Automation, or Perhaps Not (At Least for a While)

I thought this letter from Dan Hanson to Tyler Cowen was really thought provoking:

I wonder how many of the people making predictions about the future of truck drivers have ever ridden with one to see what they do?

One of the big failings of high-level analyses of future trends is that in general they either ignore or seriously underestimate the complexity of the job at a detailed level. Lots of jobs look simple or rote from a think tank or government office, but turn out to be quite complex when you dive into the details.

For example, truck drivers don’t just drive trucks. They also secure loads, including determining what to load first and last and how to tie it all down securely. They act as agents for the trunking company. They verify that what they are picking up is what is on the manifest. They are the early warning system for vehicle maintenance. They deal with the government and others at weighing stations. When sleeping in the cab, they act as security for the load. If the vehicle breaks down, they set up road flares and contact authorities. If the vehicle doesn’t handle correctly, the driver has to stop and analyze what’s wrong – blown tire, shifting load, whatever.

In addition, many truckers are sole proprietors who own their own trucks. This means they also do all the bookwork, preventative maintenance, taxes, etc. 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’ve been working in automation for 20 years. When you see how hard it is to simply digitize a paper process inside a single plant (often a multi-year project), you start to roll your eyes at ivory tower claims of entire industries being totally transformed by automation in a few years. One thing I’ve learned is a fundamentally Hayekian insight: When it comes to large scale activities, nothing about change is easy, and top-down change generally fails. Just figuring out the requirements for computerizing a job is a laborious process full of potential errors. Many automation projects fail because the people at the high levels who plan them simply do not understand the needs of the people who have to live with the results.

Take factory automation. This is the simplest environment to automate, because factories are local, closed environments that can be modified to make things simpler. A lot of the activities that go on in a factory are extremely well defined and repetitive. Factory robots are readily available that can be trained to do just about anything physically a person can do. And yet, many factories have not automated simply because there are little details about how they work that are hard to define and automate, or because they aren’t organized enough in terms of information flow, paperwork, processes, etc. It can take a team of engineers many man years to just figure out exactly what a factory needs to do to make itself ready to be automated. Often that requires changes to the physical plant, digitization of manual processes, Statistical analysis of variance in output to determine where the process is not being defined correctly, etc.

A lot of pundits have a sense that automation is accelerating in replacing jobs. In fact, I predict it will slow down, because we have been picking the low hanging fruit first. That has given us an unrealistic idea of how hard it is to fully automate a job.

Based on this I can still think of some labor-saving, but not labor-eliminating, automation roles in trucking.

  • Convoying, allowing one driver to lead multiple additional automated trucks
  • Reduction in team driving.  Currently Federal rules (e.g. for rest breaks and maximum driving times) have created incentives for teams of two drivers to move priority freight that needs to be moving constantly and not parked while the driver sleeps.  Automation might allow one person plus the automated driver to keep trucks moving continuously and safely.

One thing not mentioned by Mr. Hanson is the role of regulation.  Safe automated trucks will likely exist LONG before Federal regulatory changes will occur to allow them much use.  This is not just because there is some delay with regulators getting comfortable with the safety aspects, but because affected groups with political pull who wish to keep the status quo will use safety concerns, real or imagined, to hold up the regulatory process.

If you think I am being too pessimistic, here is a story.  The typical steam engine of the 1930's needed a driver and a fireman -- the latter's job was to make sure the furnace was correctly fueled and operating well.  When diesel locomotives came along, one benefit among many was that the fireman was no longer needed.  Seeing this on the horizon, the fireman's union was ready to dig in their heals.  They actually, boldly, took the position NOT that a diesel locomotive needed a fireman, but that it should be required to have 2 firemen!  This was partially a subject for union negotiation, but in the dysfunctional world of railroad labor regulation, it also required some regulatory changes  (as the first industry with large workforces, the government took its first shot at labor regulation in a railroad-specific manner and the result was largely dysfunctional; fortunately for the rest of industry it did a better job with labor regulation later for everyone else).  It took years to totally eliminated fireman from diesel engines.  In fact, nearly every railroad labor saving technology like this (e.g. automatic brakes rather than men on roofs turning break wheels) led to regulatory foot-dagging that allowed the new technology but resisted the reduction in personnel.

