In one of the most thought-provoking economics books of our times, A Farewell to Alms
“there was a type of employee at the beginning of the Industrial Revolution whose job and livelihood largely vanished in the early twentieth century. This was the horse. The population of working horses actually peaked in England long after the Industrial Revolution, in 1901, when 3.25 million were at work. Though they had been replaced by rail for long-distance haulage and by steam engines for driving machinery, they still plowed fields, hauled wagons and carriages short distances, pulled boats on the canals, toiled in the pits, and carried armies into battle. But the arrival of the internal combustion engine in the late nineteenth century rapidly displaced these workers, so that by 1924 there were fewer than two million. There was always a wage at which all these horses could have remained employed. But that wage was so low that it did not pay for their feed.” (page 286)
The U.S. has 15 million officially unemployed workers and additional tens of millions who aren’t working and aren’t looking for a job. Could these folks be the draft horses of the 21st century?
The cost of a low-skill worker has increased tremendously in the U.S. Let’s look at four kinds of costs:
- direct payments for wages and payroll taxes
- health insurance
- mistakes
- employment lawsuits
The minimum wage has increased steadily in the U.S. even as the average skill of a high school graduate has fallen. The federal minimum wage was increased in July 24, 2009, 1.5 years into our current economic depression. More important, perhaps, are the heavy increases in payroll taxes over the years, notably for Medicare and Social Security.
Most companies cannot culturally stomach denying health insurance to certain classes of worker. Apparently it is okay to pay the CEO 319X what the average worker gets, but it is not okay to tell low-skill workers “You aren’t important enough for us to buy you health care in the world’s most expensive and least efficient system.”
Most subtly, and perhaps most significantly, the potential cost of a mistake by an individual worker has skyrocketed. In industrial plants, the link between individual employee action and billions in losses is fairly obvious, e.g., with the Bhopal explosion. A tiny misstep in a chip factory and a wafer containing hundreds of valuable integrated circuits becomes worthless scrap. Computer networks, however, have made the potential costs of a clueless or careless office worker dramatically higher. Suppose that a company hires a low-skill not-very-alert office worker for $10/hour. This person accepts an email invitation to follow a hyperlink. One click later and the company’s network is infected with a virus. Best case: IT department spends $50,000 cleaning up; worst case: customer lists, customer credit cards, and other private data are compromised, costing millions of dollars.
As the government has increased the number of ways in which an employee can sue an employer, the expected cost of litigation from each additional employee has gone up. The cost of trying out a worker who might not work out is much higher than formerly, especially if that worker is older, female, or belongs to a government-recognized minority group. It might be smarter to employ fewer higher skill workers because the chance of litigation is lower with 100 workers than with 200 workers.
What’s the practical implication of all this? Policies that encouraged companies to hire the unemployed after the Jimmy Carter “malaise” years may no longer be effective. Health care spending as a percentage of GDP in 1980 was 8.8 percent (source) compared to nearly 20 percent today. Only a handful of companies had Internet access and there were as yet no viruses.
Or we can rephrase the entire posting as “How comfortable would you feel working at your present job alongside someone whom you would rate as among the least competent 25 percent from your high school?”
[Update: An economist sent me this article on how U.S. firms have job openings, but can’t find skilled workers to fill them.]
I find it hard to believe we’re anywhere near that point. I’d imagine that most people are still employable as migrant workers in agriculture, for instance. They may not *want* that job, and they might take a long time to become physically fit enough to do it, but there still seem to be jobs that hire minimally skilled workers.
It’s funny, but every time I tell people that eventually, artificial intelligence will be created that will automate ALL jobs, people tell me, either (a) it’s impossible (human intelligence can’t be replicated in a machine), (b) humans will remain competitive by putting computer chips in our brains, nanobots in our bloodstreams, or otherwise “merging” with technology (an idea popularized by Ray Kurzweil), or (c) that there is nothing to worry about, because every time in the past when machines were invented that automated something, people feared employment would go down, but it didn’t. (Or at least it didn’t because of machines and automation — employment obviously went down during the Great Depression and other times due to financial mismanagement.)
