Thank You Very Much, Mr. Roboto

“Hey Philip, this is Karen from San Francisco. I’m learning loads from this podcast, greatly appreciating your insights — there aren’t many unapologetic examples of people not being defined by their highest-paying jobs or their highest degrees.

“I’d be glad to hear your exploration of the less-examined aspect of automation and the transition to the future of work. Namely: what about the Sashas?”

You might recall Sasha from a few episodes ago, the worker who deliberately flies under the radar to do the absolute bare minimum based on whatever performance metric you’re measuring.

“Much of today’s conversation focuses on upskilling, coding camps, and the like, but what of the required shifts in metrics, incentives, and perceptions? What do our labor force and society look like if all, some, or none of the Sashas are left behind?

“Thanks for sharing your experiences — looking forward to the next episode.”

Karen hits at the heart of a key concern of mine over the past few years, and especially in my time at Amazon: what seismic societal shifts will result as our regulations, laws, and cultural habits are outrun by the ever-increasing pace of technical innovation? Are we going to experience slow, oozing lava that depressurizes our country, or are we building towards a magnitude 9.0 earthquake that will suddenly destroy and reshape everything? We’ll explore all of this in today’s episode. Though I certainly don’t have all the answers, it’s increasingly important that we at least begin to have this discussion as a nation.

Welcome to Peak Salvation. Today’s episode: Thank You Very Much, Mr. Roboto.

There’s a certain sort of poison that every employer deals with, the bad apple which often can go unnoticed. I was once working with Carla, who was transferred over from Inbound to help Outbound Ship Dock for a few days. Having Carla on my spur was at first exciting. She had a youthful energy, bobbing and bouncing in rhythm to music I can only assume was playing in her head, scanning and sorting boxes with a vigor rarely seen amongst my coworkers. By way of explanation, when I asked her about how she liked her Amazon job, Carla simply said she hates being idle.

As the morning wore on, I shifted from being enthused about having such a productive coworker to being slightly concerned that it seemed Carla couldn’t contain her energy, and instead needed to always be dissipating it in some manner. In moments when the spur slowed a bit, she would proactively start closing carts that were near full. The metal doors on Amazon carts swing freely on hinges unless you keep them open by using three-inch long white plastic hooks built into the side of each cart expressly for that purpose. The common practice when moving an empty cart is to make sure its doors are latched open by these hooks so that you don’t accidentally swing a free door into someone else as you turn the cart. These hooks, unfortunately, were designed in such a way that they can often be very difficult to unhook. In the worst case, I’ve spent up to perhaps a minute kneeling on the ground next to a cart door, left hand tugging on the thick plastic hook as my right hand tried to coax the mesh of the door to yield in the opposite direction.

But these issues don’t worry Carla. Instead of unhooking cart doors, Carla would just put her entire weight into suddenly tugging the door of a full cart with both her hands, jouncing a few times until the plastic hook broke to pieces and clattered across the warehouse floor. Just the day before, Angelo had warned the entire team that some broken pieces of cart hooks had caused an associate to slip and fall — hard plastic on polished concrete can feel like ice skating. But this wasn’t Carla’s concern. Throughout the day, she kept breaking cart hooks. The double whammy, of course, was that once the cart was unloaded by some associate in a different Amazon warehouse, perhaps halfway across the country, the cart could no longer be safely rolled around and would instead need repair.

And herein lies the difficulty that every manager knows: you’re typically not tasked with finding a thoroughly rotten apple, which even a junior manager knows how to deal with. You’re usually instead faced with the dilemma: Carla doesn’t shirk work and out-produces most associates. At the same time, she willfully destroys company assets that may or may not cost more to repair than her increased productivity over others. Finally, consider the fact that you operate within a management system specifically designed not to differentiate compensation based on performance: there is neither the carrot of coaxing Carla into better behavior in pursuit of faster career growth or a bigger bonus, because neither possibility exists, nor the stick of threatening Carla with a safety or quality demerit — she was loaned into your department anyway, and would happily be locked back in to her previous job by such a demerit. In the end, you have only this: accept Carla because she ships more in boxes than she breaks in hooks, or fire Carla because the opposite is true.

