Now, everyone is weighing in on it. Two economists - Erik Brynjolfsson and Andrew McAfee, wrote a book about it that's being serialized in the Atlantic. Paul Krugman has once again raised the issue several times in columns and on his blog. Noah Smith also wrote a piece in the Atlantic about automation. Wired magazine put the issue front and center, with a techno-optimist piece by arch techno-optimist Kevin Kelly. BoingBoing summarized a lot of these and weighed in here. Tyler Cowen and Dean Baker weighed in. Several columnists have written about it. 60 Minutes did a piece recently on it. Matt Yglesias considers technological unemployment to be a myth (more on that later).
I guess I can claim some sort of foresight here. And another of my common themes - that the truly transformative technologies have already been invented, that innovation will bring increasingly diminishing returns because we've picked the low-hanging fruit first, and that we're inventing things mainly to keep our economy humming rather than to fulfill basic human needs - has been taken up by The Economist Magazine, no less. See:
Can We Ever Do Better Than The Toilet? (Washington Post)
I suppose I should imagine that my other themes might start to go mainstream - that we'll incereasingly wonder why the economy isn't getting "back to normal", why "growth" isn't returning, why more growth delivers less and less net gain; and that our net energy supplies are decreasing, that climate change threatens agricultural output and the existence of civilization, that democracies are actually police states controlled by an economic elite - might start to get even more attention. Maybe people will even realize that more innovation will not solve problems without implementing social reform.
As they say on the message boards: FIRST!
When I wrote WAPGF, my aim was actually quite simple. For years I had tried to avoid automated checkout lines, seeing them as just a sideshow or an inconvenience, but eventually I was unable to do so. Living in the dessicated rust belt, I know that when manufacturing left, our overall living standards declined and never returned to their former peak, except for a lucky few. And I knew that for years we were told by economists (the same ones who said we would all benefit from "free trade") that the "service economy" was the future of jobs, and that workers would just transition into stocking shelves and walking dogs, and the economy would go on as before. Yet clearly, these automated checkout lanes belied that idea.
At the same time, it became nearly impossible for me to buy a movie, new book, or CD anywhere but online, and stores were closing all over the place Why? Online shopping, of course. You don't need a weatherman to see which way the wind blows. And I knew that behind that artificially manipulated low unemployment rates was a tide of low-wage McJobs, accompanied by soaring tuition costs since people will pay anything to access the few remaining jobs that pay decent wages.
When you see the crumbling ruins of your formerly great city around you every day, and when everyone is perpetually broke, you know that the happy talk from politicians and economists is just blowing smoke up your ass. Still, it seems most people believe it, or are just numbed by fast food, American Idol and spectator sports. But if you look closer, you can see that we are being lied to.
Most of the arguments boil down to this:
"Yes, some existing jobs will disappear, but many new ones will emerge." or as Kevin Kelley puts it, "Each successful bit of automation generates new occupations—occupations we would not have fantasized about without the prompting of the automation...Robots create jobs that we did not even know we wanted done."
So first, as for Kevin Kelly's essay, I suppose I should congratulate him for doing the most thought work of piecing together a robot economy that works for all. He's clearly done the most thought on this past the glib pronouncements from economists about lump of labor fallacies and the like. There are two problems, however. One is so fundamental that it's impossible to miss. As this article by Fabius Maximus describes, he presumes that jobs created will balance with jobs destroyed. Highly unlikely.
Kevin Kelly, Wired, 24 December 2012 — "Imagine that 7 out of 10 working Americans got fired tomorrow. What would they all do?" Kelly provides this brilliant graphic:
And the second is that his argument rests on creating in his mind highly speculative, made-up jobs.
Kelly assumes the future must be like the past, and makes no effort to size the effect of the various factors. This leads him to this Dr. Pangloss-like forecast, assuming that new jobs will appear to replace the old. The above four boxes probably will not be of equal size.
