Economist warns of the coming robot apocalypse
November 14, 2013 10:42 PM   Subscribe

The robots are here. George Mason University economist Tyler Cowen predicts that the trend towards automation will squeeze the middle class further still, and compares its effects on American politics to a too-overlooked 1955 short story by Isaac Asimov.
posted by Jacob Knitig (81 comments total) 28 users marked this as a favorite
 
Koch and George Mason University
posted by wenat at 10:53 PM on November 14, 2013 [8 favorites]




remember that riots and protests are typically the endeavors of young hotheads, not sage (or tired) senior citizens

Except in Ireland. Elderly people have been leading the protests and the ones often most active against the government's cuts. On a more general note, these sorts of predictions rely on history unfolding in a steadyish pattern, but much of the past history of revolutions around the world has been that they tend to break out in situations where they weren't anticipated. Afterwards you can go back and connect the dots, but before, for example, the Russian Revolution, no one would have expected that social order to explode the way it did.
posted by lesbiassparrow at 10:57 PM on November 14, 2013 [3 favorites]


Which is a long winded way of me saying that collapsing resources or anything could disrupt this predicted future. Robots take money to build and require a lot of resources to construct and run, some of which might end up being in short supply in the future.
posted by lesbiassparrow at 10:59 PM on November 14, 2013 [4 favorites]


I agree that robots have long supply chains and take a lot of resources to create, just like humans. The difference is that they can be stored in a box, and re-purposed, without having to spend $100,000 to get a new 4 year degree.

Betting on supply chain issues to save human jobs is a bad bet, in my opinion. I think that supply chain disruptions would be an inconvenience, not something which can put humans back to work replacing them.

Once a job has been automated, the hard work is done... capturing the knowledge required to do the job into a machine represent-able form. The robot is just an output device at that point, outputting work units.

I would be very surprised (and interested) to see any evidence of jobs that were successfully automated being de-automated at some later time.
posted by MikeWarot at 11:10 PM on November 14, 2013 [4 favorites]


Except in Ireland.

And in nicaragua.
posted by empath at 11:12 PM on November 14, 2013


I would be very surprised (and interested) to see any evidence of jobs that were successfully automated being de-automated at some later time.

Many of the jobs off-shored to third-world countries replaced automated work in the US with cheaper human labor.
posted by empath at 11:13 PM on November 14, 2013 [13 favorites]


I think a good solution to automation is public ownership of corporations. Lay off workers, and give them grants of dividend-paying stocks. All the excess profits from automation go back to the people.

Or just to make it simpler, just guaranteed income for everyone.
posted by empath at 11:16 PM on November 14, 2013 [24 favorites]


oh, empath.
posted by Auden at 11:27 PM on November 14, 2013 [12 favorites]


With all that money GMU is getting, I am finding it more ridiculous that students have to pay +$300 a semester for a parking permit to park there. Employees of GMU have to pay as well.

/rant
posted by littlesq at 11:30 PM on November 14, 2013 [1 favorite]


As Marx pointed out, each capitalist firm is compelled to economize its use of labor in the quest for profits. What is good for the individual capitalist, in the long run, undermines capitalism through ever-building tensions ("contradiction") in society, as classes polarize and automation makes material abundance possible. Capitalism's once progressive social relations - private property, wage labor - increasingly become "fetters on production".

"At a certain stage of development, the material productive forces of society come into conflict with the existing relations of production or – this merely expresses the same thing in legal terms – with the property relations within the framework of which they have operated hitherto. From forms of development of the productive forces these relations turn into their fetters. Then begins an era of social revolution. The changes in the economic foundation lead sooner or later to the transformation of the whole immense superstructure." - Marx

I work in the software industry, and here the contradiction could not be more plain. My employer, like most, relies heavily on open-source tools, and will allow workers to contribute to them while on the clock to a point; they do this not out of goodwill or anti-capitalist spirit, but because it simply works better.

Everyone would be better off if all code were open-sourced, but the social relations of capitalism stand in the way: capitalists usually need to withhold "intellectual property" in order to turn a profit, and the vast majority workers who write the code have to sell their capacities to bosses in order to live (bourgeois fantasies of the HackerNews set aside).

As productivity throughout the economy continues it's steady climb, these social relations start to look like fetters throughout the economy. Consider how much labor was required to feed the U.S. in 1850 versus today. Some of the work moved to other parts of the world, or other sectors, but the vast majority has just been automated out of existence. Who reaped the benefits?
posted by wobdev at 11:52 PM on November 14, 2013 [26 favorites]


Who reaped the benefits?

The aristocracy, as it has always done and will ever do.
posted by flabdablet at 12:49 AM on November 15, 2013 [3 favorites]


As I've said relatedly, automation is an absolute good because it frees humans' time for fun and progress : family, learning, art, activism, inventing, etc.

At the moment, we're forestalling universal underemployment by creating bullshit work in management, administration, law enforcement, finance, etc. We attract people to these pointless jobs, and ensure their loyalty, by paying them more than almost any productive worker, except extremely skilled labor like doctors, developers, etc.

I now believe the most efficient way to bring about positive social progress is to undermine the upper middle class's busywork jobs in management, administration, finance, etc. Just shrink the fattened social class so the system collapses.

Finance is likely going to automate itself out of existence, simply because they hire intelligent educated people.

Administration and management is less obviously being automated because managers are not usually any cleverer than their underlings. Yet, administration and management has roughly three components : Bureaucratic interactions are rote activities much better done by machine. Optimization decisions are way beyond most managers education because they involve the same mathematics as finance, like stochastic calculous. And finally human factors are much better handled through collaborative technologies like wikis, stackexchange, etc. I'd therefore argue that our current managers aren't even remotely competent at their jobs, so the automation revolution might take their jobs rather quickly.

We need some sort of developer, academic discipline, or just a philosophy that seeks to build open source tools for administration and management so that start ups never take on much management. Over time, these tools should filter out to large companies and government, creating unemployment amongst the upper middle class, ideally undermining our fucked up social order.

There is plenty of money to be made in replacing management jobs too. We need developers who adopt an open source morality though, knowing they'll earn a good living form consulting, rather than capitalist who seek to maximize returns by charging per seat. Anyone want to help me write this manifesto?
posted by jeffburdges at 1:26 AM on November 15, 2013 [4 favorites]


I haven't read the book, but I've been following this on Cowen's blog posts on "Marginal Revolution".