  • Peabody

    I call them "Candy Crush jobs". The job has been replaced by technology, so the job consists of playing Candy Crush on their cell phone while sitting in a particular location.

    I could see automated trucks working essentially like more flexible railroads, good at moving full trucks between centralized depots.

  • ColoComment

    Warren, it was your mention in a blog post here of the Marc Levinson book, "The Box," that led to my purchase of the book, and reading of that fascinating story. [Thank you!]
    I recommend that book whenever the issue arises of paradigm change in an industry, because the switch to container shipping from break bulk was primarily an organic development. Each segment of the entire industry had to reach the tipping point when the higher efficiencies was recognized and the expense of changeover accepted. ...ship design and yards to build them, port & harbor facilities, longshoremen's unions, tracking software, adaptation of rail & railcars and OTR trucking, etc.
    It was only well-after the economic advantages of container shipping had been recognized, and the change was well underway, that the government got involved. And then, IIRC, it was only to bring together the various industry players to, for example, standardize the size & locking together of containers.
    Driver-less may happen, eventually, but that, too, should happen organically, from the bottom up, so that the less-efficient, the less safe, the less useful, factors may be weeded out -- which won't happen if it's a shift that is dictated by government or special interests.

  • joshv

    Sure, so you hire a guy to drive out, meet the automated truck at a rally point, drive the truck in to the pickup zone, do all the stuff listed in this article, drive back out to the rally point, and send the truck on it's way. Instead of spending 12 hours behind the wheel, doing something an AI can do, far from home, he spends two hours doing the stuff that's hard to automate, and he does it close to home.

    For breakdowns/accidents, satellite surveillance and remote control can handle most everything. Have no doubt about it, some dude will be sitting someplace monitoring the satellite feeds of 20 trucks at a time. But that's a hell of a lot cheaper than 20 drivers.

  • irandom419
  • Jesse

    I can foresee several mitigants to the problems described....truckers don't really have the ability to repair their trucks themselves, a breakdown or even just a flat tire leaves them dead in the water calling for service anyway. The truck can be made smart enough to do that itself.

    Load handling at yards can be handled sort of like ships do in harbors, there is a harbor master to pilot them in because captains don't have detailed enough knowledge of the harbor to do it safely.

    Security of the load is partially reduced because the truck doesn't need to spend the night anywhere, maybe other than the docks where it loads and unloads.

    Two questions I would still have are refueling (truck stops will want to be paid for people to handle this, probably a minor concern) and bad weather/mountainous terrain or a combination of both. The truck can't stop and chain itself up, or do the perfect thing needed when getting a tow up a slippery hill.

  • marque2

    Interestingly airplanes for long flights needed two pilots an engineer and a navigator. Navigators lost their jobs first, then the engineer.

    New planes today, probably fly safer in automated mode than with a pilot, and with a few minor tweaks could already fly pilot-less. Airplane flying is less complex than the trucking situation mentioned above. The only reason airlines haven't tried pilot-less flights yet is because they are afraid of customer reaction.

    Note that there are already cargo vessels that operate on the sea automatically.

  • John Moore

    Two of the hardest automation problems here are driving in cities, and loading and unloading, which is also mostly in cities. So, what will be automated first (and this has been obvious for awhile) is long distance trucking. The trucks will be driven by humans to what are now truck stops. From there, they'll drive themselves to a truck stop near their distant destination. There, humans will take over and deliver/load and navigate through the cities and towns.

    I expect this to be soon, unless the regulators or litigators put in, ahem, roadblocks.

  • Mercury

    "One thing not mentioned by Mr. Hanson is the role of regulation. Safe automated trucks will likely exist LONG before Federal regulatory changes will occur to allow them much use."
    ------------------------------------------------------------------------------------------------
    No, they will cram them down our throats as soon as they can just as they are doing with electric cars and the superabundance of electronic in cars.

    Other than that he has some good points about truck driving but U think that will follow a similar pattern. The big change will be normalizing the disappearance of criminal liability i.e. if a robot harms you, you may be entitled to monetary compensation but nothing beyond that in terms of justice.

  • CapitalistRoader

    ...affected groups with political pull who wish to keep the status quo will use safety concerns, real or imagined, to hold up the regulatory process.

