IT’s not the future

The Second Machine Age sacrifices sense at the altar of technology

In October 2011, Erik Brynjolfsson and his colleague Andrew McAfee published Race Against the Machine: How the Digital Revolution is Accelerating Innovation, Driving Productivity, and Irreversibly Transforming Employment and the Economy. Now the two authors – henceforth B&M – have published The Second Machine Age: Work, Progress, and Prosperity in a Time of Brilliant Technologies. Strangely enough for books on innovation, the second is a lengthy recycling of the first. Still, many commentators have praised its basic message – that IT has become astoundingly powerful, but brings labour market effects that are worrying (1).

IT as a ridiculously potent force

B&M are quick to deploy the kind of hyperbole so beloved by America’s business books, and in particular by books praising IT. By pages seven and eight, computers are doing for mental power what steam engines did to physical power, and we are told that mental power is ‘at least’ as important for progress as the physical sort.

These points are just asserted, not argued. Instead, we next learn that the US Bureau of Economic Analysis first identified IT as a category of business investment in 1958 – this was also, though B&M don’t mention it, the year in which the integrated circuit was first demonstrated. We also learn that Time magazine hailed the PC as ‘Machine of the Year’ in 1982. Yet rather conveniently, the authors hold that computers have only now put the world at ‘an inflection point’ – a second machine age after the Industrial Revolution. Progress was gradual; now it’s sudden. IT is the reason.

Like too many other breathless American commentators, B&M understand by the term ‘technology’ just IT, and IT alone. And, consistent with this reductionist philosophy, the authors aver that IT simply ‘races ahead’. They note that in some countries mobile telephony is more widespread than electricity or clean water. It does not occur to them that, in India for example, failures of innovation in electricity supply and drinking water often prevent millions of people from being able actually to use or even afford mobile phones.

The second machine age, it is here vouchsafed, will increase both the variety and the volume of our consumption. For with IT omnipotent, B&M believe that abundance is the new norm with capitalism, not scarcity. Also, those with the right skills can use IT to ‘create and capture value’. However, it’s a bad time to possess only ordinaryskills. Why? Because IT brings not only consumer ‘bounty’, but also ‘spread’ – society’s familiar differences of wealth, income and social mobility. Everywhere, B&M present IT as an unstoppable force, not just for consumption (good), but also for social differentiation (bad). In this respect, they think they can get away with fusing the breezy consumerism of the dotcom bubble (1997-2001) with the more doom-laden concerns that today attend inequality.

B&M say that IT excels at following rules, but accept that it is poor at recognising patterns. They are also right to say that IT is still set to achieve new and Great Things. In both books, however, the authors use the same three technological developments as sledgehammers to drive this IT-poised-for-Great-Things point home: Google driverless cars; GeoFluent software, which, in the customer-relationship-management systems run by service companies, translates online chat into different languages; and IBM’s Watson supercomputer, which has won the game show Jeopardy!.

Like too many other breathless American commentators, Brynjolfsson and McAfee understand by the term ‘technology’ just IT, and IT alone

These are advances of a sort. However, regulatory wrangles may slow the diffusion of driverless cars. Translating online chat doesn’t easily resolve customer service issues. Game shows hardly demonstrate what B&M call ‘intelligence’. Characteristically, the authors overpraise IT. In the process, they fail fully to convince.

Robots are not ‘exploding’

So what, in the opinion of B&M, has changed? Rodney Brooks’s company Rethink Robotics has made a wheeled, safe, physically trainable robot, Baxter, which begins to overcome a paradox attributed to roboticist Hans Moravec: the oddity that computers find higher-level reasoning easier than, say, straightening the magazines on a coffee table, or folding a towel. Baxter, B&M hint, can do that kind of straightening and folding. Thus ‘the volume and variety of robots in use at companies is expanding rapidly’, the authors say; or, as they quote a vice president from Texas Instruments from early 2012, the robotics market is ‘on the cusp of exploding’.