Anyway, I’ll just come out and say it… I think today’s low-skilled* people are 21st cenury draft horses — and that the minimum requirements for meaningful employment will increase from now on, eventually exceeding even the smartest people. I think AI is possible and will someday be achieved, though I think it will take a long time — at least another 50 years. I think Kurzweil and his transhumanist followers are wrong in thinking that humans will “merge” with technology and usher in some sort of utopia. I think automation of jobs will do exactly what you would expect — leave people unemployed. But people who have income from business ownership, dividends, or financial instruments (trust funds, annuities, whatever), will be just fine. People best positioned to exploit automation and reductions in employment will become extremely rich.
—
*The key skills of the modern world are sales/marketing/office politics/persuasion in general, not whatever “skill” is written in the job description.
Lawrence: Perhaps your point was that a single individual currently unemployed could probably find work by traveling to wherever the crops need harvesting. That sounds reasonable, but http://en.wikipedia.org/wiki/Agriculture_in_the_United_States says that there are a total of approximately 800,000 jobs in agriculture in the U.S. Every farm would have to multiply its workforce by approximately 20 in order to absorb the 15 million Americans who are currently unemployed.
It is only natural, the technology eliminates the need of the extra people. But the unemployed are not like horses, they are more intelligent and they have a political system that have to look over for them. Sooner or later, they will find creative way to get themselves employed. And the government will more and less likely to take care of rest. I don’t like the socialist approach, but I think it might be necessary. It might be just as important to pay those unfit employee to stay out of making catastrophic mistakes.
To me, the mere fact that the U.S. has special visas for Mexicans to come harvest crops instead of forcing farms to hire American workers in an act of standard vote-buying protectionism says that there are jobs that employers are offering at minimum wage, but unemployed American workers aren’t willing to take. Unlike H1-B visas, if the farms are already paying minimum wage, there’s a much lower return on lobbying, so it’s hard to explain this particular visa as a product of the farm lobby to lower wages if Americans would work for minimum wage.
That leads me to believe that the unemployed in general aren’t suffering from the draft-horse problem specifically, at least not yet – if they’d reduce their living standards to that of Mexican migrants, they’d still make more than their feed. Of course, that’s not saying much.
I’m not suggesting that agriculture could employ all those workers; merely that all of them have decided not to take these jobs. (15 million people could reject one job – they don’t need 15 million jobs available to reject before they’ve demonstrated that minimum wage under poor conditions is worse in their eyes than unemployment.)
The notion that it could become commonplace in the near future to debate the question “what do we do with the extra people?” is terrifying. The historical answers to that question are not pretty, and we do not at present have an especially compassionate culture.
Lean times ahead, for those outside of the limo. The effects of widespread poverty (by which I mean those deprived of both material and the power of self-determination) upon a society are well established, and will not be averted.
I’ve been thinking further on the ‘mistake’ angle. It seems like the more automation you have, the more potential you have towards a mistake causing a huge amount of damage. Phil’s examples of simple mistakes allowing viruses to do damage is a good one. Also people who have, for example, sysadmin access to systems can blow up a whole server or several with a few bad keystrokes. I’m sure things like that happen all the time, but companies try to keep mum and it’s not in the news. But in the future, with robots running around, or flying around, potentially with weapons, and some doofus gives them a bad command…
We try to develop systems that have some resiliency towards errors. Here’s an interview I read of Peter Norvig who says at Google, failure is always an option.
http://www.slate.com/blogs/blogs/thewrongstuff/archive/2010/08/03/error-message-google-research-director-peter-norvig-on-being-wrong.aspx
In general I think robotics today is where PC’s were in the 1970’s — I don’t think the world is ready for the effects on employment when this technology matures. Wal-Mart will be replacing everyone who stocks shelves with robots that stock shelves. Everyone who drives a vehicle (taxi driver, bus driver, limo driver, long-haul semi driver) will be put out of work by autonomous vehicles. But when? It’s hard to say. The technology has a long way to go before it reaches the point where it can replace people. But you can see it improve every year.
How easy would it be to see the cost of mistakes in today’s computing environment from the perspective of 1975 or 1980? Viruses back then were an idea intellectual geeks talked about. “Imagine a program that can self-replicate itself! It’s like a life form!” And today we’ve got millions of them, and they’re a nuisance. I wonder what kind of mistakes will be possible in 2030?