White-collar work has this in spades, at least in the tech world. I’ve managed many teams in my twenty-three-year career, and there’s always at least one person on every team who poses this very question: is the juice worth the squeeze? I’m an amazing programmer, but I also pour cold water on almost any new idea. I have brilliant ideas, but I also occasionally blow up in meetings and shout ad hominem attacks at whoever’s closest in the conference room. There was once, if I remember correctly, a very talented Microsoft developer who, when walking out of a team meeting where it was announced that the team would soon be growing quite a bit, said to everyone within earshot, “You know what else grows? Cancer grows.” Microsoft was famous, especially back in the ’90s, for tolerating all sorts of misbehavior in order to retain exceptionally talented employees. It was commonly said of one hard-to-get-along-with vice president in Windows that it was “better to have him pissing out of your tent than to have him outside pissing in.” The tech world is filled with examples of this, of prima donnas tolerated for their commercial value.

Managers are faced with very difficult tradeoffs all the time around the net value, positive or negative, of certain employees. The answer’s not always clear. Watching Carla and similar associates, I felt an even greater appreciation for the difficulty of managing at an Amazon warehouse. On one hand, there’s the ease of simply letting your computer tell you which employees are falling below their quota and threatening to fire them; on the other hand, you run into cases where your hands are tied, where common tools like performance incentives and the potential for better career growth don’t exist. It’s all enough to make a longtime manager like me wish there were robots: workers who don’t need incentives in order to do the right thing, workers who don’t present the type of tradeoffs posed by temperament.

For decades now, people have been saying that robots will take our jobs. It seems that with the inevitable forward march of machine learning and better robotics technology, more and more people will be displaced from their roles or replaced entirely. When computers first started playing chess, most people believed that chess was simply too complicated for computers to ever win, that the computational complexity of chess made computers inferior to human intuition and experience. When IBM’s Deep Blue subsequently beat then-reigning world chess champion Garry Kasparov in 1997, naysayers moved the goalpost. It would seem, after all, that perhaps chess was beatable by computers. But surely Go — a game far more complex than chess, requiring creativity and ingenuity, the types of traits only associated with humans — surely Go would prove computationally impossible for a computer to win. But in 2016 a computer did indeed win. Called AlphaGo, the program created by Google’s DeepMind division definitively beat world champion Lee Sedol. Once again, the goal posts were quickly moved. Skeptics have now stopped claiming computers will never beat humans at complex games, moving on to claims around new qualities seemingly unique to humanity: creativity in art and music. Surely this time, they say, we’ve at last found something where computers cannot beat humans.

I’m of the other persuasion. I’m convinced that since computers evolve much more quickly than humans (through human innovation, granted), and their evolution is strictly forward (that is, once a skill is gained by a computer it is never lost), it’s all but guaranteed that with each passing year, they’ll be able to do more and more of what, traditionally, were jobs only humans could do. Skeptics bring up examples like farming to bolster their claims that displaced people simply find new jobs. One hundred years ago, 25% of the US labor force worked in agriculture. Today that number is 10%. But it’s not like 15% of people are simply unemployed, having been kicked from farms into the streets. Instead, the story goes, they’ve moved into jobs that didn’t exist 100 years ago, like social media consultant and hedge fund manager. People who doubt computers will ever replace us call me a Luddite and say that people displaced this decade by robots will simply go do other jobs.

But I think this time things are fundamentally different because computers and robots are shifting the nature of what’s required to remain employed. When machinery first replaced agricultural workers, there were plenty of other jobs that didn’t require any additional education. A hundred years ago, the college graduation rate was 5.7% — meaning over 94% of US jobs didn’t require a college degree. In 2020, the percentage of American adults with college degrees is just under 40%.

When I lived in London next to Westminster Cathedral, there was a McDonald’s next to the cathedral that was extremely busy with tourists, especially during lunchtime. It shut down for a few weeks, and when it reopened, all the cashiers behind the counter had been replaced with computer screens that let people order and pay on their own. This made the lines much shorter and greatly increased throughput. It also notably eliminated a bunch of jobs. That McDonald’s used to have six cashiers, each operating busily during lunchtime. The important question now is whether there are six extra jobs that opened up elsewhere in the economy that don’t require college degrees. If, instead, the new jobs created, like social media marketer, require college degrees, then these cashiers would first need to get degrees. And it’s not clear to me that everyone has the capability or desire to get a college degree.

On top of that, consider one thing Karen’s question at the top of the episode touched on: what would the future look like if we upskilled or retrained everybody? A good friend of mine once stunned me into silence with a profound question: what if 100% of the country’s adults suddenly, right now, all had college degrees? Will they all be high frequency traders or social psychology professors? Truth is, there’s a mountain of work that needs doing, and the shape of the mountain is shifting.

