171 | Christopher Mims on Our Interconnected Industrial Ecology
What this covers
Christopher Mims, a technology columnist at the Wall Street Journal, sits down with Sean Carroll to examine how the global supply chain actually works—and what its obsession with efficiency has wrought. The conversation traces the chain from factory to front door, treating it not as a planned system but as a self-organizing industrial ecology of competing marketplaces that touch only at ports, warehouses, and shipping nodes. Mims argues that this network, while awe-inspiring in its mathematical precision and reach, has become dangerously fragile and deeply dehumanizing through decades of pursuing efficiency above all else. The core question threading through: can this system be deliberately re-humanized without collapse?
The discussion moves across multiple dimensions of how efficiency is engineered into the chain. Mims traces Taylorism—Frederick Winslow Taylor's time-and-motion breakdowns—as foundational to modern work structure, and shows how containerization, just-in-time ordering, and algorithmic management amplify that legacy. He examines specific points of vulnerability: how a smartphone CPU's components scatter across six countries and three continents, why empty containers pile up in the wrong ports because of US-China trade imbalances, and how de-skilling makes workers replaceable but does not eliminate the work itself. The conversation also tracks a sharp divergence in outcomes: UPS drivers, protected by Teamster representation and conscious safety practices, versus Amazon warehouse workers managed by algorithm under conditions that produce repetitive-strain injuries. Mims argues that automation changes work rather than ending it, making the choice to structure that work humanely or brutally a political decision, not a technological inevitability. He also notes a paradoxical case that consolidation into dominant firms might paradoxically make regulation easier than today's fragmented trucking industry.
Mims argues the global supply chain is a self-organizing 'industrial ecology' of distinct markets rather than a top-down system, and that the relentless pursuit of efficiency through automation and Taylorist management has created an awe-inspiring but fragile and dehumanizing structure that must be deliberately re-humanized.
- No single actor controls the supply chain; it is an interconnected set of competitive marketplaces that communicate at ports, warehouses, and trucks
- Efficiency techniques (containerization, de-skilling, just-in-time, management-by-algorithm) drive productivity but increase fragility and worker harm
- Automation changes the nature of work rather than eliminating it, leaving a choice about whether to make that work humane
Supply chains are highly optimized mathematical systems vulnerable to shocks because just-in-time efficiency eliminates inventory buffers.
- Logistics is saturated with high-stakes mathematics because a fraction-of-a-percent efficiency gain saves tens of millions of dollars; firms invest enormously (UPS spent $300 million ~20 years ago on its Orion routing software) to approximate solutions to intractable, NP-hard problems like the traveling salesman that could never be solved exactly before the heat death of the universe, with predictive analytics being one of the best mundane applications of machine learning.
“UPS spent $300 million. This was like 20 years ago. $300 million just to stand up their Orion software, which dictates like the routes of UPS drivers every day”
- Just-in-time manufacturing optimizes for both speed and predictability by shipping goods rather than storing them, which is highly efficient but means that any supply shock leaves no inventory buffer, so the system simply runs out of the affected good.
“if you don't have any extra inventory sitting around and there is a supply shock where you can't get enough of something... then you're screwed, you're gonna run out.”
- Like biological species that over-specialize and then go extinct when the ecosystem shifts, the highly optimized global supply chain has become dangerously specialized and fragile, and the remedy for such fragility—now being pursued by chipmakers building redundant fabrication capacity—is redundancy rather than maximal efficiency.
“one of the major reasons that we think species go extinct is that they over-specialize and then the ecosystem changes... the one way that you build resilience, of course, is redundancy.”
- Cod caught off Scotland is frozen and shipped roughly 20,000 miles to China to be filleted because labor costs there are far lower, then shipped back in the cold chain to a Scottish supermarket—a journey that is absurd from an energy perspective yet was economically rational when shipping was cheap, illustrating that shipping, not labor or fabrication, is the easy part of global commerce.