    That works both ways. As vehicles become increasingly automated the accident rate will drop, which will make human-caused driving errors more visible. Groups called Mothers Against Self Driving will spring up in every metro area to lobby for regulation to reduce or eliminate human drivers. Because ...if it saves the life of only one child, isn't it worth it?

  • Mike

    I disagree. All the duties of a truck driver can be automated more easily than described in the post. Also, many drivers don't touch the loads anymore (aside from flatbeds) so there isn't much tying down or securing The loads are loaded and unloaded by other workers. Customers may miss out on the face to face, but I'm sure they wouldn't care if via automation, their costs were disappearing.

  • cc

    For local deliveries, the driver for sure does himself unload the goods and maybe even stocks the shelves--ever ask the guy at the grocery where something is and he says he is just stocking drinks (or something)? He is the truck driver. At the very least he wheels a hand-truck loaded with bread or cases of canned goods into the store. I have known truck drivers who complained that for pick up/delivery even at warehouses, the guys working there wouldn't help and he had to do it all himself (or wait many hours). If delivering to a retail clothing store, do you think the girls working there are going to unload an automated truck?
    For certain classes of products there is already skimming of the easy stuff. A container gets loaded in Shanghai that is all destined for a single Walmart warehouse in Iowa (or even a single store, I'm not sure). Loading it on and off the boat is automated, then onto a truck. The container is never opened. This process is one of the reasons that international shipping is now so cheap.

  • The moment I read this post in my email, I had the shipping-container idea.

    While I think Dan makes great points in his note to Tyler, this isn't as simple as dropping a robot in the driver's seat. The availability of self-piloting vehicles will change the routes and the logistics in ways we aren't fully appreciating. And it will accelerate when the current drivers and contractors implement the technology - nothing says it has to be top down.

  • SamWah

    I recall reading an article by a long-haul high-end household goods mover who loaded his truck himself, drove to the new house, and unloaded it into the house.

  • marco73

    I'd be willing to guess that the first implementation of robot truck drivers will be in very controlled environments, such as transport of a container on a specific road from a port to a routing depot near a highway. Many ports are usually restricted access because of customs or security of hazardous loads. It might be cost effective to have robot trucks driving back and forth on a restricted road from the secure port to the secure routing depot. The restricted road might only be restricted during overnight hours, and open to human truckers during the day. At the routing depot the load is taken by a human trucker on open highways to its destination.

  • Dan Wendlick

    Prime for automation are the long-haul, overnight, "mail runs" - regularly scheduled, point-to-point routes like what FedEx or UPS ground shipping does between cities, or more generally in the less-than-truckload (LTL) market. For example, UPS distribution center loads up a trailer in Chicago from 5-9 pm, automated truck drops off trailer from Memphis, picks up newly loaded trailer and hauls back to Memphis. Meanwhile Chicago unloads and breaks down the delivered trailer by smaller city and route to smaller trucks. Since the capital is in the prime mover more than the trailer, you keep the assets on the road for a greater percentage of the time and increase return on investment - the trucker's motto: "If the wheels ain't tuning, we ain't earning."

    Shorter routes with multiple stops, "milk runs", are less likely to be automated because you need more control over what goes onto and off of the truck at each stop, even if the routes are the same on a daily basis.

  • bannedforselfcensorship

    Just some observations. We ship palletized loads. The drivers never check the load, really. They go eat lunch, smoke, or whatever. It might be different with different cargo. We have to be more careful with intermodal and have cushioning etc.

  • OldManRick

    Just out of curiosity. Who will put the chains on the automated trucks when it's snowing? Who will check the brakes in the brake check areas before a long downhill? Who will gas up the truck? How will the automated truck know to slow down to road conditions not traffic? Can an automated truck even swerve to avoid potholes, rocks, or debris on the road?

    I've dealt with the academics who figure that because they have a PHD everyone else is dumber than them in every field, no matter how much experience the other guy has in his field. It's not pretty.

    Automate trains first to prove that it can be done on "closed circuits" before you try something with as many variables as the open road.

  • Christian

    Already exists in large open pit mining operations for ore and waste haulage. The complexity of scheduling trucks in this restricted environment is ideal for automation. Example:https://www.cat.com/en_US/by-industry/mining/articles/improving-safety-and-productivity.html