Right now, we see compensation at the top going up, so we understandably think we should try to pull everyone up. But two things: if everyone could code, and extremely well, those jobs would instead pay minimum wage — because, as I keep reminding people, you’re not paid according to your contribution to a company; you’re paid your replacement cost. And at the same time, how much would McDonald’s jobs need to pay in a world where everyone has a college degree? They’d clearly have to pay more because everyone would have plentiful job options, which is all well and good, except what does it feel like to work at a McDonald’s when you have a bachelor’s degree in civil engineering? I’d hazard a guess it’d feel exactly like working at an Amazon warehouse when you know you can launch mobile apps yourself, lead teams of hundreds of software engineers, teach at multiple universities, and sport a fabulous flattop that instantly dates your hairstyle to the ‘80s. Depressingly, when we retrain people, could we inadvertently be setting them up to be Good Will Hunting, sweeping floors, grossly overqualified? (And it's a realistic worry. Despite what the popularity of Roombas, not to mention Rosie from 'The Jetsons,' might have led you to imagine, janitors won't be automated away anytime soon. Experts think generalized cleaning jobs will still have some longevity because they involve a lot of dynamic dexterity that robots currently find difficult.)

The good news is that we’ll be OK in the short term. As long as college-graduate work continues to pay significantly more than non-graduate work, the market is saying we need more college grads. But don’t harbor any illusion that the gravy train runs forever — remember that the endgame cannot simply be that we’ll all live well if only we all had degrees. Someone needs to sweep the floors of MIT. And the more qualified that person is, the more dissatisfied they’ll feel about their job.

Then again, even college graduates aren’t safe from the juggernaut of technological encroachment. There are startups now whose entire deal is filing all the legal papers needed to get your citizenship. These are lawyers’ jobs, where each filing used to cost an immigrant over $10,000. But it turns out the job was easily automatable, consisting mostly of form-filling and knowing where to send things. Those lawyers now need other jobs. Perhaps they, too, can become social media marketers.

Highly technical jobs also aren’t secure. Sure, you’ll read plenty of articles about how tech firms can’t hire enough software developers; it’s also clear that software truly is “eating everything,” as Marc Andreessen predicted years ago. But what’s not clear is whether humans will always be the ones to hold those jobs; and even if so, whether the bar will be raised impractically high for most people. As the industry matures, fewer companies at the top will gobble up all the elite talent and create huge platforms that are hard to replicate or displace. Lesser programmers will make decent money for a while, until two things converge to form an event horizon from which they’ll never return: the top companies will eventually launch products and services that obviate the custom, manual work that lesser programmers currently do, and AI will increasingly write code better than the average lay programmer can. GPT-3, an AI released last year by OpenAI, can already complete many lines of code, and sometimes even entire functions, once you’ve typed the beginnings of it. And just two weeks ago, Google’s DeepMind division announced that their coding bot, AlphaCode, has been quietly beating 54% of software developers in online coding competitions. How long will it be before AI begins at first fixing bugs, then writing code once you tell it what the aims are, and finally creating useful things entirely on its own? Sound impossible? Like, perhaps, beating humans at the creative and incomputable game of Go?

But let’s suppose AI doesn’t, or can’t, replace people when it comes to code. Let’s suppose the concept of programming is a skill uniquely human, no matter how good computers get at other things — that the best of computers might eventually win all coding competitions but still be unable to write commercial software. There’s still the question of whether everyday people can do the job. Consider that all the top tech companies claim they can’t hire programmers fast enough. Consider, also, the tremendous amount that software programming pays, when even average college grads with no professional experience can be paid $100,000 or more a year, 48% more in their first year of work than the median American family makes. Consider, finally, that software programming continues to not require even a GED; some of the best programmers I’ve worked with over the years have not finished college, and occasionally some haven’t even finished high school. So how is it that a job that doesn’t require even a high school degree, paying far more than most jobs starting day one, isn’t something that’s oversubscribed, something that has way too many people training for it? How is it that the top tech companies, while claiming that they can’t hire enough people, reject dozens of applicants for every single hire they make?

I’d posit this is all because of the rising tide of basal talent required to participate in our new winner-take-all economy and what it means for the type of talent that can continue to be employed. Replacing the drummer for your no-name garage band is easy. Replacing a member of K-pop band BTS is going to be extremely competitive because of their global reach — the same reach that also limits the job to one opening worldwide. Sure, there’s a bunch of talk about the “creator economy,” no doubt fueled by the select few journalists who’ve established workable Substacks and the handful of superstar success stories, but the median Twitter personality has 200 followers, and the median Instagram account has 150. The average video in YouTube’s most watched category, Entertainment, garners just under 10,000 views, earning, on average, $5.00 a video. I’d need to release an entertainment video every seventeen minutes, or forty-one videos a day, to maintain my Amazon warehouse salary as an average creator.

