Summary: Artificial intelligence has arrived. The mid-twentieth century horse apocalypse shows us what it might do to employment. It is history’s rebuttal to the pollyannas about tech. This is a follow-up to yesterday’s post — Films show us how smart machines will reshape the world.
“We tend to overestimate the effect of a technology in the short run and underestimate the effect in the long run.”
— Attributed to Roy Charles Amara as paraphrased by Robert X. Cringely.
Once there were ten million plowhorses in America.
Most lost their jobs.
Many became glue or dog food.
The next and largest wave of automation is slowly becoming visible. But the coming of this revolution were visible to some people long ago. One of the first was James Blish, as described in his science fiction novel A Life for the Stars
.The cab came floating down out of the sky at the intersection and maneuvered itself to rest at the curb next to them with a finicky precision. There was, of course, nobody in it; like everything else in the world requiring an I.Q. of less than 150, it was computer-controlled.
The world-wide dominance of such machines, Chris’s father had often said, had been one of the chief contributors to the present and apparently permanent depression: the coming of semi-intelligent machines into business and technology had created a second Industrial Revolution, in which only the most highly creative human beings, and those most fitted at administration, found themselves with any skills to sell which were worth the world’s money to buy.
Fifty-five years later this era has begun. Algorithms have already replaced many in a few skilled jobs, such as bank credit officers. Tools to automate aspects of many more skilled jobs are being tested — such as radiologists and journalists — so that (like horses) their numbers employed will drop precipitously during the next generation or so. This is only the beginning. This will reshape the world. What happened before shows us how.
The horse apocalypse.
The great mid-twentieth century horse apocalypse is described by Erik Brynjolfsson and Andrew McAfee (both of MIT) in “Will Humans Go the Way of Horses?” in Foreign Affairs, July/Aug 2015 — “Labor in the Second Machine Age.”
“For many decades, horse labor appeared impervious to technological change. Even as the telegraph supplanted the Pony Express and railroads replaced the stagecoach and the Conestoga wagon, the U.S. equine population grew seemingly without end, increasing sixfold between 1840 and 1900 to more than 21 million horses and mules. The animals were vital not only on farms but also in the country’s rapidly growing urban centers, where they carried goods and people on hackney carriages and horse-drawn omnibuses.
“But then, with the introduction and spread of the internal combustion engine, the trend rapidly reversed. As engines found their way into automobiles in the city and tractors in the countryside, horses became largely irrelevant. By 1960, the United States counted just three million horses, a decline of nearly 88% in just over half a century. If there had been a debate in the early 1900s about the fate of the horse in the face of new industrial technologies, someone might have formulated a “lump of equine labor fallacy,” based on the animal’s resilience up till then. But the fallacy itself would soon be proved false: once the right technology came along, most horses were doomed as labor.
“Is a similar tipping point possible for human labor? Are autonomous vehicles, self-service kiosks, warehouse robots, and supercomputers the harbingers of a wave of technological progress that will finally sweep humans out of the economy?”
This insight was made first by Wassily Leontief (1973 Nobel Laureate in Economics) in his famous “Long-Term Impact of Technology on Employment and Unemployment“, National Academy of Engineering (1983). He describes the great wave of job extinction — of horses. Cutting their wages could not save their jobs, nor will it save our.
“Any worker who now performs his task by following specific instructions can, in principle, be replaced by a machine. That means the role of humans as the most important factor of production is bound to diminish in the same way that the role of horses in agricultural production was first diminished and then eliminated by the introduction of tractors. The general theoretical proposition that the worker who loses his job in one industry will necessarily be able to find employment, possibly after appropriate retraining, in some other industry is as invalid as would be the assertion that horses who lost their jobs in transportation and agriculture could necessarily have been put to another economically productive use.
“Reduction in the price of labor — that is, in the real wage rates — can and in certain instances did postpone its replacement by machines for the same reason that a reduction of oats rations allocated to horses could delay their replacement by tractors. But this would be only a temporary slowdown in the process; improvements in the efficiency of tractors and other inanimate means of production can be expected to proceed without any limits, while reductions in feed rations or wages have definite limits.”
See a longer excerpt here from this great paper. It is well worth reading.