My understanding is that, in rural areas, you can’t find white people or black people that will work in agricultural jobs at $12, so we end up importing illegal aliens: the work doesn’t require remarkable skill, but it does require a certain stick-to-itiveness that is hard to find. Farmers might find more US-ians who could do this kind of work if they paid more like $20 or $30 but maybe not.
As for health insurance, this is entirely capricious. The largest companies can (sorta) afford to insure minimum wage workers; on the other hand, the most talented bunch of I.T. people that I have ever worked with was at a company w/o health insurance. The owners of the company, any many of the employees, had spouses who worked at the two major educational institutions in a university town, so they saw it as something expensive that they couldn’t afford. Unfortunately, I couldn’t afford to work there.
Entrepreneurs, consultants and other talented people may find that health insurance is unavailable at any reasonable price on the market. In fact, there seems to be so little interest in selling insurance to individuals that health insurance people will sometimes hang up on the phone on you when you call up trying to buy their product. Now, if you’re a lawyer or a dentist and you’re making $250k+ a year it might be affordable, but if you’ve got that much money you might as well save it and just pay out of pocket when you need it.
The farms do not pay an hourly wage. They pay by the work done. For instance, if you are hoeing the weeds out of a row of sugarbeets, you get paid by the row.
What happened in the last few years technologically that would render 15 million Americans obsolete?
The answer in your horse scenario was the internal combustion engine. Is there such an answer now?
Mike J, you ask “What happened in the last few years technologically that would render 15 million Americans obsolete?” Nothing happened in the last few years. But if you take out the words “in the last few years”, then you get a meaningful answer: information technology. Computers, the internet, cellphones, etc. The reason why “the last few years” is misleading is because the US economy has been almost stagnant since the 1970’s (except for the very rich), and in decline since about 2000 — however this was all masked by debt. Everyone borrowed, and maintained the appearance of the same or greater standard of living. It wasn’t until the financial system blew up in 2008 and the lending stopped dead it its tracks that the real state of the economy was exposed.
Information technology creates “winner take all” markets that concentrate wealth. It’s analogous to how the industrial revolution created ‘economies of scale’ that concentrated wealth. And gave us Rockefeller, Carnegie, etc. Only now it’s Lloyd Blankstein et al. 🙂
There’s never been more work to do. As soon as oil prices start soaring America will get busy building the energy infrastructure of the future. And we’ll
need all of those “horses” back.
the first world has a serious people problem. as we migrate to globalized production, automation, virtualization, and other means of reducing demand for workers, we still have all these mouths to feed. a short-term solution for the first world will be to severely constrain immigration…but that won’t address the long-term environmental issues associated with population growth
A likely scenario is that a much smaller number of people will live very high qualities of life. Take the horse example. There are far fewer horses now but they arguably live better lives as the toys of rich people than as draft horses. I don’t know the history, perhaps there were horse pograms or something but I’d guess they were just bread more slowly and their numbers eventually dwindled to a sustainable level. It is not unlikely that the same scenario will happen with people. In fact you could argue the process has already begun (and at this point rather benignly) with first world country birth rates and in many cases populations in decline. The system is highly non-linear of course so it could be that this could happen catastrophically and violently or smoothly and happily, but looking forward its pretty likely to happen.
Of course you could picture more imaginative scenarios. I’m sure there are some domestic animals which are more prevalent than their wild counterparts were (dogs vs. wolves maybe?) certainly our favorite monoculture crops and food animals. Or most wildly imagine that technology allows colonization of space, then you could have a vastly lower ratio of humans to technology, while having way more people than previously imaginable.
Mike J: The technology is the computer and just like the ICE I’m sure it will take some time to have its full effect.
The microprocessor.
In the late 90s the USA had about 3% unemployment. It was sort of a natural test of the question: Do most people want to work? The answer was, of course, yes. What has changed so dramatically in the last 10 years that unemployment has suddenly become “structural”? The financial part of the economy ruined it for everybody else. Structural issues exist of course, but they are not the major reason so many people can’t get work.
if IT is hot why are IT jobs disappearing fast ?
http://research.stlouisfed.org/fred2/series/USINFO?cid=11
I’ve got a mule and her name is Sal
Fifteen miles on the Erie Canal
—
What happened in the last few years technologically that would render 15 million Americans obsolete?
—
2 billion Asians with access to modern means of production.
Mike J: Why the dramatic and recent change? That’s an excellent question and I also appreciate the other folks trying to answer it.