This postindustrial economy will keep expanding, even though most of the work is done by bots, because part of your task tomorrow will be to find, make, and complete new things to do, new things that will later become repetitive jobs for the robots. In the coming years robot-driven cars and trucks will become ubiquitous; this automation will spawn the new human occupation of trip optimizer, a person who tweaks the traffic system for optimal energy and time usage. Routine robo-surgery will necessitate the new skills of keeping machines sterile. When automatic self-tracking of all your activities becomes the normal thing to do, a new breed of professional analysts will arise to help you make sense of the data. And of course we will need a whole army of robot nannies, dedicated to keeping your personal bots up and running. Each of these new vocations will in turn be taken over by robots later.
The real revolution erupts when everyone has personal workbots, the descendants of Baxter, at their beck and call. Imagine you run a small organic farm. Your fleet of worker bots do all the weeding, pest control, and harvesting of produce, as directed by an overseer bot, embodied by a mesh of probes in the soil. One day your task might be to research which variety of heirloom tomato to plant; the next day it might be to update your custom labels. The bots perform everything else that can be measured.Um, does anyone really believe this claptrap? I would love for Kelly to walk through some of the abandoned suburban real estate schemes in southern California, and maybe volunteer at a nearby homeless shelter or soup kitchen. Maybe that will disabuse him of seeing "trip optimizers" and "professional analysts" everywhere. Are we all going to own a robot-controlled farm? (what will all the Mexicans do?). We can't even afford non-robot controlled farms right now. Please, Kevin, spend some more time outside of Silicon Valley!
Oh, and incidentally, the autoclaves that keep instruments sterile are already being automated. So is drug dispensing - ever hear of a Pyxis machine? I have.
Some of Kelly's assertions are patently false:
That may be true of making stuff, but a lot of jobs left in the world for humans are service jobs. I ask Brooks to walk with me through a local McDonald’s and point out the jobs that his kind of robots can replace. He demurs and suggests it might be 30 years before robots will cook for us.Here's the reality TODAY, courtesy of Marginal Revolution:
The Robot Restaurant opened in Harbin in June and has taken the F&B industry in China further into the mechanized world. Robot Restaurant staffs a total of 20 robots as waiters, cooks and busboys. Turns out Noodle Bot might need to expand its repertoire if it hopes to compete with Robot Restaurant’s 18 different kinds of service robots.Is Kelly ignorant or deliberately obfuscating? And in Japan:
Upon arrival, Usher Robot welcomes customers to the restaurant and directs them to the seating area. Patrons can then place their order, which is relayed by humans to one of the four the robot chefs who are able to cook various styles of dumplings and noodles. The robot chefs even determine the temperature and ingredients for each dish and usually take about 3 minutes to prepare the average order. These robot chefs are no slouches either. The kitchen staff is able to prepare a menu of over 30 dishes–perfect for a family dinner.
Waitress robots carry the food to customers by following a track that uses sensors placed under the floor for spatial awareness. Additionally, each robot comes equipped with its own sensors, helping it to avoid obstructions such as a kid that’s in its way.
The robots were designed and built by a local firm, the Harbin Haohai Robot Company. Each robot costs between 200,000 to 300,000 Chinese yuan (US$31,500 – US$47,000) with an additional 5 million yuan (US$790,000) invested into the restaurant itself. With the average Robot Restaurant meal costing less than 62 yuan (US$10), the restaurant is not meant to earn Harbin Haohai money. Instead, it turns out the restaurant is just a brilliant piece of marketing.
Efficiency is paramount at Kura: absent are the traditional sushi chefs and their painstaking attention to detail. In their place are sushi-making robots and an emphasis on efficiency.http://www.nytimes.com/2010/12/31/business/global/31sushi.html?src=me&ref=business&_r=0
Absent, too, are flocks of waiters. They have been largely replaced by conveyors belts that carry sushi to diners and remote managers who monitor Kura’s 262 restaurants from three control centers across Japan. ("We see gaps of over a meter between your sushi plates — please fix," a manager said recently by telephone to a Kura restaurant 10 miles away.)
