The Droids We’re Looking For

Honestly, I haven’t yet read Thomas Piketty’s best-selling economic text Capital in the Twenty-First Century, but I am familiar enough to believe in its main thrust: that wealth is becoming more concentrated in fewer and fewer individuals, and that sustained inequality is bad.  His basic thesis, as I understand it, is that capital sees a higher rate of return than the overall economy, so those who own capital (e.g., employers, or, more classically, the landed gentry, whose wealth may be appreciating in everything from real estate to NBA franchises to intellectual property to equities markets) will see their wealth increase faster than those who earn wages, which are tied to macroeconomic growth.  (Notably, he points out that this wasn’t true in the first half of the 20th Century as a result of two world wars destroying so much capital — if your money is tied up in an English tank factory, and that factory gets bombed by ze Germans, you’re not seeing huge returns on your investment).   It’s not 100% clear that this argument holds weight, due to some controversy in his research, but it does seem conclusive that inequality is rising right now, and will be for the foreseeable future.

I, however, have a different argument for why that is the case.  You guessed it: robot utopia.

Bender_Rodriguez
Or as my dad would say, “Ro-butt utopia”

The basic premise of my theory comes from the fact that robots (or, more generally, automation) are gradually replacing the labor force, leaving more people without steady income and a concentrated few white-collar automation magnates with the majority of the wealth.  There is a lot of debate about whether robots are really taking people’s jobs; in the short term, it seems like increased automation has actually produced jobs, as unemployment unequivocally fell through the first computing boom of the ’90s, for instance.  However, the types of jobs have changed; in particular, we’ve seen a significant decrease in skilled manual labor as manufacturing jobs have moved overseas or been replaced by machines.  The workforce has shifted dramatically away from manufacturing and toward service as a result.  (The immediate cause has more to do with shipping manufacturing overseas, but the reason that this is viable is that in places like Bangladesh, people can legally be treated like machines; in they future, when the people of even the least developed nations have been afforded basic human rights, multinationals won’t pay thousands of people $10.50 / hour plus health insurance and benefits with a 5-minute break every hour and an 8-hour workday in a well-ventilated and ergonomically designed factory to sew our underwear — they’re going to employ three people at $10.50 an hour to supervise the machines doing it, because machines don’t need wages, or benefits, or breaks, or workdays, or ventilation.)  This is exactly what’s already happened in the domestic farming industry, where we now have two or three guys with GPS-guided machines that plant and harvest crops on land that once would have employed hundreds — and even though it displaces agricultural workers, this is a good thing, because humanity can feed and clothe itself more efficiently without exploiting the lower class through back-breaking labor (the same reason it’s ridiculous that an advanced, egalitarian society would build an army out of clone humans, when they can obviously build a droid army instead.  “They’re super loyal and they grow up in 10 years instead of 20” isn’t a convincing argument when you can build a totally programmable droid in 5 minutes and you don’t have to pay it a pension when the war is over, not to mention the moral atrocity that is breeding a human for the explicit purpose of having it get blown up).

Nothing runs like a deere
Pictured: Efficiency

The argument espoused by the anti-robot-utopians is that, for every job that is destroyed by automation, another is created, and to date that has been more-or-less true (there is an argument that, with a decreased workforce and increased automation after the financial crisis, companies were able to sustain profits with fewer people, which is why we’ve seen sticky unemployment numbers even as we have recovered from the recession).  But this has been true solely because there are un-automated industries left for displaced workers to go into — and typically, these industries pay less, hence rising inequality over the last few decades.  When farming automated, people were able to go into manufacturing; as manufacturing has automated, people have gone into service industries (whether consulting, marketing, or waiting tables).  Many older employees have been forced out of the workforce entirely, while the younger generation has reaped the benefit of the new booming automation industry.

