In 1924 fascist leader Benito Mussolini won the Italian general elections for the first time. Just like many of their compatriots Italian couple Carlo and Giovanna Zannoni flew their country to escape the new horrendous regime. They settled in Moutiers, a mining town in French Lorraine. Thanks to its coal and iron ore mines Lorraine attracted workers from all over Europe. Carlo spent the rest of his life as an iron ore mineworker. In 1945 he was joined by his son Amilcar. Amilcar worked 14 years in the mine to finally discover the passion of his life, iron sculpture. He became a famous artist and made a living out of his talent. His work was deeply influenced by his experience in the depths of the Earth. One sculpture in particular fascinates me. It is titled "The Man Nibbled by the Machine." This is an astounding piece in which a man is devoured by a complex machinery. No flesh is left on his right leg. His left chest is wide open. So is the left part of his head. The heart, the left lung, the left brain have disappeared. They have been replaced by gears, spinning wheels, crossed axes etc... The man's body and brain are eaten by the machine. Amilcar's sculpture (which predates Terminator!) is a truly powerful metaphor: the vanishing flesh and organs suggest that man is losing his race against the machines. The nibbled body conveys a grim view of the future of man at work.
The fear of machines is of course not new. Man has always had a complex relationship to machines. He invents them to be more productive. At the same time though machines destroy jobs. The same question is asked again and gain: Are new jobs created faster and in greater numbers than suppressed ones? From Leonardo da Vinci to Harvard University Schumpeter, from the Luddites to MIT Brynjolfsson and McAfee the same anguish is expressed. What is the future of man in a world where machines do the job? This anguish is stronger than ever today. Arms and legs are no longer alone at stake: as anticipated in Zannoni's sculpture, the brain is now the target of a new breed of machines powered by artificial intelligence. Machines have never been so tangent to man. More and more people worry that this time is different. They fear that the disruption brought by digital automation will not be compensated by long run benefits as it used to be the case in the past. This is what MIT economists Brynjolfsson and McAfee call the great decoupling. Indeed, over the last thirty years,
"Economic abundance, as exemplified by GDP and productivity, has remained on an upward trajectory, but the income and job prospects for typical workers have faltered."
In 2013, after adjusting for inflation, US workers were earning roughly 15% less than in 1973. During the same time period productivity doubled, and real estate prices, healthcare costs and education tuitions rose sharply.
The angst is predicated on the fact that digital technology contributes to the replacement of more types of human labor than ever before. Radiologists are for instance at risk of losing their jobs to image analysis algorithms. Any job that rely heavily on data is under siege as machines get better and better at handling massive amounts of data. To understand how we ended up there let us look at the previous episodes since the 18th century.
The first and second industrial revolutions from the late 18th century to the early 20th century marked a turning point. Machines blossomed everywhere: from the spinning jenny, the steam engine, the milling machine, the cotton gin, the blast furnace to the internal combustion engine new manufacturing processes were put in place. New sources of energy became available such as oil and electricity. Hand production methods were replaced by machines that took over the dirty jobs. Mass production and assembly lines allowed growth to kick in and, despite the revolts of Luddites and the like, transformed agrarian societies into industrial ones. Even though the social conditions were more often than not terrible more jobs were created than lost, jobs that did not exist before. It is however fair to say that the newly minted jobs became awfully dull. Henry Ford could not be more explicit when he said:
"Why is it every time I ask for a pair of hands, they come with a brain attached?"
In the movie Modern Times Little Tramp, the character that made Charlie Chaplin famous, struggles to keep up with the infernal pace of the assembly line but finally suffers from a nervous breakdown. Thanks to machines perspiration did decrease but at the cost of pathetic and unhealthy dullness.
In the 20th century automated machines were deployed to take over dull jobs. The progress made by robotics, a branch of mechanical engineering, electrical engineering and computer science, contributed to the development of machines than can operate autonomously. In 1961, the first industrial robot, the Unimate, was installed in an American factory. Robots are now everywhere to be seen in assembly lines. They are faster and more accurate than human beings. For instance in automobile factories robots can execute fifty to hundred welding spots per second. Tesla's assembly line in Fremont, California is powered by close to 200 robots. Moreover robots do not perspire. They can work long hours without having to take a coffee break or a nap. Robotic automation helped increase productivity and wealth per capita. The wave lifted (almost) all boats and gave birth to the middle class.
The 21st century does not seem to hold the same promises. Its first years have been calamitous to say the least. Observers and commentators start wondering whether Nobel Prize Wassily Leontief was not right when he predicted that
"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."