The stuff about how we're seeing a massive rise in inequality: very convincing. The stuff about how we may be seeing a permanent and unprecedented fall in the number of available jobs: fairly convincing, deeply worrying. The stuff about how all this is because of technological changes which mean that an elite of super-brainy, super-diligent workers can do the same work as many dumb peons: not convincing at all.

For one thing: where is the surge in productivity? I would expect output per worker should be soaring if this was true. To a degree, Cowen can finesse this by saying "well, the low productivity of the dumb peons is exactly or over-compensating for the rising productivity of the cognitive elite, so it's invisible", but I don't find that explanation very convincing. A productivity change this big ought to be highly visible somewhere.

Matthew Yglesias did an interesting review of "Average is Over":
...scratch the surface and you'll see that Cowen actually thinks there's a policy reform agenda that, if implemented in a timely fashion, would channel much more of the gains of future automation and computerization to the median household...

...The reason Cowen focuses more on the somewhat dour outlook for the median household than on these possible solutions to the problem is that he doubts the political system will deliver any of these solutions. He notes that none of them are particularly on the partisan agenda of either political party, that the nature of the U.S. political system makes large changes generally unlikely, that an aging population is less likely to embrace radical changes, and that elites have a lot of ways of reenforcing their control over the political process.

That seems like a reasonable forecast to me. But it's a very different forecast from the forecast that automation and the rise of the machines means that "average is over." The actual forecast is that the political system will be under the control of a relatively narrow elite who will stomp on the interests of the median household.
posted by TheophileEscargot at 1:32 AM on November 15, 2013 [20 favorites]


Our C-suite overlords replace people with machines not because the machines are more productive per dollar invested; they're not. However, machines don't complain about their working conditions (other than by breaking down if not properly maintained) and above all they don't organize.

Replacing people with machines is done for the very simple reason that doing so undermines the power of labour unions, which in turn allows the corporations to get away with buying far cheaper human labour in Special Economic Zones offshore.
posted by flabdablet at 1:42 AM on November 15, 2013 [5 favorites]


Today robots are very expensive and pretty dumb. In the future, robots will be cheap and still pretty dumb. But there are lots of jobs, even service jobs, that really don't require human level intelligence-- for example, most of the assembly line work at a typical Chipotle could be automated today, but you'd need two or three robots costing probably $200k+ each. When those robots get down to say $20k, then the prospect of a mostly automated restaurant becomes much more attractive.

While today robots are too expensive to replace fast food workers, they aren't too expensive to replace truckers. As the prices of robots drop, it will become economical to do this for more and more jobs.
posted by Pyry at 1:53 AM on November 15, 2013


Who cares, flabdablet? All that matters is that machines are more productive per human hour spent, making the world a better place in the long term.

Yes, all the underemployed must eventually take back control, whether by vote or by protest or by guillotine or by boycott or by assassinations using little poison coated $2 Toy's R Us micro-quadrotors.

It's the satisfied upper middle class that really holds back the "revolution", but their jobs are actually easier to automate than say truck driving or speaking on a phone.
posted by jeffburdges at 2:00 AM on November 15, 2013 [2 favorites]


As I've said relatedly, automation is an absolute good because it frees humans' time for fun and progress : family, learning, art, activism, inventing, etc.

Well, except, it also frees them from a paycheck, which facilitates that fun and progress. Unless the profits from automation are given back to the displaced worker rather than to the capitalists, automation leads not to leisure but desperation.
posted by GenjiandProust at 2:03 AM on November 15, 2013 [19 favorites]


Another odd idea of Cowen's is that there are or were a large number of "Zero Marginal Product" workers in the US economy. These are workers who are on the payroll, or were pre-recession, but achieve literally nothing for their employers, workers who are a complete waste of their salary. The existence of these zero-productivity workers is part of his explanation for why the overall productivity figures don't back up his narrative, that a cognitive elite are causing inequality by being so darn super-productive with their computers.

That's a very odd belief for a right-wing "freshwater" economist like him. Markets and businesses and capitalism are supposed to be efficient. Yet he apparently believes that for the decade and a half of "great moderation" before the recession, across all sectors of the economy, there was a massive market failure where businesses just refused to lay off completely useless workers, even though it would have caused their profits and shareholder value to skyrocket.

Personally, I don't believe there were that many "ZMP" workers, and I don't believe that soaring inequality is mostly due to technology. I think wages for most have been depressed by falling unionization, and increased competition from offshore workers. I think soaring inequality is because a managerial elite have used the principal-agent problem to divert the some of money saved by switching to non-union and offshore labour into their own pockets, rather than turning it all over to shareholders.

Weirdly, that makes me more of a market fundamentalist than he is. He believes in a massive scale, long-lasting market failure. I only believe in a partial and long-studied market failure, the principal-agent problem (in this case, shareholders not being able to fully control their managers).
posted by TheophileEscargot at 2:16 AM on November 15, 2013 [14 favorites]


Having looked at Cowen's book, I would say this article really does summarize all of its major points. I found it depressingly convincing. I will be following this thread closely.
posted by wittgenstein at 2:16 AM on November 15, 2013


i don't think he's right about the politics of the situation - he's outlining a period of radical change in american work and i don't believe conservatives appreciate radical changes - also, if one takes a look at who has been protesting and being disruptive, it's clearly the extreme right wing

he also forgets that desperate people do desperate things and there will be an increase in desperate people in his scenario

i expect more social turmoil, not less
posted by pyramid termite at 2:38 AM on November 15, 2013 [2 favorites]


We could trivially address unemployment by shortening the work week, GenjiandProust. All past work week reductions benefited the economy, workers, etc. All that's lacking is political will. We should create that political will by targeting the upper middle class overseer jobs and overpaid busywork jobs for automation.

There is considerable evidence for near-zero marginal product workers in agriculture before the industrial revolution, TheophileEscargot, namely the fact that agriculture lost so many workers during the industrial revolution without causing starvation. Institutions rarely recognize these minimally productive workers until forced to.

We're currently saturated with near-zero marginal product workers across our entire economy. Just consider what most jobs actually entail. There is no new industrial revolution to lure them into productive work however this time. We should instead shorten the work week to spread the useful work around and reduce inequality. I've no approach to doing this except to force the issue by using automation to undermining our ability to invent busywork.