All this brings up the question of when we might approach Peak Programmer, the point in humanity at which the most programmers will ever be employed. As the bar of talent required to make an impact continues to rise, shutting pretty good engineers out of the most elite jobs, AI will cull the pool from the bottom, replacing more and more mediocre engineers entirely. This dynamic, a rising bar at the top and automation chasing from the bottom, will mean that we will not only one day reach Peak Programmer — we should also expect to reach Peak Human, the moment in history when we reach the maximum number of humans who will ever be employed.

–

Computers used to serve our needs. We told the TI-81 graphing calculator to give us the answer to some simple equation, and it did. We told Microsoft Word to put a sentence here or there, and it did. But increasingly, we now serve our robot overlords.

Think about what Uber drivers are really doing. They press a button to tell the Big Computer in the Cloud, that Great Switchboard in the Sky, that they’re ready to drive. The algorithm then tells them where to pick up the first customer. Once the customer boards, software tells the driver exactly where to go, down to even what turns to make, accounting for traffic and other dynamics the driver can’t possibly know themselves. The Great Switchboard even decides when to surge an area, causing dozens of human drivers to change their plans for the day and show up in some random zone in the city. We don’t tell computers what to do now; they tell g.

Or perhaps a more subtle example that better typifies how slow and nearly undetectable our robot takeover is happening: there are apps you can install that use your smartphone’s camera, along with some fancy AI, to not only tell you what type of plant or flower you might be looking at, but also whether it needs things like more water or less sunlight. It can remind you to water different plants at different times, and diagnose botanical problems that arise. At this point, whose thumb is green? Who’s the actual gardener? The AI is basically growing all sorts of plants worldwide, telling its hundreds of thousands of minions each what to do in order to maximize success. The pattern is never going to be that you’re tapped on the shoulder one day by C3PO and asked to step aside so that he can do the diplomatic negotiation. No, it starts with you in the driver’s seat, asking a computer to translate some words … then offer some alternative phrasings … then tell you how previous negotiators have gotten past a certain sticking point … then give you templates for conversations to maximize success. This is the gradient that we are all on: we think we’re using computers, but we slowly find we can’t live without them, and then we finally wake up one day to realize we’re simply doing what they tell us.

My time at Amazon coincides with just one such subtle transition. For most of BFI4’s history, Ship Dock workers manually scanned each package and put it into the right cart according to its destination. My first day on the job happened to also be when they wanted new associates to be trained on a different way of loading carts, called “directed palletization.” You can bet an engineer came up with that name. Palletization is fancy-speak for filling a cart or building up a pallet full of boxes all headed the same place. It’s “directed” because, starting this peak season, Mother Brain would now tell you where to put every package. The way it works is this: you happen to walk under one of several overhead scanners in your spur while carrying a package with its label facing upwards. A computer reads the label even without needing for you to stop, and flashes a green button in front of the cart the box belongs in. Once you’ve placed the box in the cart it told you to, you press the green button so the computer knows you’re ready to move on. This process is not only much faster, because you don’t need to fumble with a hand scanner or think about which carts go to which destinations, it’s also frankly easier. It basically takes the thinking, or, as I mentioned in the first episode, the mistaking, out of the whole process. You pick up a box from the spur — really, any box — and you just start walking, up or down the spur, it doesn’t matter. A green light soon comes on, which you should head directly towards like some sort of Nick Carraway towards Daisy’s dock in The Great Gatsby. Then plunk down the box, tap the button, and repeat. At this point, why am I even in the picture?

The same dynamic can be found writ large when looking at the entire Amazon warehouse. I have no idea what the big picture is, or why I’m doing a specific something at any particular time. In fact, the only reason that pickers exist at all, given that they already interface with completely robotized shelves of goods, is that robotic hands haven’t developed yet to the agility and adroitness of human hands. But do you doubt that’s coming? Or do you doubt that Amazon will fire me as soon as such robotic hands become reliable? They don’t unionize. They don’t require lunch breaks. They don’t create podcasts causing others to consider the arc our society is taking. And they certainly don’t pee themselves, or talk to the New York Times whenever something untoward happens. Every single person in an Amazon warehouse is simply being told what to do by The Great Algorithm in the Cloud, that at first helpful, then eventually fearsome Mother Brain. Even the managers are told, by their little laptops, which employees are underperforming. They’d otherwise have little idea; they, too, are simply doing the bidding of computers running in some dark datacenter hundreds of miles away.