Now that dirty, dangerous and dull jobs have been swallowed by machines and robots, the technologists' ambition is to conquer the brain. The goal is to disseminate androids that can not only walk but also think and make decisions. It is an old dream of course. Jacques Vaucanson, a French inventor born in Grenoble in 1709, cherished the hope of an artificial man. He built famous androids that could play the flute and the tambourine. They played the instruments so well that Vaucanson's inventions were welcomed either with fear or skepticism. Vaucanson could not pursue his ventures. He was asked by the authorities to work on the turnaround of the French ailing silk industry.
Vaucanson's dream is well and alive though. Google Director of Engineering Ray Kurzweil is crystal clear about his goal, and he is sure that it will be reached:
"I have been consistent in predicting that AIs will match human intelligence in all of the ways in which humans are now superior by 2029. They will then be able to apply their enormous speed and scale and total recall to all of human knowledge. I believe that recent advances should give us a lot of confidence that we will meet or beat that goal."
Because of the exponential factors at work (computing power, storage, speed, Kurzweil's law of accelerating returns etc...) there is no doubt according to Kurzweil that we will be able to reverse engineer the brain and hack its algorithmic component, in other words its mind part1. We will not only replicate the brain/mind but outperform it. We will replicate it thanks to the advances in artificial intelligence. The algorithmic part of the brain will be made downloadable. No flesh allowed! We will outperform the brain/mind because computers and algorithms do not forget. They can store far more data than an individual brain can. They do not make mistakes. They are unbiased and above all they will be able to share information via the cloud, a sort of powerful collective algorithmic mind. Kurzweil is convinced that we will then reach a singularity point, a point in time when there will be no difference between man and machine. Our intelligence will have become nonbiological as if Amilcar Zannoni's worst nightmare of a man nibbled by machines had come true. If one follows Kurzweil's trans-humanistic argument it is not jobs that will disappear but man as we know it. According to Kurzweil's friend and MIT Professor Marvin Minsky there is not much to worry about since after all the brain is "a meat machine" and the body "a bloody mess of organic matter." The bad news is that we will not have any jobs anymore but the good news is that we will not care as we will be something or someone else!
In a sense we have come full circle. Henry Ford wanted a body only, no brain attached. Ray Kurzweil wants the opposite : mind only, no body (and brain) attached! While both perspectives look scary, their simplistic views of man complexity is rather good news. Indeed, we do intuitively feel that a body without a brain and a mind does not make much sense. A brain and a mind without a body does not make much sense either. This split view of mind, brain and body is what University of Southern California Professor of Neurobiology Antonio R. Damasio calls Descartes' Error. Descartes, a 17th century French philosopher, wrote the famous statement "Cogito ergo sum." It says it all: to be one (only) needs to think. Descartes posited the dualism between the mind and the body. In other words the body is not of great value, and although not made of matter the mind is what matters. Kurzweil would wholeheartedly approve. Damasio agrees to disagree. Body, brain and mind cannot be considered separately. For instance, when a threatening dog barks at us, our pulse quickens, our blood pressure rises, and adrenaline is released. The complex reactions the body triggers in response to the barking dog are called emotions in neurobiology. Then comes the expression of feelings as our brain notices the changes in our body conditions. Following the emotions in front of the barking dog we experience the feeling of fear, a mental outcome. Feelings (at the brain level) are formed by emotions (at the body level). Damasio summarizes it eloquently: "Mental activity, from its simplest aspects to its most sublime, requires both brain and body proper." These body/brain/mind processes have been shaped and perfected by millions of years of evolution. Mother Nature is indeed a great tinkerer. She takes the time, makes mistakes and does not set up any target date to reach goals that, by the way, she does not have in the first place, a behavior that is the exact opposite of Kurzweil's. Kurzweil has a goal, a target date (he loves making predictions) and one huge bet on artificial intelligence. Mother Nature places multiple bets without having any specific goal (she plays the unknown instead of the known). This is what makes her so good at what she does. Kurzweil and the like piggyback on the known (at least what they think they know), and reduce it to a computer power problem. The rest, the unknown is just something that has not been computed yet but will sooner or later be. Brute force instead of soft serendipity! They forget that all what is computed and computable does not necessarily matter, and, above all, that all what matters does not always have to be computed. They overlook the pervasiveness of randomness, and how nature (including man) has dealt with it. This is a rich fabric of mistakes, trial, errors and of course successes in the great Augustinian tradition: "Si enim fallor sum". All humans, taken together as they are, flesh, brain and bones, are worth a lot more than any central cloud made of our supposedly perspiration-proof downloaded minds. To put in very simple terms, I perspire therefore I am! To put it in Hungarian philosopher Michael Polanyi's words:
"We can know more than we can tell... The skill of a driver cannot be replaced by a thorough schooling in the theory of the motorcar; the knowledge I have of my own body differs altogether from the knowledge of its physiology; and the rules of rhyming and prosody do not tell me what a poem told me."