Absent, too, are the exorbitant prices of conventional sushi restaurants. At a Kura, a sushi plate goes for 100 yen, or about $1.22.
Such measures are helping Kura stay afloat even though the country’s once-profligate diners have tightened their belts in response to two decades of little economic growth and stagnant wages.
Kelley argues we'll work with robots, not compete with them:
And once we can cowork with robots right next to us, it’s inevitable that our tasks will bleed together, and soon our old work will become theirs—and our new work will become something we can hardly imagine.Here's the problem with futurists: they don't understand economics. That's why they thought we'd all be living in space colonies even though it makes absolutely no economic sense to do so (who's going to pay the the transport costs? What will the jobs be? How much do the fuel and materials for these stations cost?, is there a cheaper alternative?, etc.). Capitalism is all about creating the most output for the least amount of workers. In effect, it is a system designed to destroy jobs, yet at the same time it abhors redistribution, refuses to restrict hours, and insists that everyone sell their labor power. Automation replaces people, or allows you to do more stuff with less people. If you need to hire people to work with the robots, you defeat the entire economic purpose of automation! You replace people, get it? And for those people to be absorbed into another sector of the economy, another sector must be growing. Right now, no sector is growing except for health care, and it will not absorb all the displaced workers, especially since it is on a fanatical drive for "lean" process improvements (something I know about first-hand). It's the argument, as Martin Ford intelligently put it, that when tractors replace farmhands, everyone will get jobs driving the tractors. I'm sure the sushi restaurant described above does not employ more workers than it would otherwise.
It is a safe bet that the highest-earning professions in the year 2050 will depend on automations and machines that have not been invented yet. That is, we can’t see these jobs from here, because we can’t yet see the machines and technologies that will make them possible.A safe bet, eh? I'll take that bet. Say, weren't people like Kelly saying we'd all have jobs like space explorer , nuclear engineer, moon colonist and rocket scientist back in 1955? I thought so. As I've said before, if the entirety of your argument is an appeal to something "we can hardly imagine," your argument isn't very good.
You’ll be paid in the future based on how well you work with robots. Ninety percent of your coworkers will be unseen machines.He's half right. But if ninety percent of my coworkers are machines, and unemployment has not gone up by 90 percent, then that means the economy is almost ten times larger! Does anyone expect that kind of growth by 2050? Won't oil be running scarce in 2030?
If I can publish and buy books online, that eliminates a lot of middlemen. Sure it's easier to publish books, but where will the wages come to buy them? I think people like Kelly assume we'll just buy each other's books. As Martin Ford points out:
Every product and service produced by the economy ultimately gets purchased (consumed) by someone. In economic terms, "demand" means a desire or need for something – backed by the ability and willingness to pay for it. There are only two entities that create final demand for products and services: individual people and governments. (And we know that government can’t be the demand solution in the long run). Individual consumer spending is typically around 70% of GDP in the United States.So who has the better command of economics, Kelly or Ford? Kelly or actual economists Brynjolfsson and McAfee? Who is basing their arguments in the reality we live in right now, rather than vague appeals to imagination and possibility? Who is using facts? Make up your own mind.
Of course, businesses also purchase things, but that is NOT final demand. Businesses buy inputs that are used to produce something else. If there is no demand for what the business is producing, it will shut down and stop buying inputs. A business may sell to another business, but somewhere down the line, that chain has to end at a person (or a government) buying something just because they want it or need it.
The point here is that a worker is also a consumer (and may support other consumers). These people drive final demand. When a worker is replaced by a machine, that machine does not go out and consume. The machine may use energy, resources and spare parts, but again, those are business inputs—not final demand. If there is no one to buy what the machine is producing, it will get shut down. Think of on industrial robot being used by an auto manufacturer. The robot will not continue running if no one is buying cars.*
So if we automate all the jobs, or most of the jobs, or if we drive wages so low that very few people have any discretionary income, then it is difficult to see how a modern mass-market economy can continue to thrive.
