, or barrels of oil, etc? I think this isn't so black and white, but really is more of a shade of grey. Some things humans can do well autonomous trucks will do poorly. Some things humans do very poorly autonomous trucks will do very well. One of the other things is simple economics. There will likely never be autonomous truck unions picketing for shorter hours or higher pay. Note that I've got nothing against truckers unions, but think the sheer economics of the issue to carriers will inevitably force the issue eventually. The first carrier to seriously use autonomous trucking will likely be capable of delivering products to customers for cheaper, creating an imbalance in the industry. Every other company will be forced to compete, with a downward race on wages.
Both of those can be solved pretty easily, there's already a fairly strong business of roadside assistance and towing for trucks the operating company could either contract with a network of those to provide repair or refueling services or just call one from a list whenever a truck breaks down or runs out of gas (the second should be exceedingly rare).
I could see road-side battery exchange - assistance arrives with fully charged batteries, swaps them for the empty batteries, truck rolls on. Faster than pumping fuel.
* Mechanical: pull over and wait for a service truck. Not much different from what happens now. Likely fewer issues overall as these trucks will probably be electric.
* Route planning isn't that hard of a problem; design the system such that running out of gas/battery is a whatever-sigma chance. Automation means lots and lots of data about the roads to base this on.
You could still have a single driver/coordinator leading a convoy of autonomous trucks and handling all the duties that would normally be done by each individual driver. One driver per 5 trucks is still downsizing 80% of drivers.
Truck drivers often work for freight forwarders who arrange the loads for them, including combining loads and shielding them from unreliable customers.
All that knowledge about waiting times, load times, customer reliability etc. can be gathered by the freight forwarder.
Traditional freight forwarders are having trouble with this because they often don't have the technology to manage and understand this data. Instead they rely on lots of personnel and deep hierarchies including many subcontractors.
At sennder.com we're trying to change this.
I'm expecting an autonomous truck to be a supplement to a driver rather than a replacement. It becomes a tool to make him more productive for his customers.
I support your statement 100%. Lets analyze a simple problem. What happen if you throw a mannequin in front of a truck? If it stop then there is a very easy way to pirate them. Human are always needed in very complex decision making circumstances.
Pirating trucks with humans in them is not much more difficult. Humans also tend to stop when you put obstacles in front of them and I doubt they would defend their cargo if a gun is waved in front of them. The reason that highway robbery is not common today is not because it's difficult to do.
The autonomous truck will trigger an alert when it's forced to make an emergency stop, and continually record its surroundings and broadcast its position. A watchdog process will note that the vehicle has failed to reach the next waypoint in the expected amount of time.
Sure, and that could happen now, but if I had to stereotype drivers(which is how criminals frequently operate), truckers would be the last kind of person I would want to mess with. Surely a huge percentage of truckers, especially overnight truckers, are packing heat.
I think the first couple rounds of autonomous trucking will be to basically eliminate long haul trucking and a transition to a model where humans drive the last 50-100 miles from a hub to the customer where most of that knowledge is most needed.
Refrigerated loads often do not ship multi modal. My understanding is that refrigeration is not guaranteed, and that time to the customer is very important.
Some of these cases apply much more to contractor truckers (majority), but apply less to employee truckers (think Publix, Kroger, USPS) where the paperwork is already finished, no upsell opportunities, another office handle payment (even with contractors this is becoming a front office thing).
> require serious judgement to be able to maneuver large trucks around them
I think this is one aspect you get incorrect. With LIDAR and SONAR a computer can know an exact distance from corner to corner, can know and set the exact angles to hit a spot; while with a different set of eyes looking at a tree beside you. We already have consumer cars that have proximity warnings and trailer paths.
I think you're vastly underestimating the enormous pressures that will be put to bear on solving literally every other problem if it means even the possibility of eliminating a majority of the drivers entirely and their huge cost to the bottom line which in modern days is placed almost literally above all other aspects of the business even to the detriment of the fundamental service being provided as long as it's bearable by the market.
Everything and more on that list is true. However, the move to autonomous Trucking will start with the larger trucking companies, and all of the decisions about how to get a load which loads to take etc etc are made by the company not the driver.
I'm by no means an expert on this, but I do follow some trucking vlogs on youtube, but it amazes me how inefficient the loading and unloading process is. Sometimes truckers have to waiting many hours just to get a place at a loading dock and then again many hours just so that a dedicated team can unload the trailer. What a waste of driver, truck, and trailer resources.
The autonomous truck will most likely happen but my prediction is that the real revolution will be automated loading/unloading. And the automated forklifts will unload into fully automated warehouses. And fully automated pickers will pick orders for delivery to your doorstep via fully automated drones. And driven from the doorstep to your fridge by a cargo roomba...