Good luck Mr Robot!
Indeed, robots do not perspire, nor do they consume by the way. This is their weakness. This is a great piece of news. Not so long ago perspiration is what used to make our jobs fragile: inspired in California, perspired in China! Perspiration combined with inspiration is what may save our jobs in the so-called race against the robots. But if we take it as a race, we will lose and Kurzweil may have it right. The terms of the race as they are currently spelled are indeed not in our favor. The race is not inevitable however. As French writer Georges Bernanos once wrote
"the danger does not lie in the machines, if it did we would be pushed into this silly dream of having to destroy them... The danger does not lie in the multiplication of machines either. It is in the growing number of men who, since they were born, are only used to what machines can deliver."
This is indeed a danger especially if we leave the machines and the robots in the wrong hands who happen to capture the bulk of productivity gains, and all the optionality (big) data give access to. Moreover we shall not forget that a whole generation has only known and used the Internet by swapping its strong currency, its data, against the weakest one can find, free! This would not really matter if digital capitalism were not a winner-take-all game. It is more than ever shaped by the rent seeking behaviors of robber robot lords. For instance, it is no accident that Google heavily invests in robotics. Robots need to be trained to be efficient at what they are supposed to do. The training takes place through data. It takes for example zillions of images to improve a robot's vision. These data do not come out of the blue. They come from us via the Faustian "free against data" swap. We are even kind enough to tag the data and upload them to the cloud. Life has never been easier for the robber robot lords!
It does not have to be so though. Digital capital is abundant. It has low marginal costs. Digital capital being more and more used in production it stands to reason that capital owners should see their returns decrease, unless of course they seek to capture rent positions and we let them do so. The race is not a race against the robots. It is a race against the robber robot lords, against the new Carnegie, Vanderbilt and the like. In his book Antifragile, Nassim Taleb praises Prometheus and mocks Epimetheus. According to Taleb, Prometheus who stole the fire from the gods is antifragile, long disorder. Prometheus enjoys upside while having no downside. The more disorder there is, the better his position becomes. Epimetheus who accepted Pandora's jar is short disorder. Only bad things can happen, and did happen after he opened the jar. I fear that we are more Epimetheus than Prometheus. Indeed, the jar we have been offered is in all senses of the word (especially in the Internet sense) a cookie jar. We have accepted it without knowing what uses would be made out of our data. The robber robot lords have stolen the fire of our data, namely the many valuable options these data carry that can and will be monetized. We have opened the jar, and, as things stand, we have no way of knowing what the robber robot lords (will) do with our data. Again, this issue is no longer a minor one: it concerns billions of people on Earth whose data are siphoned by the robber robot lords.
It is time to throw the cookie jar away and regain the control of our strong (data) currency. I do not think that robots are our best enemies for we do not for instance regret having machines do the job in mines instead of children. By the same token I do not think we should regret having robots do the hazardous, dirty, dull and time intensive jobs, be they blue collar or white collar. Automation frees up time. It liberates our mind/brain/body (which is far more than an algorithm) so that we can "perinspire" on important issues. If the decoupling between economic growth and job creation is confirmed, we cannot satisfy ourselves of Brynjolfsson and McAfee's simplistic dirty secret of economics:
"technology progress does grow the economy and create wealth, but there is no economic law that says everyone will benefit."
The key issue is not that there is a decoupling. The key issue is to whom this decoupling benefits. Sadly enough, Brynjolfsson and McAfee are right, economics has become a truly dismal science of growth full of dirty secrets. It has not always been the case though. Economics used to be called political economy. The great minds behind it were concerned with the distribution of wealth, and not so much about its maximization. That is what mattered to them, how humans can live and strive in a balanced society. They knew for a fact that, even if it helps, a growing cake is not a remedy to inequalities, especially when the cake is cooked in a winner-take-all kitchen. They kept a close eye on the distribution of wealth for they knew that an unequal and unbalanced society is unsustainable. Today's society is at risk of growing inequalities. It may become more unequal in a digital gilded age that is predicated on increasing returns to scale and winner-take-all effects. It is unbalanced and may become more unbalanced in a digital gilded age during which we entered, on an unprecedented scale, into Faustian "free against data" swaps with robber robot lords. Data are the gems of the digital gilded age, and there is no economics dirty secret saying these gems should go for free and benefit the predatory few while being stolen from the gullible many. We are not e-monkeys: we do not accept e-peanuts!
Again the race is not against the machines and the robots. The race is against those who own them to avoid that they end up owning us.
The race is about the (re)distribution of wealth in what we hope will be a sustainable digital gilded age.