But is it really all bad? I came into this Amazon job with a heart full of concern around technical unemployment, around what I thought was a problem of robots replacing people. While some of the work in BFI4 is both challenging and rewarding, most of it, if I’m honest, is dehumanizing and should be displaced. The only reason that people are involved in key steps of Amazon’s fulfillment process is that robots haven’t yet gotten reliable at some of the steps: the manual dexterity required for picking, the dynamism of handling a large variety of packages adroitly. But most of the work is physically taxing, repetitive to the point of often causing injury, and mentally devoid of stimulation or challenge.

There are indentured servants in Indian quarries whose only job is to break rocks with their hands. No tools. Just taking one large rock and smashing it against another, hoping one breaks before too long. These workers often become multigenerational within the quarry, their elementary-school-aged children breaking rocks alongside them, their entire family crouched in a circle low to the ground around a small pile of stones — their work for the day — because they owe a debt to the owners of the quarry which compounds with interest that grows faster than they can pay off in a lifetime. The sins of the father are visited upon the sons. And so it goes.

This is exactly the work that robots should be doing. I know that jobs in Amazon’s fulfillment centers aren’t exactly breaking stones with your bare hands. Or perhaps I speak to soon. Aren’t they? Repetitive, sometimes dangerous work is being done by humans only because, at the moment, robots cost more for that specific job. In the US, families don’t break rocks in quarries with their bare hands — but that’s because American labor costs more than machines that automatically break those rocks. Not so in India. And not so in many other cases for many other jobs all around the world. I, for one, welcome our robot overlords, if they come in peace to relieve us from these types of jobs. Lives are meant for more than the moral equivalent of breaking rocks all day with your bare hands.

My only hesitation is what happens to the people replaced. On one hand, good riddance if we free people from repetitive, dangerous, unfulfilling work. On the other hand, will our future generate ever more fulfilling work for even larger populations while not requiring increasing amounts of talent or education that more and more are unable to match? Not everyone is capable of becoming an artificial intelligence programmer. And even that job is going to go away eventually. AlphaZero, the AI from Google, taught itself how to beat the world’s top chess champions after simply being shown the rules of the game. Is it hard to imagine the high-paying job of software programmers will one day also be outmoded as computers increasingly teach themselves? I became an Amazon fulfillment center associate because of my concern around technical unemployment, but I’ve found no answers yet.

A future where nearly everyone works at an Amazon fulfillment center seems dystopic — I want robots to free humanity from that. In fact, if we indeed reach Peak Human at some point, Amazon might not only become America’s largest employer in 2022 — it might go permanently in the history books as America’s largest employer ever, past and future, as robots begin to cull its workforce job by job. Worse yet, the higher the wages associates get paid, the greater the financial incentive to Amazon to speed the adoption of robots to eliminate those jobs. Good riddance, yes, because these jobs are hard and prone to injury. But what next for these associates, if even today’s top-paying jobs seem credibly under threat?

And herein is the key question that most people disagree with me on. They say people like me have been around for centuries, and they rightly cite examples like the Luddites, where, in the past, claims that machines would permanently eliminate jobs never panned out. And they’d be right. My line of thinking falls under what’s called the “lump of labor fallacy” — the belief that there’s a sum total amount of work that humanity needs done, and that any bit of work done by machines necessarily means a human somewhere has just lost their means of making of living. Historically, it has never happened that people ran out of jobs because machines took over something they used to do. People have always moved on to other jobs. Farmers moved to cities. Lumberjacks became truck drivers. That sort of thing.

I concede that the burden of proof lies entirely on me because Luddites have been wrong in the past, always. But I think this time is different. (And I’m self-aware enough to know how deluded that sounds to anyone who believes me wrong.) Not all associates can become Amazon programmers. Or perhaps you think they can, and it’s a bad example. Not all lumberjacks can become astrophysicists — I hope that’s not controversial. And my contention is that, as computers get better and better at all sorts of jobs humans currently do, the remaining human jobs will look less like truck driving and more like astrophysics.