Incidentally, Kelly also believes that the single biggest problem facing humanity is that there are not enough people. Seriously. He's also associated with the "bright green" environmentalists, and here he is on the neoliberal Chicago-school economics podcast Econtalk expounding his panglossian view of the future. This just cements my perception that these guys are just moles in the environmental movement for the high-tech neoliberal investor crowd (who always want more "innovation" to keep profits high and more people to keep wages low).
We can talk possibilities all day long, but we need to consider facts. I present just two: since the end of 2008, 28,000 new net jobs were created. That's barely one-quarter of what we need in a single month to keep up with population growth produced in FOUR YEARS.
And according to the UN, the number of jobless people around the world rose by 4 million in 2012 to 197 million and is expected to grow further:
In a report, the International Labour Organization (ILO) said the worst affected were youth: nearly 13% of the under 24s were unemployed. It said global unemployment was projected to rise 5.1 million this year and by a further 3 million in 2014.The trend reflected a downturn in economic growth, the document said.This was particularly the case in developed countries. The report - Global Employment Trends 2013 - said that 6% of the world's workforce were without a job in 2012.Now these facts not in dispute. But we're told that's it's just a cyclical downturn, like the seasons; that times are tough and jobs are scarce in recession, but that things will get back to normal once the economy picks up again. We're also told we must wait, wait, wait, and we will emerge from this downturn stronger than ever, just like all the other times, or the magically promised robot-enabled jobs we can't imagine will appear. Hope springs eternal.
But are we waiting for Godot?
There is a debate over whether automation has exacerbated inequality or whether it is primarily due to policy changes. See this:
'Skill-Biased Technological Change and Rising Wage Inequality' (Economist's View)
The policy change crowd points out - correctly - that incomes have stagnated for thirty years, longer than the new automation technologies have been around. But of course such debates are asinine - economists are doing what they always do - split into warring camps and insist that only one of them has THE cause. Of course, both play a role as "force multipliers." It's like trying to explain the single thing that caused Rome's fall - Christianity, or civil wars, or barbarian invasions, or malaria, or drought, or lead poisoning, when of course all of these factors played a role and reinforced one another.
The either/or debate is just a way of not taking the issue seriously. And yes, the answers are policy answers, but they go much further than economists are willing to consider. As Cory Doctorow points out:
But here's the thing that neither of these articles -- or even Bruce's acid observations -- touches on: once technology creates abundance, what possibilities exist for distributing the fruits of that abundance such that the benefits are more evenly felt? We've been talking about an increase in productivity producing an increase in leisure for a long time, but instead, the "winner take all" world of Brynjolfsson and McAfee often seems to produce a "winner" class that works itself into an early grave by running 100-hour work weeks at astounding payscales, and a much larger "loser" class that works itself into an early grave by working 100-hour weeks in shitty, marginal, grey-economy jobs, trying to stitch together something like an income.Not only are we not making structural reforms, but we're actually going backward. We would need an expansion of non-salary-income, yet the safety net and things like unemployment insurance and social security are under attack and in danger. We need a broad, equitable distribution of income, instead we have the most skewed income since the Great Depression, and America has the most inequality in the developed world. We need a reduction in hours, and yet professionals work practically a 16 hour day, with service job hours reduced only to avoid paying benefits. We need to redistribute income downward, and we're redistributing it upward to the already rich. We need to broaden stock ownership, instead it gets more concentrated. We need an expansion of government policy to help this along, instead we get debt crises, layoffs and austerity. And the people who lose their jobs are scapegoated and demonized by pundits and politicians on the payroll of the capital ownership class as lazy and unproductive "water drinkers" or some such nonsense.
In America, anyone who proposes an increase in overall quality of life through public schools, health programs, libraries, or even Internet access, is immediately branded a socialist and dismissed out of hand.
