The Second Machine Age
Erik Brynjolfsson & Andrew McAfee
‘Every day,’ Brynjolfsson and McAfee write in a new preface to The Second Machine Age, ‘we come across examples of science fiction becoming reality.’
However, the two authors admit that advancing technological progress means losers as well as winners. There are still millions of people who have simply stopped looking for work, or are working in jobs way below their potential.
The pair see human desire and need for work as forming one of the big debates of the future.
A new machine age
The early part of the book rests on an analysis by anthropologist Ian Morris, who highlighted that for most of human history, technological progress was ‘achingly slow, almost invisible’.
Then, 200 years ago, the Industrial Revolution meant that for the first time, technology – not politics, religion or population – drove human progress.
Now, the authors say, we are in the second machine age, in which ‘computers and other digital advances are doing for mental power … what the steam engine and its descendants did for muscle power.’
In 1987, Nobel-winning economist Robert Solow said: ‘We see the computer age everywhere, except in the productivity statistics.’ In fact, Brynjolfsson and McAfee note, there always seems to be a substantial lag between the introduction of a technology and increasing productivity.
Despite the introduction of electricity to American factories in the 1890s, there was no labour productivity surge for another 20 years. Initially, factories simply replaced steam engines with electrical plants, and maintained existing processes. Not surprisingly, there wasn’t much increase in productivity.
In the 1980s, IT was still a small part of the economy, and not until the 1990s were there big productivity improvements thanks to its use.
Technology and inequality
Until 1973, a growing economy was like a tide raising all boats, with wages growing across the board. Since then, incomes in the US and most developed countries have jumped for 10-20% of the population with advanced skills and education, and fallen or stagnated for the rest.
While routine tasks are increasingly automated, leaving less-skilled people out of a job, work such as big data analytics and rapid product development have only increased the need for people with reasoning, creative or design skills.
It could be argued that such an economic environment is fair, because those who have the most positive impact on society are rewarded.
The authors do not buy this, for the simple reason that many people are not just losing out in relative terms, but seeing absolute falls in income at the same time that technology races forward. Housing, healthcare and college tuition are each significantly more costly than they used to be.
The usual remedy put forth for ‘technological unemployment’ is to provide a guaranteed income to ensure people still have enough money to keep the economy going, even if they’re not working.
Brynjolfsson and McAfee are opposed to this, as work provides many psychological benefits beyond earning a living.
Instead of a universal wage, the authors look to where humans can retain an edge over machines. ‘We’ve never seen a truly creative machine, or an entrepreneurial one, or an innovative one.’
It is the combination of human creativity and computer data-crunching that offers the most exciting advances in all areas.
That covers white-collar occupations; what about more physical work? Moravec’s Paradox is the surprising fact that computers can do high-level reasoning easily, but the simplest sensorimotor skills need massive computational power. For the foreseeable future, cleaners, cooks and hairdressers are secure.’
In 1930, Keynes imagined a time when the ‘economic problem’ was solved. In this society, people might only work 15 hours a week. Keynes failed to see that human wants are endless, creating unlimited markets for goods and services.
Oxford University researchers Frey and Osborne predict up to 47% of US jobs will be automated in the next 20 years by machine learning and robotics. This may turn out to be an exaggeration.
If you factor in the variety of tasks most jobs require, and the human interaction aspects, it turns out that only 9% of jobs are at risk of automation. The second machine age may be more human than we think.
Copyright © Tom Butler-Bowdon, 2017. The above is an abridged extract from 50 Economics Classics by Tom Butler-Bowdon, published by Nicholas Brealey Publishing.