As someone who firmly believes in both the human race’s ability to develop powerful technology and the overwhelming desire to act in one’s own best interest, the idea that one day most jobs won’t be automated is short-sighted and fails to reflect the boundless potential of automation technology.  Let’s say you’re a restaurant owner.  You’re faced with the choice of hiring a human worker, to whom you pay $2.17 an hour plus the hidden cost of adding 18% to customers’ food bills, and who gets health insurance and a 15-minute break every 4 hours, and who might provide excellent service, but might also mis-enter an order, or get into a fight with another waiter, or miss a shift; alternatively, you can buy a $100,000 automated system that requires monthly maintenance, but it never takes a break, doesn’t need health insurance, doesn’t get tipped, never mis-enters an order, never fights with the other robot-waiters, responds immediately to customers’ needs without needing to be flagged down, can serve twice as many tables, and never misses a shift.  Which do you go for?  You will buy the robotic system; you’ve decreased the cost to customers (their food is a bit more expensive, because it includes service, but they don’t have to tip), you can stay open twenty-four hours for the same cost as staying open 8 or 16 hours, you’ve increased the baseline level of service, and you’ve mitigated the risk of poor service at a relatively flat cost, so you’re acting in your own best interest to buy it, regardless of how many service-industry employees you put out of business.

robocop
Depending on the robot, you can also decrease your dine-and-dash risk.

And that’s just one example; now apply that to everything from cab drivers (the self-driving car is probably less than 10 years away from viability) to consultants (imagine Siri or Watson in 50 years), and tell me what options are left for the majority of the workforce.

If you’re thinking of huge productivity gains as our existing workforce transitions to managing the automated one, think again.  While this is the case now (and is a principal cause of the rise in productivity over the last two decades), automation is still a young industry, which means that many other industries haven’t been automated, but it also means that the automation industry itself hasn’t been automated. This means that, for every automated task, we still need a human to oversee that automation.  In the restaurant example above, we would have a manager of some sort to ensure that nothing was going wrong — customers understand how to use the system, it hasn’t broken down, the robot is refilling people’s drinks correctly, etc.  We still have humans on automated factory floors ensuring that the process is working as intended.  My favorite automation project, the self-driving car, will probably need to have humans who can override the automation in emergencies for at least 10-20 years before the kinks have been worked out and these machines can be trusted. But as the technology matures, we won’t need human oversight anymore; automated restaurants and self-driving cars won’t seem like risky gimmicks, they’ll be commonplace, because the systems will perform unequivocally better than a human could.  Bugs will eventually get worked out to the point where the technology is mature, and the automation oversight process itself will be automated — if something breaks down, it will eventually be automatically detected and dealt with.

terminator
Eventually they’ll have to automate the process of firing people who have been replaced by robots.

So where does that leave people?  For the most part, real “work” as we know it today will be done for us; we won’t need jobs as we do now, because those jobs are based on the premise that you produce something for society, and society rewards you with the ability to purchase the products of its other members, but we won’t be able to produce anything for society as well as a machine can.  People will have time to focus on whatever most interests them, whether that be art, learning, or sloth.  But at the same time, we’ll still live in a world of limited resources; classical economics doesn’t just go out the window, and there will be only a small group of people who produce anything of value to society — principally, the people who own the automation-producing capital.

This is the singular crisis of the 21st Century, to me.  If done correctly, it can lead to the robot utopia; if not, it could lead to inequality, poverty, and ultimately the demise of civilization as we know it.  To do this correctly, society will need to acknowledge that we are free enough from want that we can support non-productive human life (or potentially redefine productivity).  It is not reasonable to expect that every member of society will be qualified to work on the increasingly complex systems we will have in place, and if we do not accept a comfortable minimum quality of life for such people, chaos may reign (whether it be from class warfare, subjugation of the masses, or the ultimate demise of the lower class, whose existence will be driving the need for the automation barons’ products).  At the same time, while no one person needs to work, there will likely still be a need for people to work, and we must continue to incentivize this work in order to allow society to advance rather than falling into a classic tragedy of the commons scenario.  Why would anyone risk their capital or work toward innovative solutions to society’s inefficiencies if they weren’t properly rewarded for it?  If we cannot incentivize the 1% who continue to produce, we will stagnate or revert as our systems break down around us, because it will always be somebody else’s problem.