On the other hand, the Internet-age's sweetest dividend is the creative possibilities: the chance to sit in your little grass shack or organic farm or urban crackerbox and use the tubes to carry on debate; to contribute to software and Wikipedia; to crowdsource capital for your creativity; to find makers who have solved 90% of the problem that's nagging you and who will help you solve the remaining ten percent; to access a library of human creativity and knowledge without parallel; to have your art and creativity accessible to all, and to find the mutants who're wired the same as you and jam with them.
As for the Yglesias article, even his chart does not explain what he means to, as several commenters pointed out:
Hatch WhitfieldAnd the comments turned into a huge debate, mainly between Jed Rothwell on the side arguing that automation is a problem, and Jasper In Boston on the cornucopian side. Rothwell has a significant insight that I think bears mentioning. I have embedded and edited some the relevant portions of the discussion below, with some items underlined.
It seems to me that this graph actually presents evidence in favor of something like technological displacement.
Look at the red line, which represents aggregate worker hours: in 2002, it was at about 97, while in 2012 it was at about 100, for an increase of 3 (as far as I can tell, these are percentage points). Now look at the blue line, which is real GDP: over the same time period, that went from roughly 103 to about 121, an increase of approximately 18. So there was actually a much larger increase in GDP than in worker hours, which means that something else must be driving productivity other than labor. Alternatively, we could say that we achieve the same level of output in 2012 with less labor input compared to 2002.
Matt's right that there is a lot of covariance between the two curves, but that's hardly surprising, and even people who believe that technology is displacing workers would agree that large-scale macroeconomic trends will strongly affect labor demand -- it would be difficult to believe data that didn't show a strong relation between the two. But visual inspection also shows that the overall linear trend for GDP (i.e. the regression line) has a clearly larger slope than the trend for aggregate hours.
Of course none of those "proves" that technology is displacing workers, as the data don't address that issue. But it certainly doesn't indicate that there is no technology effect either, as Matt seems to think. And it definitely suggests that something has been going on over the last 10 years (at least) that reduces the demand for labor.
Actually, the red and blue lines are growing increasingly further apart. It would be sad if they were not, because that would mean that productivity is stagnant.
The problem is not that the red line and blue line are growing apart, the problem is that, while the red line has more than recovered since the recession, the blue line is still below recession levels.
The question is not whether technology displaces workers, that is obvious. Look at farming. The question is whether there will continue to be new fields to employ displaced workers.
Stagnant wages in America indicate that instead of new high value fields employing displaced workers, displaced workers are instead accepting lower wages. Displaced workers are not being employed by newly invented jobs, but by jobs that existed in the past, but that people were not willing to do in the past because of better options.
Automation is great, because it means better products and services, and increased leisure - the question is how to prevent the increased leisure from existing as mass unemployment along with a small number of high hour, high skill jobs.
Scary chart showing that hiring is disconnecting from output.
Notice that production began to recover steeply before hours hit bottom and has continued to rise even while hours stabilized. All that money invested in productivity assistance, and all that work piled on existing staff is paying off for employers.
So Mr. Yglesias's chart shows the opposite of what he announced it shows.
Jed RothwellRead the entire thing here: The Myth of Technological Unemployment (Slate). A good debate - make up your own mind.
Here is a list of all the jobs in the U.S.:
Look at that list, and ask yourself how many of those jobs could not be done by the kinds of robots that are capable of driving automobiles (Google), or beating the world champions at Jeopardy (IBM). Those robots cost a fortune now but in a generation they will cost a few thousand dollars each.
Self driving cars and trucks alone would eliminate millions of jobs in category 53. Look at the robots now working in warehouses. They could easily eliminate grocery store shelf stockers and clerks, 1.8 million (43-5081) How many sales clerks do you need buying things online? Sayonara most of 13 million sales related jobs. Robots could do most fast-food jobs already, and they will soon be capable of most cooking jobs. 2.8 million jobs gone.
Waiters and Waitresses at restaurants: 2.3 million. A generation from now when robots can do this, people will think it is creepy to ask a person to do it. It would be like asking a person to wash dishes by hand when you have a dishwasher right there. It would be demeaning.
Yes, some existing jobs will disappear, but many new ones will emerge.