The inefficiency isn't the problem, it is a symptom of the real problem, which is lack of trust.
It's been a while since I drove trucks but I recall the absolute worst loads were deliveries to grocery warehouses. Because of DOT Hours of Service restrictions, it sometimes made economic sense to refuse these loads and instead sit idle a day with no work. You had to carefully budget your working hours over a rolling 8-day window and wasting hours at a grocery warehouse kills your earnings for days afterwards.
Grocery warehouses took forever, mostly because of their insistence that I wait while they counted everything and broke down pallets of stuff and re-stacked it. The actual unloading was quick, it was the hours afterwards before they would acknowledge receipt of goods and sign your paperwork to release you. They also used 'lumpers', third-party laborers (who they also did not trust) to count and re-stack correctly so their own employees would have to come double-check everything. Some would expect me to pay the lumpers, which I would always refuse to do, which further delayed things. I think some drivers just paid to get out of there faster.
Contrast this with more sophisticated operations such as a Wal-Mart distribution center where you could drop your loaded trailer and collect an empty one, then leave immediately for your next load. Wal-Mart wasn't worried about under-counted freight or breakage, they are big enough to just declare it unchallenged and the supplier accepts the loss and moves on. Trust doesn't factor into the process.
It wasn't all bad, you could offer to unload yourself for extra pay on some loads, but I only did this very rarely, once when delivering electric cars (GEM cars, like a super-deluxe street legal golf cart) to a warehouse. I drove them out of the trailer onto the loading dock and got paid an extra $100 for about two minutes of work! Easy to count too...
Ya I agree in theory but disagree in practice. I moved furniture for 2.5 years and delivered appliances for 6 months for a warehouse from about 2001-2003 while I was trying to get my shareware business going. This meant sometimes loading and unloading full 53 foot semi trailers of furniture and boxes (typically 2 or 3 households worth) with just 2 or 3 people, up to 2 or 3 times a day if we were working in the warehouse itself and had a few stragglers who could help.
For a while there I came up with roughly one invention per day that would radically reduce the amount of labor we performed or protect us from the sun and elements. For example, in big cities they use lifts to windows and chutes on stairs to get boxes up and down tall buildings. But they don't use anything in smaller cities except elbow grease usually.
Anyway, whenever I brought up a new idea, it was soundly rejected by pretty much everyone. The standard answer being "that would put us out of a job!"
There is massive resistance to any kind of innovation in the shipping industry. My gut feeling is that it won't change at all until it's too late, and then the whole thing will come crashing down when it's replaced by self-driving forklifts and drones. It's one of the few industries that still (under)employs millions of people, and I think that we should keep that in mind before we work so hard to automate it into oblivion.
But automation is inevitable and will come within 10 years or so, so if we really want to help society, we should be focussing on education and vo-tech training in order to find meaningful work for everyone who's going to be misplaced. I know in my heart that that won't be enough though, which is why I'm a huge proponent of universal basic income.
I did removals for a couple of years as a young man too. I disagree with your conclusion. There are so many different physical characteristics between each day's jobs - parking setups, strange paths through gardens, different types of building layouts and contents, different types of unusual belongings, the behavior of customers in one of the most stressful activities in most people's lives (possibly combined with other stressful circumstances such as divorce, bereavement or repossession), the need to pack and handle both very fragile items and very heavy and bulky items...
People who reckon you're going to build a robot that can handle this huge diversity of demands at a cost that can beat low paid manual workers are being very techno-utopian. Same with cleaning hotel rooms,
doing roadworks, building up and knocking down special events etc. Not to mention the immense amount of work in reconfiguring physical infrastructure if we ever get remotely serious about climate change. If anyone was making a betting market on achieving automation of those kinds of jobs within a specific timeframe I'd be an eager counter party.
Now if you're a radiographer, legal assistant, project manager, sports reporter, then yeah the times are a changin'...
I always thought it was odd that we standardized shipping containers and pallets, but not truck trailers.
Shipping containers and pallets have owners, but nobody minds leaving them behind at the point of delivery because they are interchangeable and they circulate around.
If truckers could pull up to a loading dock, unhook the trailer and leave it behind, and drive away with an empty trailer, that would be pretty cool.
> If truckers could pull up to a loading dock, unhook the trailer and leave it behind, and drive away with an empty trailer,
That's called a 'hook and drop' load and they are very common. Usually used within a large customer's trailer fleet, such as ones who have distribution centers that then fan out to stores.
and the impact of electronic logbooks - drivers who could have driven an extra hour on Monday and an hour less on
Tuesday now have to match their route and good stopping places with the DOT hours structure, or face penalties
for violations.
For a person driving a car, a penalty
could be a fine and point(s) against their license, but
for a commercial driver, loss of their CDL would mean
changing careers.