Sure, the surplus of our society, if we don’t blow it all up in a variety of wars, will create demand for more astrophysicists at various fancy national labs. But my contention is that, as with what’s already happening with software roles in top tech companies today, those jobs will prove hard to fill not because they aren’t attractive enough, but because too few people can do the job, even given UBI, free college, and perhaps even years of one-on-one tutoring. For every one of us, there is a level of work which is simply undoable. And this time around, unlike any other time in history, the basal level of talent or competence that justifies a salary is rising beyond our capability.

I used to not believe this. I was great at school, easily picking up any subject. Until, that is, I reached Calculus 3 in college. There had, up to that point in my education, never been something I couldn’t understand; any shortfall in grades or test scores was attributable to lack of effort. But Calculus 3 was fundamentally different. For the first time in my life, I could feel that I was on the very boundary of how far my reasoning could stretch, right at the very limit of the level of abstraction I could handle and still understand. I then accepted — truly accepted, not just at an intellectual level, but an emotional level — that there are things I could never understand, no matter how patiently they were explained to me. Advanced physics would never be in my future. Back in Newton’s day, I could easily have been an apprentice and helped him drop a few apples here and there, timing their descent and thinking some thoughts around the speed at which things fall. By Einstein’s time, I already could only be useful in cleaning off the chalkboards and dusting out the erasers. But now, past Stephen Hawking’s era, I can’t contribute to physics in any way. The threshold of talent required to be useful in physics has exceeded my ability, even if I were to dedicate the rest of my life to its study. I believe we each have our own Calculus 3. I believe all jobs have their Newton era, their Einstein era, and yes, inevitably at some point, their Hawking era.

–

In the meantime, levity. Jim, an associate who loved using the misremembered phrase “done and said” frequently in the many stories he relayed about his five years at BFI4, had a good laugh telling me about a robot that was recently removed from Induct #5, where it was supposed to set light packages, mostly bubble-wrap envelopes, label-side up onto a conveyor belt. It felt hilarious to Jim that, when all was done and said, the robot messed up so often and needed so much rescue from nearby associates that it was summarily removed. This, for Jim, was a moment of triumph, proof positive of the folly of hubris. But I wasn’t laughing. Those folks from Corporate, the ones who look silly insisting on wearing blue manager vests while being unable to answer any practical questions on the warehouse floor, weren’t complete idiots. You can bet they didn’t just throw out that robot and repent of their foolishness. No, v2 is coming, and if not it, v3, v4, v-whatever-it-takes — until it’s v-surprise-you’re-fired.

Even if you don’t believe me — and trust me, I have plenty of friends who rapturously tell me that there will be so many amazing jobs, fulfilling jobs that exercise both mind and body, each the right amount, jobs I can’t even dream of — I ask that you at least briefly consider what we should do as a society if I’m right. Will moral opprobrium be justified when even industrious people cannot find jobs, people who are willing to retrain, “up-skill,” or whatever other trendy word buffers its speaker from the reality that not everyone will always be able to do everything? “Ready” and “willing” are but two legs in the three-legged stool of gainful employment. What shall we do then, other than feel a little guilty about judging the unemployable? Shall they live? Eat? Have health care? What about Netflix — is that a bridge too far, given the latest shows aren’t a necessity but a form of luxury? In fact, a luxury which the unemployable would then be enjoying while those still employable, the very ones providing the means by which all others live, are busy at their labors?

The answers aren’t at all easy. I believe we’ll need a universal basic income, UBI, to guarantee a floor of dignity and comfort. I don’t pretend to understand how to thread the needle of the difficult politics around this, especially as my prognostication stipulates a future of “makers” and “takers,” using the latter term here only in its descriptive, not judgmental, sense. But beside the politics of getting it done, and the cultural shift that would be necessary to precipitate it, I’ve also become convinced by my job at an Amazon warehouse that a UBI will be a necessary but not sufficient condition of future human flourishing. I’ll touch on this in an upcoming episode.

Then again, many people have doubts on even the feasibility of UBI, taking an almost Malthusian stance that hungry mouths will simply multiply like locusts to consume whatever we produce, a perspective that “the poor will always be with us.” But a great surprise over the past decades since the Green Revolution made food more plentiful, clean water became available to literally billions more people, and education has been supplied more universally, was that populations actually stabilize in size and, over time, shrink as people rise in their standard of living. It all makes sense biologically, but only in hindsight. When infant mortality is high, you naturally want to produce many more children; conversely, when your environment is safe, and when more economic opportunity is available, you want to invest as much energy as you can into individual children to make them maximally successful. By continuing to improve health and education worldwide, we not only lift people directly out of poverty, but also reduce the total amount of resources humanity needs in order to achieve a solid standard of living for everyone.