Also, long-lasting market failures are ubiquitous throughout our economy, well GM, Microsoft, etc. all still exist. All this talk about market forces basically asserts that "Biological evolution occasionally optimizes something, never mind that whole sexual selection for colorful tail feathers waste", except obviously biological evolution is not dominated by large individuals, while markets are dominated by large corporations. How could anyone not expect rampant market failures?
posted by jeffburdges at 2:39 AM on November 15, 2013 [1 favorite]


The "marginal" is important in ZMP: it theorizes that the workers provide no net benefit, so that employing them is value neutral rather than a literal waste of money. It's not a market failure.
posted by topynate at 2:51 AM on November 15, 2013 [7 favorites]


jeffburdges: 'Management' is a very broad church, which functions do you propose could be automated?
posted by freya_lamb at 3:01 AM on November 15, 2013 [1 favorite]


No robot would do my job. The office politics would kill them.
posted by Mario Speedwagon at 4:07 AM on November 15, 2013 [6 favorites]


you'll know the robots are here when frivolous robotics startups are valued at a billion dollars (instead of big marketing data) and everyone is going back and learning C instead of Ruby or whatever.

all of this is to just distract you a moment longer so they can steal more of what productivity increases there have been.
posted by ennui.bz at 4:26 AM on November 15, 2013 [1 favorite]


human factors are much better handled through collaborative technologies like wikis, stackexchange, etc.

With respect jeffburges, I think this comment says more about your own narrow experience than anything even close to universal.
The general public is pretty skeptical of wikipedia, knows nothing of the bizzarro tech pissing contest that is stackexchange and generally doesn't have the skills or inclination to engage with either.
Human to human interactions are hard to automate if you care about the outcomes. Of course, countless call centre IVR systems demonstrate they can be trivially automated if you don't.
However, I'm cynical enough to believe your other comments on management.
posted by bystander at 4:30 AM on November 15, 2013 [5 favorites]


We’re on the verge of having computer systems that understand the entirety of human “natural language"

Is there any evidence for that, or is he just extrapolating from Google Translate and over-excited university press releases? We've been told we were on that verge for sixty-odd years.
posted by Segundus at 4:35 AM on November 15, 2013 [11 favorites]


And I don't consider the Asimov story either overlooked or dystopian.

So there.
posted by Segundus at 4:37 AM on November 15, 2013 [1 favorite]


All that matters is that machines are more productive per human hour spent, making the world a better place in the long term.

There's an implicit assumption there that life is necessarily better if there's less paid work in it, which I don't think is necessarily true. If it were true, the C-suite would already be empty; any member of the 1% could easily afford to quit tomorrow and spend the rest of their lives sipping piña coladas by the lagoon. And job satisfaction is by no means an upper-management-only phenomenon.

Machines doing shitty dangerous repetitive work with a high risk of injury that a person would only take on from stark necessity: sure, as long as we also work out a sustainable way to rid the world of stark necessity. But machines displacing enjoyable, satisfying work? Thanks but no thanks.
posted by flabdablet at 4:45 AM on November 15, 2013 [8 favorites]


We've been told we were on that verge for sixty-odd years.

I can personally attest to having heard predictions that strong AI is ten years away for the last forty.
posted by flabdablet at 4:49 AM on November 15, 2013 [3 favorites]


or how about this: why is the biggest "tech" company (and repository of hundreds of billions of dollars of otherwise productive capital) of the last 15 years an advertising and marketing firm that dabbles in robot cars?

Cowen wants you to debate a point re: robots which is almost entirely counterfactual.
posted by ennui.bz at 5:02 AM on November 15, 2013 [1 favorite]


We could trivially address unemployment by shortening the work week, GenjiandProust. All past work week reductions benefited the economy, workers, etc. All that's lacking is political will. We should create that political will by targeting the upper middle class overseer jobs and overpaid busywork jobs for automation.

Only if wages rise. Suppose I am satisfied with my standard of living at $40K/year, working a theoretical 40hrs/week. Your scenario seems to assume that I will continue to be paid $40K/year for, say 30hrs/week. What seems more likely is that the capitalists would want to pay me $30K/year for those 30hrs/week and pocket the rest (minus the cost of automation). Which means a rather steep reduction in my standard of living (including probably all funding for leisure activities) of finding a second job which will probably eat up the leisure time created. So, again, unless the benefits of automation are permanently marked for the worker, workers will suffer.
posted by GenjiandProust at 5:09 AM on November 15, 2013 [5 favorites]


or this: "globalization" started in the 80's and 90's with business executives deciding, instead of investing in modern manufacturing machinery i.e. robots, they would lay off US workers and pay to have old machines shipped to China and Mexico where they could earn a profit from cheaper labor.

there is an alternate reality where industrial corporations in the US and Japan spent the end of the 20th century competing over manufacturing automation and energy efficiency/energy alternatives (see oil embargo). what would the productivity increases have looked like then?
posted by ennui.bz at 5:20 AM on November 15, 2013 [1 favorite]


Aaargh. On my phone and on my way to work, so can't link, but in re "jobs to china not robots" McKinsey did a study that shows global manufacturing employment in decline with even Asia experiencing loss due to automation, also two MIT econ professors have done a ton of work on productivity decoupling/the productivity gap -- in other words, we are now seeing GDP growth without employment growth, also Izabella Kaminska at the financial times has like a 25 part series on this on FT's Alphabille blog (requires log in). I linked to a bunch of her source material in a post I made a couple months ago. Also Paul Krugman has been giving some serious thought to this issue as well, just had an op we on it this weekend.
posted by Diablevert at 5:44 AM on November 15, 2013 [6 favorites]


but in re "jobs to china not robots" McKinsey did a study that shows global manufacturing employment in decline with even Asia experiencing loss due to automation

my point was that business was disinclined to invest in machines 30 years ago. sure there have been productivity improvements and as wages in China have increase, manufacturers have been investing in improved productivity but the trend is 30 years late in many ways. Cowen wants you to think, but we have PCs and the internet -----> productivity. but a world where the robots are here is one flooded with cheap FPGAs not cellphone SOCs.
posted by ennui.bz at 5:55 AM on November 15, 2013


I have to say, even as someone who wishes that conservatism and low-tax fervor was on the rise, Cowen is staggeringly wrong at his prediction of a new conservative America that is "already significant and observable in today's America." This makes me doubt that he's looking at data without bias.
posted by corb at 5:59 AM on November 15, 2013 [1 favorite]


(robots)...can be stored in a box, and re-purposed, without having to spend $100,000 to get a new 4 year degree.
This is nonsense. Any industrial robot is astronomically more specialised that a low wage human. I can produce slightly below average spot welds after a single hours instruction, can saw a plank in half after 30 seconds instruction, can reproduce with an embarrassingly short copulation time and have the ability to keep running on a wide range of cheap available inputs.
And can write a reasonable mefi comment straight after.
I'll buy that robots can produce better consistent quality and possibly work out cheaper per action over a long career, but the idea they offer flexibility in any meaningful way is cray-cray, as the kids say.
posted by bystander at 6:08 AM on November 15, 2013 [12 favorites]


...unless that was a dig at the cost of education, which I will agree has reached absurd levels for simple careers.
posted by bystander at 6:13 AM on November 15, 2013 [1 favorite]


1690 headline: economists predict coming factory apocalypse.