Some prefer their eyes closed.
The need to adapt is not obvious to everybody. In her (otherwise excellent) 1989 book In The Age Of The Smart Machine: The Future Of Work And Power
In their 2004 book, The New Division of Labor: How Computers Are Creating the Next Job Market
In January 2016 Elizabeth Garbee at Slate wrote “This Is Not the Fourth Industrial Revolution” — “The meaningless phrase got tossed around a lot at this year’s World Economic Forum.”
Conclusions
Automation will be one of the great challenges of the 21st century. While fantastic, it is no more dangerous than other challenges we have successfully surmounted. But it could cause devastating social turmoil before we successfully adapt. The sooner we start to prepare, the less traumatic will be the transition. Don’t listen to the people who say that because AI revolution has been predicted for so long, it will not come.
“On September 23 {in 1066 William the Conqueror’s} fleet hove in sight, and all came safely to anchor in Pevensey Bay. There was no opposition to the landing. The local fyrd had been called out this year 4 times already to watch the coast, and having, in true English style, come to the conclusion that the danger was past because it had not yet arrived had gone back to their homes.”
— From A History of the English-Speaking Peoples
by Winston Churchill.
For More Information
The new industrial revolution has begun. New research shows more robots = fewer jobs. Also see the famous book by Wassily Leontief (Nobel laureate in economics), The Future Impact of Automation on Workers
If you liked this post, like us on Facebook and follow us on Twitter. See all posts about robots and automation, and especially these…
- A warning about the robot revolution from a great economist.
- How Robots & Algorithms Are Taking Over.
- Economists show the perils and potential of the coming robot revolution.
- Three visions of our future after the robot revolution.
- The coming Great Extinction – of jobs.
Books about the coming great wave of automation.
Rise of the Robots: Technology and the Threat of a Jobless Future
The Second Machine Age: Work, Progress, and Prosperity in a Time of Brilliant Technologies
This is an interesting problem. I have often wrestled with the binary opposition of two paths of thinking about it. On the one hand, many predictors suggest that as with all other disruptive innovations in the past, such as the automobile, the computer, etc. that new jobs will come to replace the old jobs that are lost (blacksmiths, bridle makers, etc.). On the other hand it is my subjective thought that this is different (always a dangerous thing to say, but perhaps in this case true?).
With AI, you are now replacing humans in jobs, not just replacing the jobs themselves. Perhaps other newer jobs might come along that humans will do, but i sense that this may be a long time coming, and many will still remain displaced.
Given the state of AI at this time, it may still be a way off, but as with William the Conqueror, it will eventually come, and will we be ready, and will be able to deal with the consequences? Perhaps some thinking outside the box might be necessary- for all displaced workers, some sort of basic income- a concept i have held in poor regard as a general principal- however if a larger percentage of the population is displaced from the labor market, it might become necessary to think about something along those lines. Interesting times…
Sorry to double dip here, but there was another thought about this that has crossed my mind often in the past. As a healthcare provider, we have seen a progressive increase in the relative proportion of GDP dedicated to the healthcare sector. While it is described as a problem which is squeezing out capital to be used for other things, perhaps increases in service jobs, which would include healthcare, could create a means of providing employment, even as employment in manufacturing or transportation disappears.
Again, just thinking out loud, maybe it is not such a terrible thing after all, if it manages to keep people gainfully employed while AI and robots provide other goods and services that they are capable of handling. I know this sounds rather “progressive” in nature, and i am really not progressive wearing my usual hat, however AI may be soon presenting us with difficult choices, and problematic situations, and some solutions may become necessary that would not be otherwise palatable. Just putting it out for discussion.
Barry,
“perhaps increases in service jobs”
Their time too will come. The pressure to automate health care will be immense as the Boomers age and consume vast amounts of health care. We can’t afford the existing health care costs.
Also, When people speak of jobs lost but not replaced — they’re speaking of good jobs. Our plutocrats are already hiring servants. If automation continues to increase the wealth of the 1%, personal servants might become status displays of conspicuous consumption for the mega-rich. Like at Pemberley in Pride and Prejudice, hordes of menials might tend their palaces. Turning their faces to the wall as their lords pass, and other displays of class (since that is their actual role).