Here is where we return to Piketty’s ideas about inequality.  We are left in a world where a tiny few have everything, and everyone else has nothing — so how do we get around it?  Piketty suggests a wealth tax — tax the wealthiest cut of society at a rate that still incentivizes them to gain wealth, and use the money from that tax to instate a safety net for the remainder of the population.  Note that this is not communism, nor even true socialism — there is no collective ownership or central planning agency, and the goal is not income equality, which would disincentivize  the wealthy from producing.  Instead, the idea is to provide everyone with enough capital to live comfortably (an amount that hopefully grows over time, as society continues to innovate and improve), and which they can choose to spend freely, while maintaining upward mobility through the possibility of investing that capital in themselves or production of a different kind (e.g., art).  The rich will still be rich; they will have more than the average citizen, which provides incentives for average citizens to want to become rich themselves, and those who do not possess the necessary skills to continue producing have the choice of investing in themselves to gain them or to continue living comfortably, but not richly, in society.

This is just one of many possible solutions, and there’s obviously a lot that has been left unexplored here.  It would be impossible to say how society will advance over the next 50-100 years, both domestically and globally, and there are an infinite number of possible futures that don’t involve a robot utopia.  At the same time, there is no doubt that automation has the potential to radically change the world that we live in, and if it does, we will need to address those changes in a way that pushes society forward, not backward.

2 thoughts on “The Droids We’re Looking For”

  1. Oh man tackling dys/u-topian future and its relation to automation? Some brave new world shit.

    I used to think about this a lot. My very first class in college was, I shit you not, “Technical Visions of a Utopia,” and it was basically reading and watching a lot of different views (usually dystopian) of the future and how technology will play a roll. By the end of the class, my conclusion was, “Well… it’s hard for me not to see us not being totally fucked in a couple years, but people have been saying that for centuries, so I guess it’s best to not think about it anymore.” And I try not to. Still, since I don’t really want to comment on cruises and Europe, and I’m itching to deposit my two cents SOMEWHERE, I’ll think about it once again.

    So, yes, let’s think about advancing robots to the extent of our imaginations. And let’s try to avoid communism/socialism. What’s left for humans?
    -Rent. They’re not making any more land, with the exception of Dubai and the occasional volcano, so rent will continue to be a way for people to make money. And it’ll also be probably a big expenditure for most everyone.
    -Entertainment. People will, presumably, always want stories. And I doubt All Of My Circuits will really come to fruition. I think the idea of humor and irony is probably a harder thing to program into a computer than love, so we’ll need writers, thank god. Also professional athletes. We’ll keep paying those guys a ton of money, since the human flaws are the only things that make sports worth watching, let alone not-ties.
    -Risk. People will likely always have SOME form of risk in their lives that they will want to mitigate/be compensated for. Wrapped up in this point is a very curious one in the form of the stock market. What happens when HFT becomes so “efficient” that it doesn’t make sense for people to invest anymore? When it’s only robots? The eventual conclusion, I feel, is that one will be programmed better than the others and will actually “win” the stock market, especially if all of the decisions of the companies traded are made by Watson/Siri as CEO and CFO. With no human error or Steve Jobs-deaths to fluctuate the market, I imagine it becomes actually pretty easy to solve. Of course, I don’t know when we’ll let robots become CEOs. As a society, we’ll be much more comfortable being driven around at 100 mph by robots than by leading a company we’ve invested $10k in, I’m sure.
    -Energy. Like land, for now, a big issue. When nuclear plants started popping up after WW2, people were heralding it as the final solution to energy, that electricity would become “too cheap to meter”. That obviously didn’t happen, but it’ll need to, eventually. Probably not in the form of nuclear, but we both know too much about this area to get into a discussion in the comments on a blog about it.
    -Information. I think of this as the combination of Rent and Energy. The seeds of this are beginning now in the Net Neutrality ordeal, and like any attempt to control the “press” in the past, it will eventually ruin those who try to suppress it. There may be dark days before it happens, but eventually, information wants to be freely spread. If the CONTROL of information becomes automated, however, and there’s no one real with real stakes, then information could be halted, I imagine. Damnit, I think I’m moving down the dystopian future-hell-scape again.
    -Crime? Probably not. I imagine we can become very good at stopping crime.
    What else? That’s not enough to sustain a society, is it? Esp not a despotic one. There’ve gotta be other things humans will control in the very long term.