Changing technology invalidates old economic models. Capitalism/communism had a good run, but they will die off as surely as feudalism and mercantilism did, and for the same reason. They were a product of the technology of their eras.

It'll be ugly for a while, transitions from oneeconomy to another usually are (see workhouses, 16 hour factory jobs, etc from our last transition). But the uglyness wilo pass and we will wind up better off. What we need to be doing is not bewailing the coming death of capitalism/communism, but rather trying to figure out how to minimize the fallout from the transition.
posted by sotonohito at 7:07 AM on November 15, 2013 [1 favorite]


The jobs in China thing was more directed at empath than you, ennui. But personally I don't see how what business decided was cheaper to do 30 years ago means anything about what's possible and logical today.

I'll buy that robots can produce better consistent quality and possibly work out cheaper per action over a long career, but the idea they offer flexibility in any meaningful way is cray-cray, as the kids say.

Meet Baxter. Is he limited, sure. But he's also only Version 2.0. This is closer than we think.

In re: "we've been hearing all if this for 60 years, we're not there yet" Kevin Drum had a good point on this about the nature of exponential growth in his article on this --- if you start pouring a shot glass of water into a giant hole, and then a year later you pour two shot glasses, then a year later a quarter cup ..... well for 50 years you get nothing but a very slightly damp patch of dirt. And in 60 years you get Lake Michigan.

I don't think we're so close to cracking natural language that IBM's Watson is going to replace Kevin Kline in the role of Polonius. But to do stuff like, say, taking a list of symptoms and searching medical encyclopaedias and studies to come up with a diagnosis and potentially a treatment plan? Yeah, I think we're awful close to that. And in the broad sense, diagnosing disease and coming up with a treatment plan is a doctor's job. It's not all of what they do, but it's a lot, and it is the component of what they do that requires intellectual skill and high levels of training. I don't think you can automate the entirety of the medical profession. But there's a post on the front page of mefi right now that points to how robots could replace a whole bunch of lab techs. This doesn't need to be a one to one thing, where human doctor gets replaced by robot doctor. Instead what we'll get is "tasks that take up 25% of human's day now done by robot". Either way it works out to fewer doctors.

Also, here's an article from the MIT tech review that discusses some of the productivity decoupling work. And here's the McKinsey report --- key graphic is the very first one, showing the number of people employed in manufacturing worldwide dropped 24 percent between 1996 and 2006.

And here's krugman's op-ed. I'd urge anyone seriously interested in this to go through the rigamarole of getting an FT login and reading Kaminiska's Beyond Scarcity series. (The very first entry's a bit impenetrable, but she links to tons of interesting work that's being done on this.)

It's funny, I've been interested in this for a while, and lately I feel like it's been breaking out all over -- the post on the revival of Marxism from yesterday is essentially a post on this, because the key part of the automation is that all the rewards of growth flow to capital and not labor. I am not sure that Marxism will provide and adequate guidebook for solutions, however, when a breaking a strike doesn't mean bringing in scabs and the Pinkertons but just shipping in a few Baxters. Ladies and gentlemen, weaponize your sabots, here we go again...
posted by Diablevert at 7:09 AM on November 15, 2013 [10 favorites]


Betting on supply chain issues to save human jobs is a bad bet, in my opinion. I think that supply chain disruptions would be an inconvenience, not something which can put humans back to work replacing them.

The argument that automation might somehow magically create just as many new jobs ancillary to producing robots is profoundly illogical.

If it worked out that way, it would turn out there had been no benefit to automating labor in the first place. The incentive to automate is decreased cost in human capital and the potential to achieve the management sciences dream of reducing all working people to interchangeable units like light-bulbs (so managers can be freer to hire and fire to meet short-term business needs). If the only way to accomplish that were to support an expensive new market that gave just as many people jobs with comparable wages to those being eliminated, there'd be no net gain in efficiency and there'd be no point in automating.

It's a lie/self-deception to think that we can automate jobs away but make up the losses with gains in jobs related to automating jobs away. If it worked that way, there'd be no economic benefit to automating.
posted by saulgoodman at 7:11 AM on November 15, 2013 [6 favorites]


I can't help looking at this with an ecological designer's eye. Yes, all these forces are in motion, but has anyone stopped to ask the simplest, most important question? What are we trying to do here?

Looking at outcomes, it appears the objective is to maximize inequality, eliminate financial security (and even viability) for the bulk of the population, while destroying the biological systems that make civilization possible. Without a clear goal, we are pursuing the default goal with all it's casualties.

I think it's worthwhile to ask what the economy is there to do, serve those ruthless and unscrupulous enough to grab as much as they can get away with, or provide a secure, respectable life for all without trashing the nest?
posted by Dodecadermaldenticles at 7:20 AM on November 15, 2013 [3 favorites]


I work in the software industry, and here the contradiction could not be more plain. My employer, like most, relies heavily on open-source tools, and will allow workers to contribute to them while on the clock to a point; they do this not out of goodwill or anti-capitalist spirit, but because it simply works better.

I work in the same industry, but I've always regarded software development and the open source movement as rather compatible with Marxism. The means of production (code) is collectively owned by the workers and the boss class has very little ability to tie workers to their production task. It isn't like a factory where the worker has little means to go build their own factory or fail competing piecemeal against the assembly line.

On a more topical note, one of the most fanciful ideas out there is automation of software development. I love tools that make custom logic more available to the masses but the notion of automating programming is about the most Utopian thing I can't think of, which is to say the least likely to be done in my lifetime.
posted by dgran at 7:28 AM on November 15, 2013


serve those ruthless and unscrupulous enough to grab as much as they can get away with

Pretty much.