“it was frozen, it went to China, it was filleted, and then it's shipped back in that same cold chain to Scotland, because the cost of labor is so much lower there. So it's like this 20,000-mile journey for your cod”
- A single complex component like a smartphone CPU traverses a globe-spanning, multi-country chain—ultra-pure quartz sand from Appalachia, melted into crystal possibly in Australia, sliced and etched into wafers at TSMC in Taiwan, packaged in Malaysia, assembled in Shenzhen and possibly India or Vietnam—making the system maximally vulnerable to trade wars, actual wars, and pandemics, so that it is almost a miracle the system has not failed.
“if you were trying to design the most Rube Goldbergian and the most vulnerable to trade wars, actual wars, global pandemics, I mean, honestly, it is a miracle that we haven't run out of everything by now”
- The current supply chain crisis is driven as much by an explosion in demand as by supply constraints: lockdowns shifted a few percentage points of consumption from services to goods (which are biased toward East Asian e-commerce supply chains), shippers cancelled orders early in the pandemic fearing a recession, and a permanent lifestyle change toward working and living at home ('family in-sourcing') created demand that far outstrips a system that cannot ramp up quickly.
“as much as we have a supply of shipping crunch, we have an explosion in demand for shipping. This is because lockdowns led so many more of us to buy things via e-commerce”
Worker dignity requires agency to resist unsafe conditions; automation without worker power becomes Taylorist surveillance that drives burnout.
- The utopian vision that automation will free everyone to be poets is elitist and misreads what automation does, because for most people the dignity of work—the instinctual satisfaction of doing something and seeing a result—is essential, and ending work would leave many idle and unhappy; the goal should instead be making work humane and dignified while decoupling people's sense of life-meaning from their jobs.
“this whole idea that we're gonna get fully-automated luxury communism or whatever you wanna call it. And everybody is just gonna be a poet, I got news for you, like half the people are just gonna sit around smoking weed all day and be very unhappy. Like doing stuff, the dignity of work is extremely important.”
- The Toyota Production System worked partly because, within Japan's lifelong-employment culture, it empowered workers to halt the line (the Andon flag) when something was wrong; when Amazon executives imported these ideas, the worker-empowerment component was removed and replaced with surveillance, mutating a humane system into Bezos-ism.
“there it's more about like, 'Let's empower the worker.'... That exact sort of idea, set of ideas, got imported into Amazon by some executives who I talked to in the book, but it got mutated in some ways and became Bezos-ism. They took out the worker empowerment and just put more surveillance.”
- Whether efficiency improvements help or harm workers depends entirely on worker agency: in unionized industries like auto manufacturing, workers can push back on unsafe or unsustainable changes and will embrace efficiency that is also safer and more humane, whereas where workers have no agency, as in Amazon warehouses, the relentless repetitive pace—driven by surrounding automation—produces musculoskeletal disorders like carpal tunnel because the body cannot repeat the same action ten hours a day.
“everything depends on how much agency that worker has”
- Bezos-ism is Taylorism updated with pervasive real-time technological surveillance, so that workers are managed by algorithm rather than a human boss: an Amazon fulfillment-center algorithm flags time 'off task' and issues automated warnings and write-ups, a model now copied across fast food, call centers, and shift scheduling, and the resulting lack of any human flexibility ('no button for a bad day') is a driver of the Great Resignation.
“management by algorithm... if somebody falls behind in an Amazon Fulfillment Center, it's not their boss who's like, 'Hey, work faster,'... it's an algorithm”
- Frederick Winslow Taylor is to industrial capitalism what Marx is to communism—the first management consultant whose time-and-motion studies broke jobs into minute timed tasks to speed up work—and his ideas are unfamiliar only because they have become the invisible 'water we swim in'; the approach was so brutal when implemented at the Watertown Armory that workers struck and Congress banned it from federal facilities, yet it now structures nearly all work.