Barry,
Nicely stated.
I’m always puzzled that people regard this time it’s different with suspicion with regards to tech. The tech revolutions have been inflection points in history after which things were different. For example –tech, esp modern contraceptives, changed the role of women unlike anything seen in history. Does anyone expect past patterns to reassert themselves?
The future is already here.
“Palantir: the ‘special ops’ tech giant that wields as much real-world power as Google” in The Guardian — “Peter Thiel’s CIA-backed, data-mining firm honed its ‘crime predicting’ techniques against insurgents in Iraq. The same methods are now being sold to police departments. Will they inflame already tense relations between the public and the police?”
Mike,
Color me skeptical that Palantir’s software is that powerful. It’s just one of the long list of tech marvels that don’t seem to have accomplished much in Iraq. We’ll see if they do much for US police — other than boost tje wealth of Silicon Valley’s plutocrats.
Mike,
Re-reading this story, my guess is that it is in effect a press lease from Palantir. It’s loaded with dubious or exaggerated claims, like this: “Bernie Madoff was imprisoned with the help of Palantir.” It’s software may have been of use sorting thru the records, but imprisoning him required going into his trading room and seeing that it was empty — there was nothing there. The trading records were fake.
Mike,
Palantir has deployed this kind of public relations blitz before. See “Killer App” by Shane Harris in The Washingtonian, Jan 2012 — “Have a bunch of Silicon Valley geeks at Palantir Technologies figured out how to stop terrorists?”
Stephen Arnold did some debunking of its claims: “Palantir Applies Lipstick, Much Lipstick“.
The article discusses Palantir’s tools at work in 2011. Six years later we can look around and see their revolutionary impact — where?
FM-
I was an early adopter from 2008-09. It is augmented reality not artificial.
In my mind, this is a larger threat than the robots.
The impact is on civil liberties and privacy.
Perhaps the remaining creatives, inventors and others who still have jobs will adopt the jobless as pets as has happened to horses. I am not serious but it seems like an interesting idea for a movie.
As nations become more advanced their birth rates diminish. Is it possible that the world’s human population will shrink until it comes into balance with the number of productive jobs available?
I sometimes wonder what a fully automated nursing home will look like. Hopefully I will never find out.
Cities In Flight. It’s interesting the parallels with today’s world, as I always considered Blish’s future firmly bedded in dust bowl/depression America, but robots…
The thing about this that I think is fascinating is whose jobs will be replaced. In the past it’s always been the manual workers that have gone up against the wall, no real power, no voice no recourse except strike, protest and violence. This time the scythe will be passing through many of the articulate, voting, well educated and monied middle class. Not inclined to strike or riot, but voting… 2nd term for Trump, promising not to allow wholesale automation of white collar jobs?
It will be interesting to see what government responses are.
Then there’s immigration, currently rich countries import a lot of skilled and semi-skilled people, poorer countries needing the money sent home by immigrants. The argument largely being that population growth is declining and more people are need to replace us old gits as we retire.
But if we’re being replaced by robots, who’s going need to hire?
Steve,
(1) “2nd term for Trump, promising not to allow wholesale automation of white collar jobs?”
Not likely. Trump’s election populism has been revealed as a con, staffing his administration with plutocrats and generals who administer standard GOP policies. Trump’s first three policies have been to strip tens of millions of health care, ramp up our foreign wars, and cut taxes on the rich.
But more generally, can people vote to prevent automation? This was a common response during previous waves of automation — with a uniform failure rate. As described in the apocryphal origin of the word “sabotage”: workers (from Netherlands or Lyon, or luddites in England) would throw their “sabots” (clogs) in the machines to break them). One of the fanciful bot confidently stated beliefs of Brynjolfsson and McAfee is that this will work in the 21st century. They don’t explain how. Or why such measure might work despite their history of failure.
(2) “he argument largely being that population growth is declining and more people are need to replace us old gits as we retire. But if we’re being replaced by robots, who’s going need to hire?”
Exactly right. Economists, like generals, usually have their eyes fixed firmly on the past — and so usually find the future a surprise. I have said that slow or negative population growth is the key to success in the 21st century:
> But more generally, can people vote to prevent automation?