Obviously the Collapse of 2008 precipitated a lot of layoffs and restructuring. Companies often need a strong push to effect change and the meltdown and subsequent taxpayer-funded bailout would have woken up any business manager.
If I had to pick one thing I would say the increasing utility of Internet. Just ten years ago the Internet wasn’t all that useful for people trying to work. Very few could get email on their phones and even fewer could get access to documents outside of the company’s private network.
Let’s look just at Google’s services. I can use Gmail and Google Calendar for free, thus eliminating the need to pay a team of IT specialists (and a witchdoctor for good measure) to keep Microsoft Exchange Server alive. I can access my Google Docs from anyway. Today I was having trouble checking it at JetBlue’s automated kiosk in Logan Airport. The alternative was a line of perhaps 300 people waiting to talk to human agents. I needed my confirmation number to use the machine and skip the line. I got the number from my Gmail on my Google G1 Android phone. Think about all of the people whose job it was to keep track of information by filing it or producing it upon request. A lot of those jobs are permanently gone thanks to Google’s ability to roar through gigabytes of email and documents.
Obviously Google has opened the door to a lot of new enterprises and reduced the amount of capital required for a lot of startup companies (since they won’t need an IT department), but it has also closed some doors to those whose skills were rather basic.
I think a pertinent question to ask is whether the 15 million Americans are all similar in some key way, or they’re very diverse. If they’re all similar, then that would support “the draft horse” theory. If they’re diverse, then doesn’t that mean we’re not talking about any particular issue of technological obsolescence, but rather a widespread shrinking of the entire economy, perhaps due more to outside forces (like globalization) than inside forces (like disruptive technology)?
Obviously relevant: The Doubling of the Global Workforce
I don’t understand why, under these conditions, capitalism would be even slightly desirable. A system where you need a “job” to live but there is otherwise no actual need for your labor is a system that doesn’t make any sense at all. We’ve kept people yoked to this “job” nonsense by inventing more and more ridiculous busy-work for them to do, from pet-grooming to manufacturing pointless tanks that rust away, but now you claim that there is no more busy-work left. We can no longer invent some humiliating nonsense for people to do.
So please tell me why we have to have money and jobs. You are explicitly saying that not everyone is required to work all day every day; so let’s split the jobs up, everyone work a few hours a day, and the rest of the time let’s hang out, write poetry, have sex, play games.
Mike J: Are America’s long-term unemployed similar or diverse? I’m sure that they are as diverse as the employed. The only thing that they have in common, really, is that no employer believes that they can be hired at a profit, at least not here in the U.S. under prevailing customers and laws.
notebook: I don’t think that I said that there aren’t jobs out there. Almost every family, for example, would be happy to have additional help with cooking, cleaning, and babysitting. The fact that they don’t hire additional help with these chores shows that they aren’t able to find sufficiently high quality workers at a sufficiently low cost.
I’ve seen this point alluded to in the comments but not said clearly: People can learn to be different things. The identities we take on for our jobs- factory worker, teacher, claims adjuster, etc. are transitory. A draft horse by contrast is still a draft horse when it clocks out. It’s possibilities for reinvention are limited not even to the horse’s imagination but to it’s human owner for whom the horse’s survival and reproduction are of limited consequence.
Are there any historical “draft horse” examples involving real people?
—
Are there any historical “draft horse” examples involving real people?
—
Someone with a 70 IQ would have been of some utility in agrarian/gatherer societies.
Their utility is quite diminished in today’s economy.
bjcefola: The best example of a “draft horse” would be a person who had concluded that it was more sensible to depend on family members and government handouts than to seek employment. Those are the folks that the Bureau of Labor Statistics calls “discouraged” and there are about 1.2 million in the U.S. right now (see ftp://ftp.bls.gov/pub/suppl/empsit.cpseea38.txt ).
A more enterprising “draft horse” would be someone who completes a program of education, can’t find a job, and decides to go back to school and get additional credentials.
Also, see this blog post:
Structural Unemployment: The Economists Just Don’t Get It
http://econfuture.wordpress.com/2010/08/04/structural-unemployment-the-economists-just-dont-get-it/
Wayne,
I think you are exactly correct in your prediction about automation and its impact on unemployment. But here you have it wrong:
“But people who have income from business ownership, dividends, or financial instruments (trust funds, annuities, whatever), will be just fine. People best positioned to exploit automation and reductions in employment will become extremely rich.”