Google’s December 2013 purchase of Boston Dynamics, the roboticists who make BigDog, might reinforce this view of robots enjoying explosive growth. Yet statistics on robots, which B&M cannot anywhere be bothered to present, tell a different story. The world’s total supply of industrial robots did rise by 12 per cent, to 179,000 units, in 2013. And, like the rest of IT has done for years, robots do play a mediating role in redundancies today. Yet the worldwide tally of industrial robots forecast to be sold in 2014 – 200,000 – is actually fewer than the average number of new jobs created, monthly, in the still spluttering US economy:

Annual industrial robots supplied, thousands, world, 2003-2014

Source: International Federation of Robotics, June 2014

Monthly non-farm jobs created, thousands, US, January-June 2014

Source: Bureau of Labor Statistics

It is true that capitalism is developing new and different robots. Some robots now work alongside car workers, rather than replacing them; others are emerging as service robots. Yet robots are not coming very rapidly, even in China (2).

It’s clear that once B&M leave their fairytale world of IT-led consumer bounty for the gritty realities of the factory floor, they all too easily imagine things.

Is innovation really just about Moore’s law, digitisation, and combination?

Like many others, B&M come over all love-struck about the exponential function, and, in IT, its expression as Moore’s law. In a 1965 article, Gordon Moore, co-founder of Intel Corporation, famously noted that the number of transistors in a minimum-cost integrated circuit had been doubling every 12 months, and predicted that this rate of improvement would continue. Since then, Moore’s prediction has come true, provided that the period for doubling is taken to be 18 months.

To be fair, B&M rightly note that Moore’s ‘law’ isn’t a law of physics. Yet in a kooky nod to the even-kookier futurologist Raymond Kurzweil (The Age of Spiritual Machines: When Computers Exceed Human Intelligence, 2000), the authors invite us to accept a tale from sixth-century Indian mythology. Imagine a chessboard whose squares are successively given one, two, four, eight grains of rice and so on. Well, 32 ‘doublings’ in ‘digital gear’ since, er, 1958 takes us to 2006. So in 2014, B&M insist, IT is now well into ‘the second half’ of its own developmental chessboard – a place which makes science fiction into reality, and which leaves intuition and experience behind. IT, in short, is now at a moment when ‘things get weird’, and ‘most of us have trouble keeping up’.

So although B&M say Moore’s law isn’t a law of nature, in fact they follow the geometrical progressions that once entranced Thomas Malthus. They turn accelerating computer power into, yes, a law of nature. Indeed for B&M, the ‘many dimensions’ of Moore’s law apply not just to chips, but also to gigabytes-per-dollar efficiencies in the hard drives of computers, to the speeds achieved by broadband telecommunications, and to the speeds and energy efficiencies attained by supercomputers. For B&M, Moore’s law ensures that Siri speech recognition on Apple’s iPhone 4, like the ‘ever-cheaper sensors’ on Microsoft’s Kinect, Baxter robots and Google cars, will bring us to a time when ‘what’s come before is no longer a particularly reliable guide to what will happen next’. Awesome!

Thankfully, the authors’ prose is not quite as glib as that of, say, Eric Schmidt and Jared Cohen in The New Digital Age. And to their credit, B&M do have a glimmer that, during the Depression of the 1930s, people already speculated that robots would wage war, commit crimes and displace workers. However the authors entirely miss how exaggerated fears about robots are growing today. The collapse of investment in the US likewise gives them no pause: no, IT is going everywhere and conquering everything (3).

To justify their upbeat impressionism, B&M add, to their revered exponential function, ‘the digitisation of just about everything’. Turning information into noughts and ones, they say, ‘increases understanding’. Harking back to Carl Shapiro and Hal Varian’s 1998 book Information Rules: A Strategic Guide to the Network Economy, the authors insist that information is different. Anyone can share it; to reproduce, transmit or receive it costs nothing; the more people who use an IT system, the better it is (the famous ‘network effect’).