This trend, of wealth reducing populations, is in fact so consistent that it’s become problematic in many developed nations, including, soon, our own. The US birth rate is already well under the replacement rate, meaning that, absent massive loosening of our immigration policy or a radical change in reproductive culture, we will have a shrinking population in coming years. This is all fine and good, you might say, especially if you’re the type who loves to be the last one to move into a great neighborhood and close the door behind you. But the problem is that nearly a quarter of the US budget is for social security alone, and another 21% for Medicare and Medicaid; these benefits, disproportionately going to the elderly, together form nearly half of all government spending, funded entirely from taxes. A booming population of retirees, combined with a massive reduction in the working population that pays income taxes, means that we’re headed for a big surprise unless we change something radically. This surprise, for instance, has already hit Japan, causing significant issues in its economy.

It’s unlikely we’ll change the behavior of wealthy countries to reverse their declining birth rate; we could increase our immigration, especially of young working individuals, but we’d also be competing with many other wealthy nations for the same workers; and, if you haven’t guessed it already, we could speed the development and adoption of robots so the labor that our society depends on is retained through a combination of workers and machines.

Science fiction author William Gibson is attributed as saying, “The future is already here; it’s just not evenly distributed.” In the same spirit, as mentioned in a previous episode, there’s plenty of “bounty”; the problem is its concentration, the lack of “spread.” The Milken Institute notes that the net worth of Forbes billionaires quintupled since 2000 while median American family income has fallen. The bounty in America is great, and it’s only increasing as technology fuels production and efficiencies never before possible. We should enact policies that ensure that the spread of our bounty lifts all boats in this country, while at the same time not killing the golden goose of innovation and market competition that has fueled most of the greatest improvements in the human condition over the past several hundred years.

There can be a certain pleasure in work which I hope robots don’t eliminate. When you load boxes one-by-one into an eighteen-wheeler, Amazon has a neat strategy where you build a solid wall out of boxes (as if each box was a brick) perhaps six feet high, across the width of the truck, every three feet or so. As you finish each wall of boxes, using preferably the bigger and heavier boxes to do so, you start throwing smaller boxes over the wall to fill the gap. The wall of boxes thus holds the little boxes from spilling out towards you. You can even throw so many small boxes over the wall that they reach the ceiling of the truck, about perhaps ten feet high. Once the small boxes seem like they might spill over the edge of the last wall you built, you move towards the back door of the truck a few feet and start building your next wall, repeating this process until the entire truck is filled with boxes. It turns out that I really like this task for a few reasons. Sure, it’s a lot of lifting; and in fact, it’s even more tiring to throw hundreds of boxes several feet above your head. However, the work is quiet, because though the truck is docked with the warehouse, you’re sort of technically “outside” the warehouse a bit when you’re deep in the truck. So things are quieter for once. I enjoy the solitude, and the Tetris of wall-building when there are so many different box profiles to account for. I could spend all day doing this task, and it had become one of my favorite assignments in all of Ship Dock. Hard doesn’t have to mean unpleasant. Work is sometimes its own reward, if you could but remove all the baggage surrounding it.

Perhaps the associate who epitomizes this ethic the most, the greatest Amazon warehouse fanboy I’d meet in all of Peak, is the same Jim who thought it hilarious that a robot inductor was more trouble than it was worth at the time. Jim’s the type of salt-of-the-earth worker all too likely to use phrases like “Git ’er done!” He believes in hard work and the value of a day’s wages born from the labor of his hands. I was raking for Jim one day in Induct #5 when he told me all about his experiences at Amazon. Unlike most associates I met, Jim was more than happy to talk nonstop, answering all sorts of questions I had about his experiences at various Amazon facilities. He often used his catchphrase, “done and said,” as a way to conclude a point.

“That’s what I decided when all was done and said.”

“All done and said, he basically had some sort of a problem with me.”

“It comes down to this: when all is done and said, they can just find another job if they don’t like this one.”