What you are missing is the collapse in consumer demand and the financial implosion. Think about it: You think mortgage defaults are bad now? What will happen when unemployment is 50%? Who will be buying products and services? Where will corporate profits come from?
Hope you will read the book, The Lights in the Tunnel….think about it!
One more….
Did Advancing Technology Contribute to the Financial Crisis?
http://econfuture.wordpress.com/2010/04/06/did-advancing-technology-contribute-to-the-financial-crisis/
What’s the new equivalent of the internal combustion engine?
The old industrial revolution was about replacing human and animal muscle power with inanimate power. The new industrial revolution is about replacing human ability to monitor what’s going on from moment to moment, to decide what to do, and to anticipate and solve problems in order to complete tasks.
If we take digital computers and the internet as the equivalent of the steam engine, the “new internal combustion engine” is a combination of five technologies: digital radio (cell phones), internet search, statistical machine reasoning, machine vision/sensing, and robotics. Combine these five things in the right ways and the resulting devices can do nearly any task in any industry. (Except sport or artistic performance – what would be the point?, and some of the professions. But not everybody can be a start, by definition, and not everybody can be a materials scientist, judge, or politician.)
Once prototypes exist–well, we know how to cut the cost of machines. Bring the cost low enough, and no human can compete – especially since machines will work 24×7 and with absolutely reliable quality. (Think of medieval scribes versus the printing press.)
All five component technologies are in commercial use today, but three of them exist only in embryonic form. The new composite general-purpose technology still lies in the future. The current high level of unemployment is due to lack of demand, pure and simple.
There is a mismatch between skills and aptitude wanted and available. This isn’t important, because the coming collapse of the dollar will make factory workers wanted, and the coming collapse of government funds will make symbol analysts less wanted.
Think of 500 billion dollars worth of high paid local, state, and federal workers getting laid off, at the same time as 500 billion dollars of imported goods needs to be replaced, when the foreigners get tired of buying our treasury bills.
People used to look forward to being out of work due to AI. Here’s Oscar Willde “The soul of man under socialism”:
The chief advantage that would result from the establishment of Socialism is, undoubtedly, the fact that Socialism would relieve us from th[e] sordid necessity of living for others…. So while Humanity will be amusing itself, or enjoying cultivated leisure – which, and not labour, is the aim of man – or making beautiful things, or reading beautiful things, or simply contemplating the world with admiration and delight, machinery will be doing all the necessary and unpleasant work. The fact is, that civilisation requires slaves. The Greeks were quite right there. Unless there are slaves to do the ugly, horrible, uninteresting work, culture and contemplation become almost impossible. Human slavery is wrong, insecure, and demoralising. On mechanical slavery, on the slavery of the machine, the future of the world depends.
Those who believe that human-level artificial intelligence is anywhere near realization are mistaken. The aural and visual signal-processing and pattern recognition capabilities of the brain are awesome compared to those of any computer system. Though OCR software has been around for about forty years, it still cannot deal with anything but very clean input in a limited range of fonts, and especially not with handwriting. Speech recognition software is similarly limited, unable to cope with natural language even in a noise-free environment, and utterly useless in a noisy environment. Machine translation, though touted for decades as being just around the corner, is still nearly useless for most real-world texts, as anyone can easily learn by trying out the various services on the web.
Being fairly fluent in Japanese, I was asked in the early ’80s whether I might be able to translate an electronics-related patent that had been faxed on coarse resolution to my employer’s legal department, apparently after having already been faxed and copied several times within Japan. My initial reaction was that it was almost completely illegible — 90+% of the characters resembling spiders squashed onto the paper — but that I’d see what I could puzzle out. On the first reading I could decipher perhaps 20%, working backwards in the manner of “if I were to smear the characters for ‘amplifier’ might they not look like those three there, and that might make sense in what I think the context may be”. On the second reading recognition rose to about 60%. The third time through, I felt that I had deciphered nearly 100% of the text, but I must emphasize the ‘felt’, as the more closely I looked at any individual spider-squashing, the less confidence I had in my interpretation of it. So I transliterated my reading of the Japanese into romaji and emailed it to the firm’s Tokyo office for their opinion. Their reply expressed awe and amazement — I had read it correctly.