Bread and circuses have always been relatively low-cost.
posted by flabdablet at 7:35 AM on November 15, 2013 [1 favorite]


I've always regarded software development and the open source movement as rather compatible with Marxism

...especially as contrasted with top-down hierarchical corporate control, which is rather compatible with Stalinism.
posted by flabdablet at 7:36 AM on November 15, 2013 [1 favorite]


one of the most fanciful ideas out there is automation of software development

First saw that one thirty years ago.
posted by flabdablet at 7:40 AM on November 15, 2013


Seems to me that in a world with a population at 7 billion and heading towards 10 billion, we're going to get to a point where disposable slaves are cheaper than robots.
posted by vibrotronica at 7:42 AM on November 15, 2013


saulgoodman: I've encountered your argument before about not being able to put everyone to work making robots. Neither side is "illogical", they just make different assumptions. You are assuming 1. we have limited demand for the product and 2. the robots are no more efficient than workers.

E.g.: If 1 robot does the work of 10 people, clearly there is no benefit if it takes 10 people to make the robot. But if it takes 5 people to make the robot, then why would 10 people not just build 2 robots and everyone gets twice as much stuff?

I'm not saying everyone is going to be employed making robots in the future (although you could argue this is what the whole IT industry is...). But even if we were, the right question seems to me whether it's more productive and we can consume everything we produce, not whether it makes sense economically.
posted by ropeladder at 7:48 AM on November 15, 2013


Vibrotronica: population growth is shrinking fast around the world and much of Asia and Europe are headed for big demographic busts.

Anyway disposable foreign labor has been much cheaper than robots for years. This is one of the major points of David Graeber's article, Of Flying Cars and the Declining Rate of Profit.
posted by ropeladder at 8:00 AM on November 15, 2013


E.g.: If 1 robot does the work of 10 people, clearly there is no benefit if it takes 10 people to make the robot. But if it takes 5 people to make the robot, then why would 10 people not just build 2 robots and everyone gets twice as much stuff?

We don't make robots so people can get more stuff. We make them because there are people who think they can fire some percentage of their workforce if they have a robot to do the job instead. In order for the math to work out such that this arrangement is job neutral, the people hiring the robot makers would ultimately need to be spending just as much on getting robots made as they had been spending on paying employees, or else there's no reason for anyone to use robots. If automation can't deliver fewer people working or lower wages to workers, there's no economic case to business for using them. People don't do business so we can have cool stuff or enjoy life. They do business to fill up their bank accounts with money. Who's going to buy that second robot? Where does the demand for the second robot come from when you're first robot just put half the workforce in one company out of a job?
posted by saulgoodman at 8:02 AM on November 15, 2013


Anyway disposable foreign labor has been much cheaper than robots for years.

That was once true but this advantage has been eroding. As living standards improve in the third world, labor costs there are rising. And as computing power increases, robots can be trained to do more complex jobs and safely work closely with human, so the cost of automation is dropping. The widely reported surge of growth in the US manufacturing sector is largely being driven by this shift.
posted by Diablevert at 8:20 AM on November 15, 2013 [1 favorite]


People don't do business so we can have cool stuff or enjoy life.

Well, no, but that's why we consume. And so we buy from businesses that provide us cool stuff and help us enjoy life.

Automation can deliver fewer people working and lower wages to workers, and that when it does so the question is whether those productivity effects are met with static demand (and the whole industry is reduced by the amount of productivity gained) or increased demand.

I'm not saying that technological disruption and low wage work are not or cannot be problems. I'm just arguing that there is nothing inherently illogical about saying we can all just make robots. It's just the same thing as arguing we can all just make self-driving tractors instead plowing the fields by hand. Obviously that's not practical but it's practicality has to do with field space and demand for food, not marginal costs to employers or other internal business logic.

Diablevert: You're right. I was just responding to Vibrotronica who was arguing that labor was getting cheaper as population increased.
posted by ropeladder at 8:33 AM on November 15, 2013 [1 favorite]


In skilled commercial applications, robots are not for missions in which they're cheaper than people, they're for missions in which they are better than people and may (or may not) be cheaper than the people they replaced. We don't have autopilots or high-frequency trading computers because human pilots and equity traders have high salaries.
posted by MattD at 8:33 AM on November 15, 2013 [1 favorite]


Well, no, but that's why we consume. And so we buy from businesses that provide us cool stuff and help us enjoy life.

The act of automating jobs away in the short term decreases economic demand by reducing overall employment. Less money in consumption, less demand for innovation, less economic development. Absent some miraculous economic stimulus falling from heaven, the very act of automating makes it less likely there will be demand for those new industries and services that are supposed to make up the gap.

People who lose their jobs due to automation aren't then magically freed-up to start innovating and creating new opportunities: they're unemployed. In a job market that increasingly disadvantages unemployment, offers access to fewer safety net support options, and is constantly demanding more highly-skilled yet more adaptable and flexible workers. Sorry, but high unemployment is not an economic stimulus. If forcing people into poverty stimulated innovation, most poor nations would quickly be back among the most innovative and economically prosperous in the world. But that's not how it works in the real world. In the real world, poverty brings stagnation.
posted by saulgoodman at 9:18 AM on November 15, 2013 [5 favorites]


High frequency trading is just a scam.. let's leave it out of the discussion, please.

We're going to eventually face a time when the government stops lying, and we face the fact of 50% employment, and worry how to keep it that high, while it asymptotically approaches zero.

The industrial revolution gave us the ability to all have tons of stuff we don't need, and then send it to the landfill. That's what ate up the productivity increases. This time around there aren't enough inputs available to do it again.
posted by MikeWarot at 9:21 AM on November 15, 2013 [2 favorites]


E.g.: If 1 robot does the work of 10 people, clearly there is no benefit if it takes 10 people to make the robot. But if it takes 5 people to make the robot, then why would 10 people not just build 2 robots and everyone gets twice as much stuff?

Well, and also, even if it takes 20 people to make a robot that replaces 10 people, that's 10 people's yearly salaries for every year of the life of the robot, and the people making robots are certainly making more than one robot. Even if they only made 10 robots in a year, that's 100 people's yearly salaries for every year the robot's operational. And the number of people required to make the robots will decrease as the production of the robots gets streamlined. And the years of service of the robots will increase as they get better.