Instead of punishing the individual driver (which is what's done), ought to heavily punish the company instead.
Right now, a driver can get threatened and written up if they refuse to drive, even if they would be breaking the law. There's no repercussions for a company to process through drivers.
If instead the DOT focused on the companies, they could turn the companies themselves into enforcers as well of the law.
The only group that could fall through the cracks are owner/operators. They would still have a justification and reason to skirt the law and fake the logbooks.
All of the elogs are data mined big brother style.
Guys who run the same route with the same cargo every week and usually get waved through weigh stations get stopped and checked out because they decided to take their bathroom break on a different leg of the route.
If the “Autonomous Truck Port” model becomes popular,
I wonder if we will see a highway-oriented version of
the https://en.m.wikipedia.org/wiki/Intermodal_container
with a lower tare weight but the same twist-locks and
dimensions?
This could mean that ATP-to-ATP trip
segments would be done by trailers built for containers,
so load/unload operations could be automated,
and hookup/unhook of trailers to tractors
would happen much less often.
Also, equipment to handle ocean port
and ATP container movements would be
similar in specification.
That is a very good idea, though the development and standardization of the current series of intermodal containers was about a 20-year process, and that's with _enormous_ assistance from Vietnam War military logistics planners, who desperately needed a non breakbulk solution to supply line issues that were not solvable with the tiny conex boxes they had on hand.
Intermodal are built to stack 9 units high.
I’m looking for a containerization system to replace
manual hook-and-drop of box trailers with an
automatable load, lock, and drive away process
that has a lower tare weight penalty than
intermodal. A TEU has a tare weight of 5000 pounds,
FEU is ~5500 pounds, which is a lot to carry around if
you don’t need to stack units 9 high for an ocean trip.
You'd probably be better off looking for an interface within an interface.
Just accept that the cost of changing from the existing 'intermodal' shipping container as an OUTSIDE standard when on the ocean is too high.
Having a light weight inner unit, maybe something like a pallet sub-unit or dimension-ed around being pallet compatible but expecting a long run to be conveyed out quickly, then loaded in to something else quickly would be a better place to begin from.
Then you'd only need to figure out a good boxing solution for THAT on an existing trucking platform. Re-using the same intermodal locking interface for existing trailers but with a lighter build on top could be useful, particularly if the re-packing allows for a full container load to just go somewhere.
However since unpacking is happening, automated pallet handling/routing might also be a useful economic goal since there would then be an opportunity to sort and re-pack/re-mix at the dock-side.
You'll still want them to be able to stack at least two units high or they will not be eligible for 90% of rail transit. Also having more than one container standard will create a different set of logistics problems for storage, repositioning, and allocation / return of empties.
Are you sure that the stacking height is driving the tare weight? I would guess that it could be retaining longitudinal stiffness even after some corrosion. Also theft protection is driving the material selection more than anything else.
A better solution would apply a minimum standard of living and benefits to ALL workers - irrespective of the number of hours worked everyone should get benefits (health, sick/vacation 'savings', unemployment) which follow a worker and for which pro-rated jobs at least contribute a pro-rated portion.
Also, 'on call' or 'must be available' should be considered at least minimum wage PLUS benefits idle time since the employee is being asked to be on standby and not take other work. Schedules should also be known months in advance so proper planning can be made.
Can anybody expand on the link between autonomous car driving vs autonomous truck driving in terms of challenges faced. If they are similar, not sure why there is this kind of forecasting seeing how autonomous car driving is pretty much far from over or even usable. If they are vastly different, I don't imagine trucking is any easier than cars (maybe it is?)
Depends on where the truck is driving. If we restrict them to highway or designed hubs off of highways the problem of trucks isn't much harder than cars. When we start talking about maneuvering in smaller areas like city streets and receiving areas at the destination it's definitely somewhat harder just because there's so much more vehicle to handle and keep track of.
One reason to think it'll happen much quicker in trucking is because the economic incentive is much higher for a trucking company to replace it's drivers than for a single person to use an autonomous car.
For trucking, solving less of the total problem is still immensely valuable. For example, being able to do simple platooning on freeways and having a driver available to navigate complex or congested environments can still provide significant commercial benefits in cost & efficiency. This isn't as much the case for autonomous passenger vehicles.
Additionally, the safety and compliance standards are different. For instance, ISO 26262 only applies to passenger vehicles.
The difference is in economic incentives, not necessarily challenges. Driving a truck is definitely more difficult than a car but well within the ballpark of any self-driving technology capable of autonomous city driving.