Had I not had the requisite technical background to understand the content of the document I would never have been able even to read it, because nearly every character was illegible if viewed in isolation. It was only the semantic, syntactic, and visual pattern-recognition abilities of my brain working in concert with domain expertise that made decipherment possible.
Although a Japanese of less than average ability probably would not have been able to decipher that particular document, he or she would almost probably be able to read a similarly illegible text relating to more mundane matters.
In contrast, I recently acquired the latest edition of a leading OCR software package supposedly able to recognize Japanese text. It fails even with copy scanned directly from a printed book if the scanned lines are more than about one degree off from vertical.
Hey Phil,
Has modesty prevented you from pointing out the fact that the business insider has published your opinion? 🙂
http://www.businessinsider.com/most-of-americas-unemployed-are-21st-century-draft-horses-2010-8
Robert,
I looked up The Lights in the Tunnel on Amazon
http://www.amazon.com/Lights-Tunnel-Automation-Accelerating-Technology/dp/1448659817/ref=sr_1_1?ie=UTF8&s=books&qid=1281503029&sr=8-1
and it looks like an interesting book. I’ve added it to my shopping cart. I do have this little problem that my Amazon shopping cart has 384 books in it… so not actually sure when I’ll get around to reading it. (I need to win the lottery so I can read books).
Your comment about the lack of consumer demand — actually there is a fancy term for this in economics and it is “inadequate aggregate demand”. It’s actually a function of concentration of wealth. Once you have wealth sufficiently concentrated in the rich people, then the masses no longer have enough money to buy things. Once the masses don’t have enough money to buy things, then it becomes difficult for the rich people to invest in new businesses that make things people want to buy. Anyway I think you understand all that, but I’m just making the point that it’s not something that happens *only* as a result of automation. If wealth is concentrated politically, you get the same result. What I find interesting about this is, this scenario is not necessarily catastrophic for the rich people — they can’t meaningfully invest their money but they can gamble it — and many societies can persist in this state for a long time, perhaps indefinitely. There are many societies in the world whose basic structure is a tiny tiny handful of ultra-rich people and teeming masses of poor people.
So there’s that, but I think automation adds a new twist, which is that once automation reaches a sufficiently advanced level, artificially intelligent agents can *themselves* become economic actors, buying and selling in the markets. This means there is no problem with the exit of low-skilled human labor from the economy — the economy can continue with automated agents buying and selling, and goods and services produced that they want to buy.
So in short, I think automation will remove humans from the *labor* market (but not the capital markets, and there will still be income from licensing of intellectual property, dividends, trust funds, ets — everything will fine except the labor market), and the economy will not implode but will keep on humming.
If you think this analysis is incorrect or the book Lights in the Tunnel makes a compelling case that this line of reasoning is wrong, I’d like to hear it. I think there may be problems or ‘speed bumps’ along the way as the system adapts and makes adjustments for new automation, but I don’t foresee the whole economy imploding as you describe.
To jm regarding aural and visual signal-processing and pattern recognition capabilities of software and OCR, yes I agree there are many deficiencies in AI systems today and that is because we don’t know how the brain works. It will probably take us a long time to figure it out, and we’ll figure it out in bits and pieces. And as bits and pieces are figured out, we’ll use that knowledge to automate jobs. For example once *good* vision processing really works, the “robocar” concept will work. And that means everyone who makes a living driving any kind of vehicle will be out of work as I mentioned before — that’s just one example. Anyway, my point is, how the brain works is a mystery today, but it won’t be a mystery forever. Once the algorithms are known, they can be encoded in computer hardware/software. Once the algorithms are encoded in hardware/software, then what matters is cost. Even if McDonald’s build an automated restaurant, is the automated restaurant cheaper than hiring teenagers to flip burgers? As long as teenagers are cheaper, they have jobs. But Moore’s Law doubles the information processing capacity per price every 2.18 years, and algorithms are continuously improved. It’s just a matter of time.