Jim was a rabid Amazon warehouse booster, having worked in BFI4 since it opened in 2016. He was proud to show me that five-year folks all have cardkeys where their profile photo is wrapped in an orange border. I’m told ten-year folks have blue borders, and twenty-year folks have black borders. It’s funny how, when you’re one of 950,000 people, the color of a three-millimeter border becomes a valued differentiator. Jim was not only proud of his tenure but dismissive of anyone who didn’t like the job of being a fulfillment center associate. “There’re millions of jobs out there! Heck, if you’re complainin’ all the time, why don’t you jest quwett an’ take one o’ them? Done and said, truth is Amazon’s the best job out there. That’s why they’re here. Whiners.” He’d get worked up a few times about ungrateful associates over the course of the day we worked together, each time putting special emphasis on their option to “jest quwett!”

When all is done and said, that’s not what I remember Jim for. No, no — the far more memorable thing was his dietary revelation, dropped nonchalantly almost as an afterthought in his discourse on working at Amazon Fresh prior to BFI4: “I’ve never ordered from there since.” My ears perked up, especially as I had myself ordered from there many times. “It’s disgusting. I saw them put raw salmon right on top of raw chicken” — here he made a double-hand palm-up pantomime of setting quite an Alaskan catch onto air which I could only assume was uncooked poultry — “just like that! Stuff happened all the time. Things were disorganized — they didn’t even know where to put the, it’s like, the …” Jim’s eyes rapidly scanned around below him as if the misplaced grocery items might somehow have ended up in BFI4. He gave up articulating the specific types of items that were apparently organized quite poorly — somehow so poorly, in fact, that he’d never order from there again — and instead just ended his explanation with a look of disgust, his two hands dismissing the imaginary fish he had just finished putting on chicken. I mumbled my sympathies, relaying my own experience not eating at Subway for seven years after working there. I didn’t tell him I have a good friend who held a fairly senior position in Amazon Fresh, someone I trust a lot, someone who I’m certain wouldn’t write software that made associates mix raw fish with poultry.

If you detected a shift in my bias there, from Warehouse Associate Philip to Software Engineer Philip, you’re not alone. My editor made this interesting observation when she first read the story about Jim: that my inclination, to which I was oblivious, was to feel skeptical of Jim’s story because I happened to know someone senior in Corporate who had worked in Amazon Fresh. Though nearly all my Amazon experiences so far had me empathizing with warehouse associates and questioning folks in Corporate, simply knowing one individual on the other side, the side I had spent twenty-three years living in, snapped me right back to seeing the world through The Man’s eyes, suspecting that Jim just happened to see some highly unusual mistakes. I’m stuck in Season 3 of Lost, living among The Others, confused where my loyalties lie, thrashed easily back into the community from which I came.

In the final analysis, I’m much more likely to be the guy enthused about all the amazing things the latest robots can do, the guy with friends who invent such robots, friends who no doubt don blue vests when entering a fulfillment center to christen their latest labor-saving creations. Whereas there’s no doubt in my mind where Jim’s loyalties lie: he’s a true-blooded laborer through and through, one who’d readily defend his employer against hordes of ingrates (except in matters piscatory). He’d be the one I could honestly imagine shedding real, incredulous tears when, one day at last, some blue-vest-wearing impresario of robotization announces the imminent deployment of RoboInductor v-surprise-you’re-fired. If that day is anything like what happened with the outsourcing of software development in the dot-com days, Jim would be asked by Blue Vest to kindly show his new robot partner the ropes for a few weeks before exiting the building. His bafflement, envy, and anger would be all that remained, once everything was done and said. RoboInductor, on the other hand, would never regale me with stories about Amazon Fresh, never tell whiny associates to jest quwett!, though I’m pretty sure it also wouldn’t feel the type of pride in American industry and joy of hard work which Jim exudes.

Truth is, I’m not just selflessly concerned about Jim’s financial future. I’m concerned about my own — about what history teaches us happens time and time again when one part of society takes off into the stratosphere while the rest barely subsist. Consider even the literal interpretation, going on in our society right now, where Elon Musk is busy trying to launch humanity to Mars. Is there any doubt what type of people will go in those rockets if the Earth becomes increasingly unhabitable? It’s certainly not going to be all of us. And here’s where many disaster movies are spot-on: whenever the rich begin getting into escape rafts and protective bunkers, the resultant anger and rioting always leads to dead rich people. The wealthy seem to never learn this lesson. It’s always I’m smart / they’re lazy; I worked hard / get a job; I earned it / handouts destroy initiative — just more, more, more, until one day, as in societies past, pitchforks.

We’ll soon hear the people sing, singing songs of angry men. Let’s come to the table lest we’re brought to the stake.

Coming next week: Speaking of rich people, I’m invited to the Space Needle for a private event where only multi-millionaires are allowed. There are such things. It’s as surreal as it sounds.