I'm not worried about the future.
Where will they emerge? Every "Major Occupational Group" on that list has existed for the last 150 years. There are no new Groups. People have been shifting from one group to another, but there are no new groups. If nearly everyone in every group is replaced with a robot, what other occupations will appear?
I am not worried either, but I can see that work is coming to an end. Human labor will be worthless. That's a good thing.
It would only be a good thing if we were moving towards a post-scarcity society. I don't see that happening anytime soon.
What I do see is a large, idle underclass coming into existence.
Every "Major Occupational Group" on that list has existed for the last 150 years.
This may be the most meaningless thing you've written on this entire thread. Are you seriously claiming with a straight face that the economy hasn't produced any novel forms of employment in the last 1.5 centuries? We had software engineers working in the 1930s, did we? Also, technology doesn't just create entirely new forms of employment -- technological churn has historically expanded some types of employment (even as it decimates others). Many more people work in healthcare than ever before, for instance, despite huge improvements in technology.
The economy has not produced any new "Major Occupational Group" in 150 years because every kind of labor we want done for us is already done. As I said, people have moved from one group to another, as the amount of labor ebbs and flows in different sectors. But there are no new groups, and robots will move into all groups simultaneously.
We did not have many software engineers in the 1930s, although there were some at IBM developing wonderfully complex punch card accounting systems. We had accountants doing manual book-keeping. They were later replaced by a much smaller number of programmers writing accounting software. But here's the key thing:
The core job -- book-keeping -- remained the same. We do not have any new core jobs. We do not do any fundamentally new tasks that our ancestors did not already do in 1880. There are only new ways to fulfill the same needs people fulfilled back then. People worked in management, entertainment, insurance, law enforcement, healthcare and other categories back then, and they work in the same categories today. Look at the list. You will see that every job category in the universe of employment today was represented back then. The numbers of people in each category was different, and the tools of the trade were different. The point is, computers and robots will soon be able to do just about EVERY SINGLE JOB listed in that table. Except education and decision making management. That will leave everyone else unemployed.
Regarding all those people working in healthcare: I have been in the hospital. Most of them could be replaced by robots right now and the care would be a heck of a lot better than it is. I have never seen such inefficient methods! When a nurse needed an aspirin, she would fill in pages of paperwork and then walk down a long hall, get the aspirin, give it me, and fill in more paperwork. She would take 5 minutes to do something a robot could do better in a few seconds.
In a sane society it will mean less people work and the people that do work, work less hours.
Yes! However, after a while it means that most people do no work at all, because there is no point to working 5 hours a week. You cannot keep your edge doing a professional job for so few hours.
In a sane society, we will give everyone enough money to live comfortably and enjoy life whether they work or not. We can't afford to do that now. In the future, we will not be able to afford NOT to do that. Everyone's happiness and economic security will depend upon it.
But in the society that actually exists, it means that 10% of the population will work twice as hard and make 100 times as much money and 90% of the population will be unemployed.
I fear you may be right. That is why we need to start adjusting the economy and our attitudes toward work now.
Self driving cars and trucks alone would eliminate millions of jobs in category 53
In exactly the same manner that cars themselves eliminated millions of jobs formerly held by blacksmiths and carriage drivers and horse dealers. You still haven't provided a scintilla of evidence as to why the creation of new fields won't open up as yet undreamed-of jobs, or indeed create growth in existing fields (as, for instance, the advent of automobiles increased the number of people working in oil production, or in tourism).
You have not shown why robots and computers will not be capable of doing any new job that we dream up in the future. Watson can diagnose disease better than any human doctor. Our robots will soon be capable of working faster and cheaper than any human factory worker. So what new "fields" will we discover that we can do but they cannot? There are none.
Once again unemployment is not the core problem of our economy. The core problem is WAGES. In the past those massive productivity gains made their way down to the workforce in the form of rising wages. Those wages purchased more goods and services which drove demand and led to increased hiring. Not a perfect system but one that at least worked for a good chunk of the population.