    So here I am again. Standing on the cliff of this enormous, stupid mystery. Starting down this rabbit hole is always a mistake. So I go back to what I said at first. I think it’s dumb to think about this. At the end of the day, our only hope is that Google really does care about people. So we’re fucked. Of course, people have said the same about the East India Company, Standard Oil, General Motors, and Microsoft in the past. So we’ll be fine.

    1. I think there will always (or at least in our lifetime) be ways for d00dz to make that green. And I agree that, as much as we look at this now as a “Oh no the world’s gonna change!” thing, any change that actually happens will roll out slowly, and will be dealt with as it rolls out, so that every year we’ll be saying “in 10 years we’re gonna be so screwed!” but it’ll never actually happen. Although one could also make an argument that some sort of long-term plan / foresight might ease friction throughout the process and point us to a more efficient long-term solution… Perhaps this is a job for my friend CentralPlanningBot?
      Anyway, in response to a few of your ideas re: stuff people will always do, I have some rebuttals (largely because this post is about how everything will be done by robots, so I have to find rebuttals, even if they are ridiculous).
      Rent: this sort of furthers the argument that capital accumulates, promoting inequality. Sure, those who own the land will grow rich off of the rents from it, but those who have to pay those rents have to get ducats from somewhere in order to pay them. Probably they won’t get it from land-owning, y’knowhadimean?
      Entertainment: I agree to an extent and I think that art (whether written, drawn, carved, performed, etc.) will provide a good market for this sort of thing … but we already live in a world where computers write news stories that are basically indistinguishable from human-written stuff, and one of the most-watched shows on TV is Two and a Half Men. Are you telling me that’s not written by a robot?
      Risk: I think the market is already controlled by robots now; the only thing left is to automate the automaters (tell me you weren’t an automater when you worked in the mines in DC?) I don’t think they’ll ever “win” the stock market, because they’d have to literally be able to predict the future. Although I do agree that, at a hyper-automated level, you’d basically end up with a bunch of machines running science on each other and going with the thing that has the best change of being right, which would probably converge. Who knows?
      Energy: We will need to continue to innovate in the energy department, especially in the near-term (although perhaps one day we will find the golden bullet). However, in the energy sector, most of the dangerous jobs (like mining coal or huffing natural gas to determine its purity) will be done by robots within the next 50 years. There’s no reason to risk another chilean mine collapse when you have cheap labor that can work 24 hours and as a bonus, if the mine collapses on it, it can mine its way out. Here you’re left with maybe 1-5% of the original industry, with technocrats inventing new energy magix or those supplying the capital (land / machinery / refinery / etc.), who in my narrative above are among the haves.
      Information: Your narrative is terrifying.
      Crime: I like to think the Crimebots and the Copbots will manage to keep each other in check.
      Anyway, I agree — it’s probably not enough to keep society running, specifically because the only stuff left that humans can do is a tiny little subset of all the stuff that produced money, which is why I suggest we just embrace it rather than fight it. As huge swaths of the workforce get replaced, we should basically make the people at the top who replaced them support them — wealth tax! Then those people can choose to spend “their” money on whatever goods and services are out there being provided by the technocrats and the artists and the capitalists who provide for society. Otherwise I think we head down a particularly dark path.
      But as you say, we’ll probably be fine.

Leave a Reply

Your email address will not be published. Required fields are marked *