To wkwillis, you say the coming collapse of the dollar will make factory workers wanted. No way. Employment in factories is decreasing, even in “low wage” countries like China, because of robotics. People are not aware how much industrial robotics has advanced, because industrial robots are tucked away in factories and we never see them. http://www.youtube.com/watch?v=HXjECjyvvUQ
Chemists may be facing this draft horse situation head on:
http://pipeline.corante.com/archives/2010/08/11/if_youre_not_a_chemist_what_next.php#comments
I am mainly familiar with the biotech industry but I can definitely tell you is being automated on all fronts. There will still be a need for a few proven high-quality PhDs in research and development but even their numbers will also be reduced eventually. I can’t help laugh when economists and career guides cite biotech as a growth industry in terms of employment. Profit and company capital in biotech and pharma will probably increase but there is no way the number employed in these industries will do likewise. In fact, the numbers will even be reduced.
We are just starting to see a new phenomenon: Uncompensated work. This is not totally new of course; there have always been people providing public unpaid public service: volunteer fire departments for example. However, this is beginning to expand rapidly to people with spare time but no income.
Consider self-publishing on the internet (not just writings but music and movies). The cost of publishing is now so low that getting a monetary advance is not required, but consumers expect the content to be free of charge and so no revenue comes in (except perhaps from advertising).
Amazon created a “Mechanical Turk” whereby anyone can get paid at a tiny piecework rate to do boring pattern recognition tasks. But is the payment really necessary?
An article in the journal Nature, “Predicting protein structures with a multiplayer online game”, reports on a team that solved a protein-folding problem in an interesting way. A brute-force computer algorithm takes far too long; however the computer can easily display a portion of the protein with low and high energy points shown in different colors. A human can quickly find the most likely ways to fold. 57,000 computer users (all credited as authors of the paper) played the protein folding video game and together got the answer.
So: if you have a pattern-recognition problem that is far too complex to solve by brute force, but you can create a multiplayer computer game that is isomorphic to the problem, and it’s so much fun to play that it becomes the next Farmville, you have a vast supply of free brainpower available.
Your post tipped me into writing this post on why it is absurd to fear from AIs that interacting with them will destroy the human economy, just like it is absurd to fear from foreigners that trading with them will destroy the domestic economy.
http://fare.livejournal.com/159046.html
Wayne,
Sorry for the long delay. Regarding your point about societies existing with all the wealth highly concentrated. I’m not sure that there has been a case where a country moved having a strong middle class and reasonable distribution of wealth to essentially a 3rd world model with a few very wealthy people and nearly everyone else impoverished.
The implications of that are pretty dire. Consider mortgage defaults for one thing. Would the financial system survive that? What about the tax revenue base? What about a soveign debt crisis? How would the US fund defense? Also, nearly all mass-market industries would collapse.
Would a political revolution be unthinkable in that scenario? In the 1930s, revolution was definitely considered possible — FDR worried he might be the last president.
I think the scenario laid out in “The Lights in the Tunnel” has a lot of merit and is cause for concern.
Robert: A “poor” person in the U.S. today may still have a higher standard of living than a middle class American circa 1950. Also, a lot of America’s poorer residents are either immigrants or children of recent immigrants. They are living at a much higher material standard of living than they enjoyed in their home country. Thus a headline of widening income disparity in the U.S. does not indicate a vast number of “impoverished” people by worldwide standards. It may simply indicate that the top 1% have gotten richer while the other 99% are no better (or worse) off than previously. See http://www.wisebread.com/our-high-high-standard-of-living-1 for example.
Will Americans vote for a Socialist revolution? They sort of already have!
How can the ever-expanding U.S. government fund itself when most sectors of the economy are stagnant? Eventually the expansion rate of government may slow, but that’s not a crisis except for the young people who were counting on government jobs and will have to live with their parents until age 35 instead.
Similarly your idea that “mass-market industries would collapse” doesn’t make sense. The world has never been richer than now. Mass-market U.S. companies are finding newly wealthy consumers in Mexico, India, China, Africa, and Brazil. The U.S. is not such a great place to do business anymore, but there is literally a world of opportunity out there. Look at http://seekingalpha.com/article/225893-procter-gamble-looking-ahead-to-september-2010-quarterly-results for example.
It is tempting to assume that dynamism is the natural state of the U.S. economy, but nothing stops us from stagnating for a few decades the way that England did after World War II.
Nonsense. The need for streetsweepers, caregivers, assistant cooks and so on is ilimited. Anyone wishing to work (such as Mexican illegal immigrants) easily find a job and can pay their way. JNon working people is an artificial social product. It does not exist in nature nor in a normal society.