For B&M, these debatable peculiarities of IT explain everything. Yet if IT is so unique, the exponential nature of Moore’s law so magical, and digitisation so dynamising, why did what Shapiro and Varian described as the US ‘network economy’ collapse with the Lehman Brothers crisis of 2008?

If reducing technology to IT betrays B&M’s conservative premises, so too does the third characteristic of their New Age: innovation as combining things that already exist. This doctrine, which belongs not just to gurus such as ‘complexity scholar Brian Arthur’ and ‘economist Paul Romer’, but also to many others (4), is laughable. Yes, when Einstein worked out e=mc2, he had no idea that his science would eventually assist the Manhattan Project and nuclear energy. But whatever else is said about Einstein, his theory of special relativity, first promulgated in 1905, was notjust a simple combination of what had gone before.

For B&M, by contrast, systems such as the web and Facebook merely combine the old. In the case of Facebook, there is some truth in this; but that fact doesn’t stop B&M from cheering. For them, the only problem with innovation is to use our old friend, the wisdom of crowds, also known as crowd sourcing. Yes, get enough experts and non-experts into a peer group so as to filter all the new ideas springing up everywhere, weld these filtered ideas into new companies, and go on to build more of a ‘peer economy’.

It escapes B&M that genuine innovation depends on new knowledge, which itself often depends on radically new scientific discoveries, new experiments and new prototypes – all of which combine what has gone before to only a limited extent. Do we today see sustained, transformative, mutually reinforcing innovations in transit between manufacturing, construction, energy, transport, pharmaceuticals, retailing and other services? Hardly. Perhaps the reason is that capitalism relies too much on what has gone before. And perhaps this is why B&M are so very keen to uphold combining the old as the key to making the new.

Semiconductors don’t drive inequality

Since 2005, America has suffered a slowdown in IT-based productivity increases. However, the authors believe that this relapse merely reflects a temporary pessimism that set in after the Lehman upset. When the US introduced nationwide electricity, they note, productivity gains took time to emerge. Still, in productivity terms, the three decades that followed the Depression ‘proved to be the best ones of the twentieth century’.

Could the Second World War, the destruction of capital and the political quelling of the international working class have had anything to do with the boom that succeeded the Depression? For B&M, clearly not. For them, even the fate of the mid-century US economy was determined by the flow of electrons – not by the electrons in chips, it’s true, but by those electrons bound up with the alternating current in America’s electricity grid.

Noting that for every dollar invested in IT hardware today, firms have to invest another nine in software, training and revamping business processes, B&M obliquely acknowledge that the intensified exploitation of labour – ‘business process redesign’ and also ‘incentives systems’ – is a factor in economic growth, not just technology. They also concede that vital to production today is ‘human capital’, even more than three other ‘intangible assets’: intellectual property, organisational capital and, yes, ‘user-generated capital on digital-content sites’. But such is the authors’ technological determinism that it is IT, and pretty much only IT, that always, for them, turns out to be the prime mover today.

So society’s shift from analog to digital technologies has, in the opinion of B&M, contributed not just to progress, but also to ‘an income distribution that is far more spread out than before’. The authors dimly suggest that this shift is one that has occurred in America’s ‘current economic system’ – one which for the authors is, inevitably, unambiguously a ‘free-market’ system. However, for B&M, it is relatively small IT firms such as Facebook’s Instagram, and the seven superstar billionaires it has created, that now make for income spread. Capitalist social relations in general, the criterion of profitability, the state-backed financialisation of all aspects of economic life: these are never even intuited as possible sources of inequality.

Like their compatriot Tyler Cowen, B&M assume that the ‘main driver’ of growing inequality is IT (5). Yet how, stripped of social relations, can semiconductors lead to unemployment and declining rates of social mobility, any more than they can create ‘intelligence’, or, indeed, ‘digital democracy’? It is a definite structure of human political economy, not silicon, which leads to human suffering.