Don’t know. But voting is a market, politicians want jobs, and they’ll promise things they think will get them market share. In the UK millenials don’t get the attention they want because they’ve got a history of not voting, while the over 50’s do, because there’s a lot of them and they vote. I guess we’ll see.
They may not be able to prevent it, but I think it may be possible to slow it with regulation. The EU have managed to pretty much kill GMO use or purchase with red tape. The other difference with the past is that the very people who would prime candidates for much of the benefits of automation are probably also the targets to be automated.
How many AI induced disasters would it take before there were calls for more regulation or an outright ban or the equivalent of having someone with a red flag marching in front of the AI to prevent it running amok and killing civilians or just an AI speed limit.
I agree about the population growth thing, it’s key in many areas if we are to continue to live well. IMO automation/AI is part of the solution to that conundrum.
There’s a paragraph from the second Terminator movie:
“Watching John with the machine, it was suddenly so clear. The Terminator would never stop. It would never leave him. It would never hurt him, never shout at him, or get drunk and hit him, or say it was too busy to spend time with him. It would always be there. And it would die to protect him. Of all the would-be fathers who came and went over the years, this thing, this machine was the only one that measured up. In an insane world, it was the sanest choice.”
If the above is possible it’s going to be impossible to fight it, even if we should. Perhaps the biggest question we should be asking is, can we really build something *better* than ourselves? Can we even agree on what ‘better’ means.
Finally, will religious groups demand or start the construction of ‘true believer’ AI?
Steve,
“But voting is a market, politicians want jobs, and they’ll promise things they think will get them market share.”
If it’s that certain, why the almost total failure of voters to stop — or even slow — previous waves of automation?
“The EU have managed to pretty much kill GMO use or purchase with red tape.”
I don’t believe use of genetically modified organisms is “automation” in any meaningful sense. And time will tell how long the UK regs last, and if other nations follow the UK’s lead.
“How many AI induced disasters would it take before there were calls for more regulation or an outright ban ”
I agree that is a likely future. But probably a distant one. AI is being implemented at an accelerating speed. But at an incremental pace in applications that make a large disaster unlikely. Reading sonigrams, Siri, driving cars, writing personal trust documents, writing newspaper articles, etc. Slow means that problems are detected early, at small scale. And the nature of the aps mean the problems are small.
Eventually both these things will change, with rapid — even careless — implementation as controllers of large socially critical systems. But that is probably decades in the future — after much of the employment effect of automation has hit. We tend to overestimate short-term change and underestimate long-term change.
I’d recommend just going with general histories, they’re interesting enough on their own. Though the period that I find to be the most interesting is the one after the Han. The three centuries between the Han and Sui dynasties were extremely bloody and conflict ridden, but also very dynamic. You had physical inventions like Printing, Gunpowder, and Paper, the huge cultural shift brought about by the introduction of Buddhism, and the rise of many peoples on China’s periphery. In particular, this era saw the first (and most successful) invasion of China by the Mongols, Korea’s Golden age, and emergence of Japan as we know it.
While this post certainly provides a valid warning, it doesn’t provide a complete picture. Fabius Maximus, you often talk about developing technologies in a vacumn, considering their impact only in the context of how society is right now. You seem to not take into context that other developing technologies are going to interact with each other. The future where we have Robots everywhere is going to be same future where we get a 12-inch extension of our Penises for a penny, and those two aspects are going to interact in unpredictable ways. I’m not saying that things are going hunky-dory, I’m saying that I’m not going to assume things based on human nature when I’m not sure if human nature is going to be recognizable 3 decades down the road.
FDW,
“While this post certainly provides a valid warning, it doesn’t provide a complete picture.”
This is a common rebuttal, and one of the most daft. Please show me what you have written in a thousand words or so that “provides a complete picture” of a complex phenomenon. Books — long scholarly books — are written about these things, and reviewers usually point out important aspect the authors didn’t discuss. This doesn’t not claim to be a report on the Cosmic All, just giving one historical example that illustrates one dynamic of automation.
Well, one thing that might provide a more complete picture is asking if mass Biological augmentation might end up outperforming AI. We’ve only just begun to be able to tap into our biological potential, and there’s a distinct possibility that we could hit the wall in terms of what we can do in terms of silicon.