However, while the average consumer won't pay an extra $10-??k for a first generation self driving car that can't navigate their daily commute, a trucking company will glady pay an extra $50-???k for a first generation self driving truck that can only drive on highways to distribution centers if it means eliminating the human driver. The latter is as much as 50% of the operating cost of transportation, even when long haul trucking, so as long as capital and maintenance costs are comparable there is a lot more room to profit off of each sale. Even if it takes autonomous cars many more decades years to become an every day consumer item, autonomous trucks will likely be operational much sooner just because there is so much more money to be made in the short term from a few clients like Walmart, Amazon, Costco, etc. that don't need a perfect driving experience to benefit financially.
> Many other freight-moving jobs will be created in their place, perhaps even more than will be lost, but these new jobs will be local driving and last-mile delivery jobs that—absent proactive public policy—will likely be misclassified independent contractors and have lower wages and poor working conditions.
I'm sure the industry would want that, but I wonder how many people are going to go to the trouble of getting a commercial driver's license for a shit job.
Having a CDL is particularly risky, for a number of reasons. For example the blood alcohol content necessary for a DUI is only .04 for a CDL holder, and that's regardless if you're at work or driving your personal vehicle. It's usually twice that in most States.
And even if you test below .04, if it's detectable at all law enforcement can ground you for 24 hours. And that won't look good to your employer.
For most people, if they lose their license for any reason, they can still figure out a way to get to work. But if you're making your living on your CDL, it doesn't matter if you figure out how to get to work, because you can't work.
Depending on your circumstances, it can be quite a lot of trouble and expense to get a CDL. Why go through all that and take that risk, when it's just as easy to be a helper on a roofing truck or a fast food worker.
Hence the importance of pro-active policy measures. It'll be interesting to see this play out in the next few years as protecting truck driver jobs with regulations should (hopefully?) receive bipartisan support.
I think the autopilot scenario listed is the most likely adoption path to autonomous trucks in the short term. I think the proponents of this technology are very optimistic about the adoption timescale. I also see almost no adoption in local trucking in the near future.
The platooning mechanism may see more use outside of North America, maybe in the LTL fields here - but it would be a pretty disruptive change in the way that freight is sent and routed compared to current use.
...truck makers racing to build autonomous trucks.
This trend has led to dozens of reports and news
articles suggesting that automation could effectively
eliminate the truck-driving profession.
Could you beat around the bush more then this?
'trend' 'suggesting' 'could' 'effectively'
Driving is being automated and then the truck-driving profession will be gone.
The rest of the article seems similarely bloated. With it's scenarios and diagrams and stuff.
Driving is being automated and then the truck-driving profession will be gone.
Not for a 100 years, don't believe the self-driving hype. That's what the rest of the article is about, the long distance highway truck driving jobs will (mostly) disappear but there's going to be a ton of city truck driving jobs that are likely to remain the domain of humans for decades to come.
>but there's going to be a ton of city truck driving jobs that are likely to remain the domain of humans for decades to come.
This. Someone has to wheel around that pallet jack loaded with beer. Someone has to hand the clipboard and pen to the person signing for the delivery. Someone has to tarp the load.
At some point the tech will get good enough, a switch will flip and meat based steering wheel holders sitting in front of dry vans on interstate highways will be replaced with semiconductor based steering wheel holders.
Would technicians, construction/renovation workers, and the likes be good alternative career paths for the current truck drivers?
If so, there should be policy support/subsidy to help train those who wish to shift to these options which, to my knowledge, pay pretty well and will continue to be in demand for the foreseeable future (little automation is likely, given the limits of current technology.)
You don't get many good carpenters out of unemployed truck drivers. Good carpenters are born to be good carpenters.
That's the problem with most difficult to automate jobs that don't require college degree. They sound like anybody could do it. But there is huge difference between good renovation and bad renovation. The kind of guy who can do good job consistently is rare. The job requires passion and good attitude and personality that matches the job.
It's kinda sad how many men are passionate about driving a car. It's bit like how many women used to be passionate about weaving.
Since the US might need to build over a million affordable housing units to address its homeless problems, ‘newbie’ builders, carpenters, and other technicians might help to reduce the unit budget requirements to a level which a large quantity could be built within a realistic fiscal constraint.
Artisan carpenters/master technicians would still be in high demand but those transitioned to the new careers would serve a different market and needs.
> These people have local knowledge that is not easily transferable. They know the quirks of the routes, they have relationships with customers, they learn how best to navigate through certain areas, they understand how to optimize by splitting loads or arranging for return loads at their destination, etc. They also learn which customers pay promptly, which ones provide their loads in a way that’s easy to get on the truck, which ones generally have their paperwork in order, etc. Loading docks are not all equal. Some are very ad-hoc and require serious judgement to be able to manoever large trucks around them. Never underestimate the importance of local knowledge.
I would expect that even if trucks became autonomous, a lot of truckers would survive the transition because people would still need to do a lot of that stuff.