By this time B&M feel the need to propose solutions. Chief among these is properly salaried education – though this, for B&M, should be about skills more than knowledge. And here again IT rides to the rescue. The automatic grading, by IT, of student essays would improve the quality of education and lower its cost. Also, it’s time to dispense with the Victorian three Rs, and instead enrol with the IT-based Khan Academy and Massive Open Online Courses (MOOCs).

What are the authors’ further solutions? Match employees to employers through… LinkedIn. Tax negative externalities, such as pollution, with the aid of… IT-based meters. Improve America’s declining rate of new business formation among more-than-sole-proprietor enterprises. Relax regulations. Lighten the climate for would-be immigrants. Increase state funding and prizes for science. Tax the super-rich, cut payroll taxes, give the very poor extra state money if they go out to work, and follow Richard Nixon in guaranteeing a basic income for all. Oh, and America should start charging VAT. Profound innovations.

‘In a place like Silicon Valley or a research university like MIT’, write B&M, ‘the rapid pace of innovation is particularly easy to see. Start-ups flourish, mint new millionaires and billionaires, while research labs churn out astonishing new technologies.’

Gentlemen! Even though you are only on the east coast of the US, at the Massachusetts Institute of Technology, you are too close to IT. You have forgotten all you should have learned in the dotcom bubble of 1997-2000. Get out a bit more. If you do, you will see technologies that go beyond IT. Everywhere, in fact, you will see how today’s capitalist society and culture constantly thwart the advances that technology promises.

The Second Machine Age: Work, Progress, and Prosperity in a Time of Brilliant Technologies, by Erik Brynjolfsson and Andrew McAfee, is published by WW Norton and Company. (Buy this book from Amazon(UK).)

Footnotes and references

(1) In the usual incestuous style, fans of The Second Machine Age, who are quoted at length at the book’s front, include many of the all-too-familiar American entrepreneurs and technology buffs whom the book itself quotes. Also, the Financial Times has repeatedly endorsed the basic theses of The Second Machine Age. See for example two articles by Martin Wolf, chief economics commentator, here and here.

(2) B&M believe that ‘Terry Guo of Foxconn’ (a noted supplier to Apple) has been ‘aggressively installing hundreds of thousands of robots to replace an equivalent number of human workers’. But Foxconn’s CEO is Terry Gou, not Guo. And, far from being heavily roboticised, Foxconn was in 2012 the world’s tenth largest user of labour, boasting a workforce of 1.2million people. As of late June 2013, the company had installed just 20,000 ‘robot arms and robotic tools’. We can be sure that if Foxconn had been putting in not 20,000 robots, but the hundreds of thousands of robots that B&M imagine, it would not now be reputed to be hiring another 100,000 workers to make the Apple iPhone 6.

(3) Sloppily citing the Wall Street Journal, rather than the direct figures from the US Bureau of Economic Analysis, B&M report that US spending on capital equipment and software ‘soared’ by 26 per cent. They refer here to the two years running from the second quarter of 2009 to the second quarter of 2011 – only a blip in the long run decline of investment in the US. From Q2 2011 to Q2 2012 the increase US spending on capital equipment and software was not 26 per cent, but 4.7 per cent; from Q2 2012 to Q2 2013, it was 2.9 per cent. These more recent figures are impressive – but not in the way B&M would like!

(4) See for example Frans Johansson, The Medici Effect: What Elephants and Epidemics Can Teach Us About Innovation, Harvard Business School Press, 2004; Scott Berkun, The Myths of Innovation, O’Reilly UK, 2007.

(5) Bizarrely, B&M summon Karl Marx to support this point. They maintain that Marx ‘predicted’ that mechanisation would drive workers to a subsistence wage. Marx did no such thing, and the authors have clearly not read the pages of Capital to which they refer the reader, where the word ‘subsistence’ does not appear. The B&M reference is to pages 708-9 of Capital.