I think the thing that's holding robots back is likely less about production costs and more about the fact that we're still just solving a lot of basic problems with them, they're just not all there yet. But each time one of those problems is solved, it's cheaper and easier for the next generation robot, and a lot of the hard problems are being solved with public money so the burden isn't entirely on private firms and the whims of investors, so that dam's gonna break sooner or later. And when it does...

In skilled commercial applications, robots are not for missions in which they're cheaper than people, they're for missions in which they are better than people and may (or may not) be cheaper than the people they replaced. We don't have autopilots or high-frequency trading computers because human pilots and equity traders have high salaries.

That's definitely the trend in skilled applications, but expansion into unskilled applications has always been on the roadmap, and we're getting closer every day. Things like the Kiva warehouse bots, just as one example. Some day sooner rather than later those or something like them will not only slash the workforce of big Amazon warehouses, but also knock out the back room staffs of big box stores across the country. Combine that with manufacturing and other manual labor robots, call center automation, etc... there are so many smaller cities where the floor will just fall out when those particular low-pay, low-skill jobs dry up. I live in an area like that now - big box stores, manual labor and call centers are pretty much it for the low-end jobs, and the only thing that's kept those jobs around is that they have to be local and can't be outsourced, or in the case of call centers they've just set up shop in areas in the US with low wages and cost of living instead of outsourcing (basically just to keep American accents on their end of the phone) - but once automation hits a sweet spot and moves in, that all goes away. And, to the last, they're all owned by huge companies who have the buying power to get that automation out to all their locations. That local hardware store won't be able to automate its backroom, but that local hardware store probably doesn't exist anymore anyways. Home Depot, though?

I mean, obviously, there has always been doom just around the corner from automation for more than a century now and it hasn't always panned out that way, but now we're in a particularly fragile position to handle it right as the technologies really pick up steam.
posted by jason_steakums at 9:21 AM on November 15, 2013 [4 favorites]


For me, they picked the wrong classic era SF story. A better choice would have been Frederik Pohl's The Midas Plague (short version - robots pour out a literal cornucopia of stuff, and the problem has stopped being supply and started being demand. What happens to all those goods? The answer they've settled on is that the poor have to consume them in an exhausting orgy of changing clothes and replacing kitchen appliances and so on. The rich can afford to live simply.)

That doesn't seem so far - to me at least - from a guaranteed minimum income. The reason you'll ultimately get a guaranteed minimum income, no matter how obscene it may seem to the elite right now, is because otherwise the system eventually falls apart. While corporations/the rich might have the ability to produce goods very cheaply indeed, there'll be nobody waiting to buy them. In other words, as we run out of natural demand (not the desire for goods, but the ability to acquire them) then demand itself will in some weird meta-sense become in short supply and high demand. Thus we will create more of it.

The way to do that is to give money to the poor and let them use it to buy things from the corporations' robot factories.

This won't work if Toyota gives me money and I use that money to buy a Toyota because there's nothing coming in from outside. However, it will work just fine if the government gives me money to buy a car and I buy a Toyota. What you get then is a situation where the supposedly meritocratic elite will then create another meritocracy in their upper class world. They will compete to most successfully meet the desires - and thus mop up the guaranteed income - of the commoners in the lower class world. Will I buy a Toyota with my government check or will I buy a Volvo?

I would argue that something like this is already going in in pockets of the economy. Witness the for-profit education industry which has done a lovely job of siphoning money out of the federal government through guaranteed student loans to the poor and undereducated.

Witness also the way food companies recognize that food stamp users have become their own distinct and surprisingly important market segment. Energy drink makers, for example, recognize that food stamp users are a huge part of their market and lobby hard against restrictions that would prevent food stamps from being used to purchase energy drinks and other "bad" foods.

As more people are moved out of the productive economy and kept on government support so they won't starve or start setting fires, it's just a matter of time before this principle becomes more widespread.
posted by Naberius at 9:27 AM on November 15, 2013 [5 favorites]


Naberius, Kaminska and Krugman are already on your same page.
posted by Diablevert at 9:36 AM on November 15, 2013


freya_lamb: "'Management' is a very broad church, which functions do you propose could be automated?"

I know a couple of managers that could easily be replaced by inanimate blocks of wood with faces painted on them. Mostly nobody would know the difference except for a few areas where the number of meetings would drop and productivity would go up.
posted by Hairy Lobster at 11:34 AM on November 15, 2013


This is something I think about now and then, and whether this is actually going to happen or not.

Thing is - when you look at the current situation, especially in light of the things that Prof. Richard D. Wolff speaks about here, you see that this sort of thing is really the historical trend (shit - it's a basic marxist economic analysis when the idea of extracting surplus value (profit) out of things).

You have women entering the workforce, increased immigration and offshoring which leads to wages that are slowly being depressed. Look at what one of the main arguments against liberating slaves was, to the poor white workers - essentially it's the same thing - slaves will enter the paid market, and reduce your pay rate by increasing competition.

It's how you get all these Tea Party types against women in the work force and immigration and "America First" (I suppose you could question which comes first, sexism, jingoism and nationalism or the other way around - I think Marx would argue it's the economic conditions which breed these other factors).

Regardless, what I'm getting at here: this is one more thing with the trend of capitalism to increase profit via decreasing labor costs.

So - if we take this assumption at face value, that increased automation of production (and I am assuming, without RTFA-ing, that Cowen is also discussing the trend in Japan towards robotics that do more than production, but are now being designed to work towards nursing/care roles, due to the demographic shifts occurring), that as we increase the role of robotics and automation, we find that humans have less and less to do, and this assumption, I think, goes fundamentally at odds with the moral code we have sold ourselves over the past ~400 years ("The Puritan Work Ethic"), and we have these retrograde rejects pushing an agenda that stigmatizes those who are punished via this systematic change cycle (which is ironic, since many of them have been punished by it as well), and so we have these morality lectures on people on welfare and medical care. We have this Ayn Rand belief in Self-Authorship and Bootstrapping combined with this perverse historical reading of one specific Holy Book (The Bible - not The Fountainhead or Atlas Shrugged) to make a set of long lasting cultural enforcements (which at least, in theory, had a tint of compassion rooted in it), degrade into taking the worst of those enforcements and ripping out the compassionate heart and replacing it with this ego-based "Autonomy" (funny - same root as "automation") and "self-sufficiency".