However, part of my post was an emotional reaction. One thing that I’ve been seeing a lot on this site over many months of lurking is a distinct worry that the US is going to go the way of the Roman Republic. I’ve been thinking a lot about the future myself, and I’ve come to read the Tea Leaves differently. I think the US is on the path of the Han Dynasty. I see emerging trends in technology, culture and the economy as weakening the powerbase of the global Elite, not strengthening it. And I think the issues will start to come to a boil relatively soon.
FDW,
“I think the US is on the path of the Han Dynasty.”
Thanks for that insight! I know little of China history, but will look into it. Any suggestions where to look for more about this analogy?
“one thing that might provide a more complete picture is asking if mass Biological augmentation might end up outperforming AI.”
“if mass Biological augmentation might end up outperforming AI.”
That’s certainly possible. I doubt it will be that simple. Biological augmentation might be people advantages in some domains. Human-computer fusions could be big! And AIs will almost certainly have advantages in some areas. Complex futures lie ahead. For a brief sketch about these things see Potentially horrific effects of drugs and machines making people smarter & stronger.
“there’s a distinct possibility that we could hit the wall in terms of what we can do in terms of silicon.”
I doubt that very much, in terms of hardware, software, or computer systems.
In general, I feel analogizing to China is a better idea, because China has a much more similar climate and landscape to that of the US. Not a perfect fit, but a better one than that of the Roman Republic.
I think some of there are some interesting parallels between the Han and the US. Both government brought back into fashion their style of government (Han with Meritocracy, the US with Democracy), both followed a dysfunction government that emerged victorious from lengthy wars (Han with Qin and the Warring States, US with the First Republic and the many colonial wars of the 17th and 18th centuries), both conquered vast lands (Han with Southern China, Central Asia, Korea, and Vietnam, US with, well you know), and both have lent their name to being used as an ethnonym.
FDW,
Thanks for this valuable and interesting material! I, probably like most Americans, know absurdly little about Chinese history. I’ll bet we could learn a lot from its study.
Any recommendations about books or articles exploring this similarity?
There is an important difference between people and horses. Horses don’t buy things.
The industrial revolution displaced many workers. In addition to developing new jobs, we eventually shortened the work week, enhanced labor laws and added strong unions to the mix. It was necessary to make the labor we needed worth enough to buy what we produced—and that’s what we did.
Either we will inflate the value of labor until wages can support consumption, or ordinary people will receive significant income from some source other than wages—be that a basic income guarantee, a universal profit-sharing law, or some other invention.
Eventually we will do what we must do. Unfortunately, we’re likely to try everything else we can think of first.
Coises,
“Either we will inflate the value of labor until wages can support consumption, or ordinary people will receive significant income from some source other than wages—be that a basic income guarantee, a universal profit-sharing law, or some other invention.”
Yes, those are the obvious outcomes. But then the obvious outcomes from the previous waves of automation came only after long periods of pain — social stress and violence. Hence my conclusion.
“Unfortunately, we’re likely to try everything else we can think of first.”
I agree that is the way to be. But “nothing is written.” These cheap easy predictions have a high failure rate. Remember my previous post. Smart machines are an inflection point in history. Assuming that they will have not massive and unexpected impacts is imo quite daft. But nonetheless, that’s the consensus expert opinion. Bizarre.
If the solution is to shrink the population through reduced fertility (whether imposed by government, social pressure or personal preference) there will be a difficult transition as the population ages.
China:
The rationale for the abolition is summarized by former Wall Street Journal reporter Mei Fong: “The reason China is doing this right now is because they have too many men, too many old people, and too few young people. They have this huge crushing demographic crisis as a result of the one-child policy. And if people don’t start having more children, they’re going to have a vastly diminished workforce to support a huge aging population.” China’s ratio is about five working adults to one retiree; the huge retiree community must be supported, and that will dampen future growth, according to Fong.
https://en.wikipedia.org/wiki/One-child_policy
Japan:
Based on 2012 data from the National Institute of Population and Social Security Research, Japan’s population will keep declining by about one million people every year in the coming decades, which will leave Japan with a population of 42 million in 2110. More than 40% of the population is expected to be over age 65 in 2060.
https://en.wikipedia.org/wiki/Demography_of_Japan
Rovingbroker,
” there will be a difficult transition as the population ages.”