But over the past 12+ years wages have been flat to declining for the vast majority of the workforce. The big productivity gains (and the profits that came with them) did not filter down to the workforce. They went instead to the financial sector and shaky mortgage backed securities. For most of the zero years the average person compensated for declining wages with debt. Credit cards, second mortgages etc. That well ran dry in 2008 and still wages remain flat. Companies are sitting on big profits and instead of using those profits to give their existing workforce more money (and thus more buying power) they going right back into the financial markets.
The teen years are going be even worse than the zero years.
The main reason why an increase in production must be balanced by an increase in wages is to maintain stability in the system.
When wealth is concentrated, as it often will, it causes instabilities in the market as production capacity cannot be adequately distributed in any semblance of fairness or efficiency.
The instability spills over into politics, social unrest, and culminates in a crisis of some sort (war, revolution, transfer of power, hyperinflation). In this crisis a lot of wealth is usually wasted or destroyed, but when the dust settles the society is usually more equal, if not necessarily better off.
In the last 30 years the crisis has been artificially delayed, but the underlying problem has not been addressed. Unless it is there will eventually be a crisis
Something else to consider, as a benefit of automation. As automation reduces the amount of labour needed for a certain product, it should likewise reduce the cost of that thing. In a perfectly competitive market this reduction in cost should have a corresponding reduction in price, which is great for pretty much everyone, except for the inconvenience of the displaced workers who have to find something else to do.
This would be fine if markets were free and efficient, but they are not. For example, the compensation paid to CEOs would suggest that there is an extreme shortage of qualified CEOs. This may even be the case, but the extreme compensation should have taken care of that shortage a long time ago.
If people are unemployed there can be only two reasons: Either those people have no economic value, or there is a shortage of capital to employ their economic value. If a country believes that its people have no economic value then it has lost faith in itself. If there is a shortage of capital, then technology is the solution, not the problem.
The Real Inspector Hound
i meet a lot of CEOs.
most of 'em could be replaced by a toggle switch.
Or a Magic 8 Ball.
As for productivity improvements, productivity has been pushed so high that it has created surplus productivity. That's what persistent unemployment is -- human productivity that cannot be put to use at any price.
It turns out that demand is indeed finite. And by now, all the goods and services that the market can clear are being produced with only about 80% of the global workforce. This is not a US issue but a global one.
Consider, how is it that China can supply itself with the goods its needs AND supply us with the goods we need? There are enough jobs for the US or for China, but not enough jobs for BOTH the US and China.
Something to consider, as a futurist discussion, is what the world will look like when human labour is (at least mostly) obsolete. This is sort of an end-point of capitalism, because at this point the profits of production that once was split between the labourer and the capitalist now go completely to the capitalist. But at the same time, there are no profits either, because unless someone is willing and able to buy surplus production, that surplus is worthless.
This puts the capitalist in a very dangerous position, because he is now surrounded by people that he doesn't need, but that need what he has. His options are to give it away, flee, or start killing. If dystopian novels are any indication, then the killing seems most likely - heaven forbid. Fleeing seems to be the preferred option of Objectivists.
But I have some hope for humanity that a person who finds himself with wealth that is no worth to him but great worth to his neighbors will not withold it out of spite. A man can dream.
The broader problem, (and this is the most politically incorrect thing you can say, especially around educators), is that there really are dumb people. Tens of millions of middling to dull 75-95 IQ 'high school was a challenge' folks. Who aren't going to become information architects or outsourcing project managers no matter how many hours they spend in classrooms.
Until recently, technology advances made machines stronger, faster, and more reliable than average Joes. But, even at the slow end, he was much better at mopping a floor, understanding speech, packing a box, or driving a lorry than even the best supercomputer. So, he had some major competitive advantages for just being human.
Now, (with some redesign of the workplace around 'smart' machines) we're getting to the point that a machine can do *everything* that a mediocre worker can do, and do it, net, cheaper than even paying him a mere subsistence wage.