So what happens? What happens when more and more people compete for fewer and fewer jobs - you either support the view that all people have an inalienable right to life liberty and the pursuit of happiness, in which case, the inability of these individuals to find work means that they must seek help elsewhere if they are to provision food, shelter and all basic animal necessities that guarantees the bare minimum of "life" (that's not taking into account any creature "comforts" (which they also love to blame the poor for needing, of course -- that is "pursuit of happiness")) (and it's all because that damned monkey in the middle "Liberty" playing games, using its rhetoric to deny others the fundamental right to life and existence).

So we have a culture that blames and threatens those who are suffering under the thumbs of increased automations, refusing to guarantee them basic rights, meanwhile, as we're pitted against one another, those who reap the profits of the increase in "worker productivity" (that is to say "the Capitalists") get off scot-free and blameless (not merely blameless, but the exact inverse: Praiseworthy) for the mythographies of the day proclaim them the supreme bootstrappers, the most honorable and righteous above all else.

The rest of us?

We can just fuck off and die.
posted by symbioid at 12:34 PM on November 15, 2013 [3 favorites]


Really the robots we have are just better factory machines. They're just doing more of what happened in the 19th century and taking it up another few notches. There may be a paradigm change, but I'm not seeing it yet.
posted by Segundus at 1:06 PM on November 15, 2013


TheophileEscargot: "The actual forecast is that the political system will be under the control of a relatively narrow elite who will stomp on the interests of the median household."

That's... not a forecast. It's a statement of fact, put into the future tense.

EDIT: OK, I suppose it still is, in the sense that "The sun will rise tomorrow morning" and "Water will be wet in the year 2042" are forecasts.
posted by IAmBroom at 1:17 PM on November 15, 2013 [1 favorite]


Really the robots we have are just better factory machines. They're just doing more of what happened in the 19th century and taking it up another few notches. There may be a paradigm change, but I'm not seeing it yet.

Brain v. Brawn. The industrial revolution replaced human and animal muscle power with machine power: tractor instead of horse, spinning Jenny instead of grandma's foot at the spinning wheel. This caused a massive shift into, well, paperwork. Administrative, managerial, accounting, designing, brokering stuff. Work that requires brains, to analyse and categorise and define and calculate. Not necessarily genius level stuff --- but stuff that requires the computing power of the human mind to make decisions. That's what's under threat here. That means that the general prescription for what happens when your job gets wiped out doesn't apply --- acquiring more education, more skill, will not necessarily help you.
posted by Diablevert at 1:28 PM on November 15, 2013


Robots don't create the economic theory under which they are employed. I refer to the robots that build our autos for one pattern of obsolescence. They won't create an army of unemployed workers anymore than the tractor created a nation of unemployed farmers.

It might work out that our future will be a bit less fucked up than what happened when industrialization sent people from rural areas to work in factories. Might be that Road Warrior is right around the corner. Or, maybe the new world order will be like the world of Star Trek.

In either case, I wouldn't blame the robot anymore than I would blame the cell phone. Or indoor plumbing.
posted by mule98J at 2:23 PM on November 15, 2013 [1 favorite]


Tons of robots making tons of stuff that all those out of work people won't be able to buy. The future is an odd one.
posted by Foam Pants at 5:08 PM on November 15, 2013


America: the best run company in the world.
posted by Vibrissae at 9:59 PM on November 15, 2013 [1 favorite]


The reason you'll ultimately get a guaranteed minimum income, no matter how obscene it may seem to the elite right now, is because otherwise the system eventually falls apart.

The thing is, though, that the very idea of an as-of-right guaranteed minimum income does seem obscene to the aristocracy, almost all of whom are afflicted with the delusion that their disproportionate buying power is in some way deserved. And since the aristocracy is in charge of public policy, anything that functions as a guaranteed minimum income will come with some form of arbitrary requirement attached to maintain the fiction that it's actually some kind of payment for services rendered.

This might be an obligation to pretend to engage in a meaningless hunt for nonexistent employment and submit a report on that every week, or some kind of "work for the dole" scheme involving busywork indistinguishable from that used for the "community service" criminal sanction, or simply a requirement to wear a prominently visible electronic tracking device; the detail doesn't matter very much. The point would be to make sure anybody accepting their guaranteed-available minimum income remains as low in the social pecking order as it is possible to be.

The fact that a more equitable distribution of wealth would supercharge the economy as no other measure possibly could (possibly even to the extent of keeping capitalism viable indefinitely) is irrelevant to those in charge, whose main concern is remaining more powerful than everybody else. The point is to be at the top of whatever heap exists, and if you make the whole heap flatter, that gets harder.
posted by flabdablet at 10:17 PM on November 15, 2013 [3 favorites]


America: the best run company in the world

Cite, please
posted by flabdablet at 5:49 AM on November 16, 2013


Flabdablet - Yeah, I completely agree that they'll find some kind of sumptuary law approach that makes you do something humiliating and draining to get your spending power. Right now my guess would be that it would involve big data and marketing (since the whole point of this exercise for the elite is to generate sales). Maybe your "job" is watching TV commercials for eight hours a day while lasers scan your eyeballs to track where you look. Or your car radio can't be turned off and only plays ads keyed to whatever businesses you're driving by right now.
posted by Naberius at 6:38 AM on November 16, 2013


Maybe your "job" is watching TV commercials for eight hours a day while lasers scan your eyeballs to track where you look

Google Glass without an adblocker?
posted by flabdablet at 9:39 PM on November 16, 2013


There's no reason to assume that the system requires guaranteed income. Given the use of drones and 'non-lethal' crowd control systems and ubiquitous automated surveillance, and so on, it would be just as easy to intimidate everyone into compliance and imprison and/or execute dissidents. A fully automated economy has no need for a large population to keep the elites well fed and pampered. It works just as well with a small population as a large one.
posted by empath at 2:14 AM on November 17, 2013 [1 favorite]


I agree, pretty much, with empath. Our political solutions nowadays generally come in various flavors of capitalism. Other options are available.

The pampered class can be pampered by robots, using only a smallish work force: for greasing gears and filling hoppers, maybe preparing meals, or any other tasks done better by the human hand. They need only to keep their security technicians in arms and ammo, as a way of holding the unwashed, and (strictly speaking) unnecessary, others at bay.