“I think Taylor is to modern technological industrial capitalism. But his name is not familiar because his ideas have become the water that we swim in.”
Containerization and standardization created a global logistics network by solving coordination problems through a single intermodal standard.
- Adopting containerization required overcoming a coordination problem: everyone could agree a single standard would be beneficial, but someone had to bear the high cost of going first; Malcom McLean broke this deadlock largely because a US government contract to move material into Vietnam gave him the startup capital to build purpose-built ships and standardized cranes, after which adoption cascaded worldwide.
“ultimately he won, in no small part, because the United States government needed to move an enormous amount of material into Vietnam for the Vietnam War. So he got that contract”
- Containerization radically increased port throughput: where unloading a break-bulk ship of barrels and sacks by an army of men took about two weeks, a containerized ship—even a very large one holding 10,000 forty-foot containers—can be unloaded in about 24 hours, making container shipping the most efficient way to move mass across the Earth's surface.
“Usually it used to take like two weeks to unload a ship. A containerized ship, even a big one, they can unload it in 24 hours.”
- Containerization is a deep, general phenomenon of standardization analogous to internet data packets or electrical plug voltages: by creating one standard intermodal unit that any mode of transport (ship, truck, train) and any standardized crane can handle, it enabled the entire global trade network, just as the underlying internet protocol turns data into uniformly-handled discrete packets.
“any time you wanna create a system, whether it's the Internet or a network of fungi in the forest or a shipping network, you want to containerized things, so you have a standard thing that all the different modes of transport can handle, and that's called intermodal”
- Because steel intermodal containers are built heavy to withstand ocean travel, many are not carried far by truck; instead their contents are emptied in large warehouses near ports (like California's Inland Empire) and cross-docked into much lighter semi-trailers to save fuel, except where goods travel by more-efficient rail.
“a lot of containers get emptied in these vast warehouses that are kind of invisible to us, in California's Inland Empire... And the materials cross-docked, as it's put, into the back of a long haul truck”
- Queuing theory—the mathematics of waiting in lines that grew out of Erlang/Shannon-era work on how many switchboard operators a town needs—is integral to the whole supply chain, because every node (ships waiting offshore, three-dimensional stacks of containers in a terminal, trucks lining up) is a queue, and a single fully-unloaded modern container ship would produce a line of trucks about 50 miles long, which is why ports automate yard grooming to minimize total moves.
“queuing theory, which is literally the math of waiting in lines, is absolutely integral to the entire switching because everywhere you go in the supply chain, everybody's waiting in a line”
Trucking faces a retention crisis rooted in working conditions and lack of worker power, not labor shortage or demographics.
- New drivers for big trucking carriers are in a position resembling indentured servitude: they may owe the company for the training that licensed them, locking them into a year-plus contract they cannot exit while working roughly 80 hours a week for about $45,000 a year and never seeing their families.
“it's really like indentured servitude. You may owe them for the training that you got to become a truck driver, and so you've got at least a year contract with them that you can't get out of, and you're working like 80 hours a week and you're making $45,000 a year”
- By value of goods, more than 90% of what moves through America moves on trucks, so the old phrase 'if trucks stop, America stops' is literally true, and a tight trucker labor market poses a real risk of supply-chain disruptions affecting everyone.
“by value of goods, it's like 90+% of the value of goods that moves through America, it moves on trucks. Like that old phrase, If trucks stop, America stops.”
- Truckers get poor political representation because the prominent lobbying body, the American Trucking Association, represents big carriers rather than the majority owner-operators and small companies, whose own lobby is rarely quoted—evidenced by the only trucking provision in the infrastructure bill being the ATA's request to lower the commercial-license age from 21 to 18 to expand the labor pool.
“the big lobbying body there, it's called the American Trucking Association, it represents these really big carriers, but they don't even represent the majority of the trucking industry.”