Perhaps, perhaps not. These wildly overconfident forecasts about the future have a high failure rate. As in the classic “America’s cities certainly become buried in horse manure.” We see only the doomsters’ forecasts because they make better clickbait than more pleasant forecasts (I’ve seen that in the low pageviews of posts about good news).
For details see Doomsters warned of End Times from overpopulation. Now *fewer* people are disastrous. The transition might be painful, as with most transitions in history. But it need not be extremely so.
Roving Broker,
Follow-up — tech will help with some aspects of the age wave.
“How robots could help bridge the elder-care gap” at The Conversation.
I don’t the “Silver Wave” as being as much of a problem as people think, but for a different reason. Anti-Aging technologies are moving very fast in terms of development, and it could very well mean we don’t need all those nursing robots.
And frankly, you really ought to be talking about biological augmentation a whole lot more. I haven’t heard you talk about the implications of CRISPR at all, even though it means that Genetic Engineering is going to be cheap and widely accessible. I feel you should also be taking a look at the concept of Morphological Freedom, because we’re a whole lot closer to it than many people seem to be aware.
FDW,
“Anti-Aging technologies are moving very fast in terms of development, and it could very well mean we don’t need all those nursing robots.”
It usually takes decades to bring drugs from early lab stage to widespread use. Most drugs don’t make it. Sometimes the drugs work but are too expensive for mass use (which would be politically problematic).
“you really ought to be talking about biological augmentation a whole lot more.”
Geopolitics is the study of what we should do today — based on what’s on the board now, or coming soon. Go too far out and the odds of useful predictions drops to zero. So there are few implications for current actions.
The thing about Anti-Aging is that it’s not just in the labs anymore, there are drugs on the market, like Metformin. And I think that, as Aubrey de Grey has said, once people become aware how easy Antiaging therapies are (There are only 7 ways in which the human body ages), there’s going to be huge social/political pressure to make them widely available ASAP.
And a lot of the techniques involved in rejuvenation can be applied to a lot of different things, opening the door to morphological freedom.
FDW,
Please, no
pseudoexaggeration of the science. I take Metformin, and am well aware of this story.(1) A summary about its possible effects: “Metformin – its potential anti-cancer and anti-aging effects“. Note the tentative nature of the conclusions.
“The generally accepted mechanism of metformin’s effect is stimulation of adenosine monophosphate (AMP)-activated protein kinase (AMPK). AMPK is directly activated by an increase in AMP:ATP ratio in metabolic stress conditions including hypoxia and glucose deprivation. Lately, many novel pathways, besides AMPK induction, have been revealed, which can explain some of metformin’s beneficial effects. It may help to identify new targets for treatment of diabetes and metabolic syndrome. Moreover, metformin is now attracting the attention of researchers in fields other than diabetes, as it has been shown to have anti-cancer, immunoregulatory and anti-aging effects. The aim of this review is to describe the potential anti-cancer and anti-aging properties of metformin and discuss the possible underlying mechanisms.”
(2) There have been studies tentatively showing effects, although they are outside the current FDA approval guidelines. Here is one of the trials exploring its effects: To determine if treatment with metformin (1700 mg/day) will restore the gene expression profile of older, glucose intolerant adults to that of young healthy subjects. I’ve heard of other studies in progress, but haven’t checked them out. It’s still a long road to documented effectiveness, let alone widespread use. Note the “Detailed Description” description section, similar to that of the above study.
“Metformin in Longevity Study (MILES).”
FDW,
Follow-up — Most of the anti-aging drugs are in the lab. But, as you note, Metformin is approved for human use. It’s “in the lab” only in the narrow sense of phase IV clinical trials to determine effectiveness.
FDW,
I corrected my post to read “Please, no exaggeration of the science. “Pseudo-science” was wrong (too hasty writing), for which I apologize.
FDW,
I’m doing serial typos today.
I corrected my post to read “Please, no exaggeration of the science. “Pseudo-science” was wrong (too hasty writing), for which I apologize.