I might even suppose that the pampered class could amuse themselves by piloting armed UAV's, looking for targets to strafe. This would not only provide a way for them to idle away their leisure hours, but serve as an object lesson to their indentured labor force, to illustrate on which side of the fence lies the greener grass.
posted by mule98J at 11:09 AM on November 18, 2013


Jacob Knitig: "a too-overlooked 1955 short story by Isaac Asimov."

I don't know. Franchise has been pretty frequently republished.
posted by Chrysostom at 11:20 AM on November 18, 2013


There are advantages conferred upon nations whose populations participate more voluntarily, empath and mule98J, like more technological, cultural, etc. advancement. And external pressures act as a sanity check on government policy.

Assuming no world government, we'd hopefully acquiesce to such violent robotic suppression so differently in different nations that external pressures erode the power of the more oppressive elites, as we're currently seeing in the U.S. and U.K.

Now a single world government is the scariest imaginable political system precisely because it's eliminates that sanity check. World government is worse than religious dictatorships, fascists, etc. It's so scary that almost any upbeat statements about human progress require this tacit "no world government" assumption I just employed.

Automated surveillance is an incredibly dangerous development, but the danger comes partially from how technology gets a "world wide rollout" before anyone understands the consequences. There is no reasonable analog of the "no world government" assumption for information technology, as few understand it, but conversely people do understand violence even when committed by robots.
posted by jeffburdges at 3:25 PM on November 18, 2013


Asimov on world government.
posted by Chrysostom at 6:15 PM on November 18, 2013 [1 favorite]


and (strictly speaking) unnecessary, others

No, that doesn't really work. If you simply get rid of the teeming millions you've done much the same thing as if you'd implemented universal income: you've flattened the pyramid. There is no point at all in having a huge fuck-off yacht if most everybody else has one too - the plebs have an absolutely vital social role, from the aristocracy's point of view, which is to be impressed.

external pressures act as a sanity check on government policy

Might want to re-think that one from the Pakistani point of view with respect to American government policy, and run a thought experiment about likely US drone policy if Pakistan were in fact one of the United States.

a single world government is the scariest imaginable political system

The arguments you raise against a single world government are actually arguments against a single world administration, which is not necessarily the same thing. The folks who currently run the huge multinationals are a de facto world administration that remains largely unaccountable to the public.

I frequently despair, when discussing this and related matters with people rather to my right, at their inability to let go of the idea that it's the label "government" attached to a huge and dangerous bureaucracy that makes it scary and wrong. To my way of thinking, an elected government is the least scary form of huge and dangerous bureaucracy, precisely because it does at least have to pay lip service to public accountability and can in theory be forced to respond to public pressure. Corporations, as they presently operate, really don't - except insofar as is necessary to keep the propaganda view of any retail arm positive.
posted by flabdablet at 6:48 PM on November 18, 2013 [1 favorite]


I agree completely that "the huge multinationals are a de facto world administration that remains largely unaccountable to the public", but..

We're rebelling reasonably effectively against the international treaty system through which the multinationals impose legislation upon nations, because (a) national leaders are often ultimately accountable only to their own populations and (b) treaties largely depend upon consensus before they carry any weight. Any world government would simply impost ACTA, TPP, etc. by finding enough bought, dumb, confused, etc. representatives.

Law enforcement is carried out infinitely far sanely throughout much of the rest of the world than in the U.S., especially in continental Europe. Any world government would quickly lose that via the war on drugs, war on child porn, etc.

In fact, we've infinitely more chance of regulating the multinationals if we dissolved existing world government-like institutions, like the WTO and WIPO. Imagine what the pollution situation in China might look like if Europe imposed extra import duties on products whose production involved excessive pollution?
posted by jeffburdges at 1:24 AM on November 19, 2013


You'll get no argument from me on the scariness of WTO and WIPO, not because they're world-government-like but because they are, much like the huge corporations for whose benefit they exist, both fundamentally undemocratic and far too narrow in scope. If considerations other than trade factored into the decisions made by those bodies, I think they'd make sounder decisions.

Any world government would simply impost ACTA, TPP, etc. by finding enough bought, dumb, confused, etc. representatives

I was going to pass snarky remarks about the US form of government, but then I remembered who's running my place at present. So you have a point, but I still can't see how a bought, dumb, confused etc. representative is actually worse than a non-representative.

Law enforcement is carried out infinitely far sanely throughout much of the rest of the world than in the U.S., especially in continental Europe. Any world government would quickly lose that via the war on drugs, war on child porn, etc.

There's an implicit assumption there that the US would necessarily be more dominant in any world federation than it already is, which I don't see as a given.
posted by flabdablet at 4:04 AM on November 19, 2013


At present, the world is run by consensus amongst nations mixed with military and economic force imposed by larger nations. Ain't surprising an unelected consensus based system makes better decisions than an elected majoritarian system.

Also, large nations are slowly losing power to the consensus elements. American is in dramatic decline, while China and the E.U. will likely never grow as powerful as America once was.

We must make the wold decision making bodies even more consensus based by dismantling treaty organizations like the WTO and WIPO, while making individual nations more democratic through electoral reform.

Almost any topic you might name, like climate change, workers rights, etc. actually benefits more from decentralized power because even a few "activist nations" banning offending products places serious pressure on the multinationals. Any world elected body would inherently standardize on infinitely worse rules than what the most progressive individual nations choose.
posted by jeffburdges at 4:52 AM on November 19, 2013


Ain't about America's influence on a world government I'm worried about either. All electoral systems are eventually subject to exploitation, decline into totalitarianism, etc., at least via Aarow's theorem, but usually by much worse measures.

Any world government would necessarily wind up worse off than America now, simply because abuse is intrinsic to power. And power is inherently too absolute because consent is too easy to manufacture. All elites require that external pressure before they'll behave, support progress, etc.

America is now slowly decriminalizing cannabis. Why? Ain't internal electoral pressures, voters thought that way for ages. It's the cold hard external reality that everyone else found that works better. America acquiesced to civil rights only because segregation gave the Soviet Union an edge. etc. etc. etc.

I'd imagine even the ultra-progressive places like Scandinavia became even more progressive because they pride themselves on setting an example and recognize that their culture allows for experimentation.
posted by jeffburdges at 5:12 AM on November 19, 2013


« Older Love's Secret Ascension   |   Outside of all corporate taxation zones Newer »


This thread has been archived and is closed to new comments