- There is no shortage of people wanting to enter trucking—there are three times as many people holding the required commercial driver's license as there are working truckers—but rather a burnout and retention crisis, because the typical trucker is on the road 21 days a month, works 14-hour days, is paid only by the mile (not for time stuck at warehouses), and often cannot find a legal place to park and sleep.
“There are three times as many people in the United States who have the kind of commercial driver's license required to drive a truck as there are of working truck drivers... it's 'cause they burned out”
- Because warehouse turnover can reach 100% per year and there is a finite number of working-age people, raising wages is no longer sufficient to recruit and retain workers; when organizing workers consistently say their issue is working conditions rather than pay or benefits, a tight labor market will eventually force employers like Amazon to improve conditions because they otherwise cannot run the business.
“If you have a 100% turnover per year which is not at all atypical for these warehouses, you need to figure out how to hold on to those workers longer”
Last-mile delivery conditions improve when workers unionize or gain structural power to resist unsustainable work design.
- Citing Matthew Yglesias, a paradoxical case can be made that a dominant company like Amazon owning its entire supply chain would be easier to regulate than today's hyper-fragmented industry, since trucking's fragmentation has not helped its drivers, and—given a functioning democracy—a single large entity makes setting fair rules more feasible.
“Sometimes I wish that a company like Amazon would become so dominant because then it would be easier to regulate this very fragmented industry of supply chains... trucking is super fragmented and that hasn't helped those truck drivers at all.”
- Last-mile delivery driver conditions diverge sharply with unionization and corporate structure: century-old, Teamster-unionized UPS practices 'business unionism' with conscientious safety training, while FedEx uses non-unionized sub-contracted delivery companies and Amazon copied that franchisee model, responding to safety pressure by adding in-cab surveillance cameras rather than lowering the pace of work.
“If you look at a company like UPS... it's unionized. It is what academics call business unionism... UPS is pretty conscientious about training their drivers to be safe... FedEx is all sub-contracted delivery companies, so none of their drivers are unionized. And then Amazon just copied FedEx's model”
- California passed a bill limiting the quota systems Amazon uses in its fulfillment centers, and because Amazon cannot simply exit the California market, the law will have a significant effect on warehouse working conditions there.
“California just passed a bill that limits the quota systems that Amazon uses in its fulfillment centers. And Amazon can't just pull out of California, so that's gonna have a significant effect there.”
- Whatever one's preferred mechanism (universal basic income or government subsidies), the most effective and fastest lever to raise wages and reduce work hours is making it easier for workers to unionize again, since unionization is historically how the 40-hour work week and rising wages were achieved.
“my favorite thing is we really should change our labor laws to make it easier for workers to unionize again, 'cause I'll tell you what, that's how we got the 40-hour work week, that's a really fast way to get wages to go up so that people don't have to work as hard.”
Supply chains function as decentralized biological networks with no top-down coordination, creating structural inefficiencies and fragility.
- The physical flow of containers mirrors the US-China trade imbalance: full containers flow from giant Chinese ports to the US, but because much of what the US sells back is intangible (intellectual property, software, cloud services) that needs no container, empty containers accumulate in the wrong places and there is no top-down authority to reposition them efficiently, creating a structural inefficiency that worsens shortages.
“what we're selling back to China is a lot of it's like intellectual property, and software... things that do not need to go into shipping containers, so then the problem is, Well, you got all these empty shipping containers like who's gonna ship them back”
- The global supply chain is not planned top-down but functions like a biological network of fungi, where distinct self-contained segments (ocean shipping, trucking, e-commerce) act as separate competitive marketplaces that communicate only at handoff points like ports and warehouses, with no single actor having cornered any market.
“it's not top-down, the different segments communicate with each other at the point of a port or a warehouse or a truck, as goods are moved from one to the other. But each one of them is kind of self-contained, and it is also a distinct marketplace.”