You wrote: "But, even at the slow end, he was much better at mopping a floor, understanding speech, packing a box, or driving a lorry than even the best supercomputer. So, he had some major competitive advantages for just being human."
Exactly right. That is what makes the latest advances qualitatively different from all previous technology in history. That is why the old rules about expanding occupation no longer apply. The last set of competitive advantages that people had over machines is rapidly dwindling away. And not just for one occupation, but for all of them simultaneously
We've already replaced a third of human labor with automation many times over.
This same argument occurred with the mechanical loom, the cotton gin, factories, cars, electricity, computers, Internet, etc., etc., etc. Every time we've created a new technology that makes an old job obsolete the result is more and more jobs.
It is nothing new, but that does not mean it is insignificant or unimportant. Farmers are now less than 1% of the population. The need for lawyers will soon be reduced. Taxi drivers and truck drivers will be eliminated once antonymous cars are perfected. Pharmacists are already being eliminated. Sooner or later you run out of jobs. People do not need an infinite amount of food, or transport, or pharmaceutics. There is a limit to our demand, and once machines can fulfill all demand, we will have no need for human workers.
Just because something has been happening for 200 years that does not mean it will never end, or reach a crisis. We have been burning fossil fuel for 200 years but it only now causing significant global warming. U.S. agriculture has been losing topsoil for 150 years, but it is only now causing problems in many places.
Demand is not infinitely large. We cannot expand production of food, automobiles, lawyers, doctors or druggists by a factor of 10. Nobody wants that much food, or needs 10 cars. I sure don't want any more healthcare or lawyers than I need. Material demand is limited, and once machines can meet it there will be no work for human workers.
For 30-odd years of the continuing Industrial Revolution, we accommodated increased productivity by lowering working hours from 70+ hours a week to 40, improving working conditions, and increasing pay. That lasted until roughly the Depression.
The post-WWII answer? Market, market, market. Consume, consume, consume. Increasing net 'stuff' consumption by the early 1990s to a full order of magnitude more than the 1930s. And when 75% of the country couldn't afford more and more consumption, we put it on plastic.
Now, good old Jean-Baptiste Say says that we can just keep on doing this. Add on another order-of-magnitude more consumption over the next 30 years this time (as tech improvements just keep coming faster).
Besides the Malthusian arguments against it, there's a real question of satiation. There's only so many hours in a day. Only so many shoes one can wear in a given day. And, we live in a nation wth more plastic pumpkins shoved in more ministorage complexes than any other people in history. And it's so dehumanizingly ugly that a quarter of the population has to pop antidepressants to put up with it.
If we can't and won't consume more, more, more, and can't figure out the economics of dropping well below the 40-hour-workweek/40-year-career (an ideal more discussed in the 1930s-50s than today sadly), then, yes, this technological unemployment problem is new and is real.
And as a final note, Jed Rothwell's point above that we have added no new fundamental job categories is key to the innovation debate as well. The Internet revolution is transforming how we do business, true, but that does not have the effect on people's day-to-day lives as the period from 1850-1950. We still do the same fundamental tasks, just with computers - keep track of inventory, customer service, communication, logistics, manufacturing, accounting, construction, transportation, education, publishing, etc. We're doing the same things, just differently. That's very different from the majority of people never leaving their home village, or not having electricity or antibiotics, or not being able to communicate other than by letter, or being able to move around faster than a horse could ride. The two are just qualitatively different. The debate is not that there is no innovation going on, it's that they are less transformative and less world-changing to people's lives. Today as I write this, there is a ten below wind chill outside, and I can tell you that being able to access water, including hot water, cook, and go the the bathroom inside is worth more to me than this laptop I'm typing on. That was Robert Gordon's point. Most of the new "innovations" are merely toys. Much of what the innovation optimists tout - video games, apps, Twitter, Facebook, YouTube, etc. is merely entertainment.
No entrepreneur creates jobs. They fulfill needs. And if our needs are adequately fulfilled with only thirty percent of the workforce working, what then of the other seventy percent?