- The spot-market cost to ship a container across the Pacific from a major Chinese port to the US West Coast rose from roughly $2,000 before the pandemic to $16,000-$20,000, a roughly tenfold increase, making air freight a competitive but constrained alternative.
“In the before times, it would cost you maybe $2,000. Now on the spot market, it'll cost you $16,000, $20,000.”
- Even as Chinese wages rise, manufacturing stays in China because making complex modern goods requires accumulated expertise that pools in unexpected places (e.g., the world's microchips are packaged in Malaysia); manufacturers can only relocate one piece at a time starting with the least complex (hand assembly), which paradoxically lengthens and complicates supply chains and increases their vulnerability.
“there is a stickiness to things still being made there because it requires expertise to make many things in our modern world... they have to do it one piece at a time, starting with the least complicated part”
Automation changes job nature rather than eliminating work; productivity gains expand human material demands across history.
- De-skilling is taking the knowledge of a skilled human, embodying it in a piece of automation (which need not be very smart), and thereby making a once-tricky job easy enough for any new worker—seen from the Luddites' automatic looms to the automated fast-food grill; this makes recruitment trivial, which removes the incentive to retain workers until the employer runs out of new people to recruit.
“de-skilling is essentially when you take the knowledge of the human, you embody it in a machine, a piece of automation... when you de-skill a task, you make it very easy to recruit new workers which gives you less incentive to retain them”
- Automation does not eliminate work but polarizes the labor force, hollowing out the middle class by pushing many people down into algorithm-managed jobs while pushing many others up into creative, problem-solving roths that machines cannot do; the durable way to 'future-proof' oneself is therefore creativity, problem-solving, and education.
“what we're seeing in the nature of work in general is this kind of polarization where... the American middle class is getting hollowed out... But also many, many people have been pushed up that income scale and are doing jobs that no machine can do.”
- Across history the prediction that productivity gains would shrink work has always been wrong because human demands continually expand—keeping the productivity of 1950s living would allow a three-day week, but we instead built more complex material culture (dialysis machines, mRNA vaccines, the internet) whose embodied energy and complexity represents extra labor; machines therefore do not take jobs but change the nature of jobs for good or ill, giving us the opportunity to bend that change toward humane, dignified work.
“if we all were content with leading the life that our ancestors had in the 1950s, given the increase in productivity since then like, 'Yeah, we could all work like three days a week.' But we wouldn't have kidney dialysis machines or MRNA vaccines or the Internet”
Modern consumption patterns shifted demand toward goods requiring complex global supply chains vulnerable to disruption.
- Economists call it 'multiplexing' when one good is shared and reused efficiently (like an office coffeemaker among many workers); the shift to working from home broke that multiplexing, requiring a separate good for each newly-isolated person and thereby multiplying durable-goods demand.
“Economists call this multiplexing, when one good can be used over and over again... when we were all going to offices and sharing a coffeemaker, we were multiplexing that coffeemaker.”
Harbor pilots exemplify how supply chain safety depends on human expertise and personal risk-taking often invisible to consumers.
- Nearly all goods you own depend on a harbor pilot who risked their life: to board a moving container ship they must time an ocean swell and leap from a tiny pilot boat onto a rope ladder, where falling means near-certain death, giving harbor pilots roughly a 1-in-20 chance of dying over a career, and they steer a skyscraper-mass vessel into a port with tolerances of about a meter.
“over the course of a harbor pilot's career, there's a one in 20 chance that they'll die on the job.”
- Hunter-gatherer societies were the 'original affluent society': a !Kung San tribesperson needs only about four hours to gather a day's calories and do chores, spending the rest socializing—suggesting modern wealth and productivity could be redistributed to let everyone work less, restoring civil society whose decline (as in 'Bowling Alone') stems partly from people being too busy.
“hunter-gatherer societies are the original affluent society. The number of hours it takes a Kung San tribes person to gather enough calories for the day and do their chores or whatever is like four, and the rest of the time they're gonna sit around chewing the fat with their neighbors.”