The Second Machine Age, Work, Progress, and Prosperity in a Time of Brilliant Technologies;

Erik Brynjolfsson and Andrew McAfee; 2014

Review by Graham Mulligan

The trajectory of human history has changed. Expressed as a graph it has gone along a slow but improving line that travelled slightly upwards until the Industrial Revolution or the First Machine Age. The graph then starts to accelerate upward as the physical environment becomes altered by human control of physical energy. The authors of this book see a new ‘inflection point’ where the curve bends upward more rapidly, exponentially driven by the increasing powers of computing. This has occurred in less than a single lifetime. This is the Second Machine Age where robots and artificial intelligence and wonderful new stuff happens.

Brynjolfsson and McAfee develop the argument that three forces are combining to create the inflection point they describe. The power of exponential improvement, commonly known as Moore’s Law ,makes change almost unimaginable. The invention of chess, as retold by Ray Kurzweil in The Age of Spiritual Machines, serves as illustration. Imagine one grain of rice on one square, two grains on the next square, four on the next, eight on the next, until all the squares on the chessboard are filled by doubling the previous amount. The result is a very large number. Moore’s Law doubles the computing power of machines every 18 months.

The second force Brynjolfsson and McAfee credit is digitization. Turning information and media into bits and bytes has a unique feature that powers it. Digital information can be reproduced without loss and for negligible cost. Applications are built up from layers of such powerful systems leading to the creation of wonderful new tools and entertainments. One of my personal favourites is language-translation tools like Google Translate but many more as well that live as apps on my smart phone. While studying Mandarin I discovered an app that uses the camera on my smart phone to send an image of a Traditional or Simplified character to a website for immediate decoding.

The third force in Brynjolfsson and McAfee’s inflection trilogy is innovation. The upward slope of the line in the progress graph from the Industrial Revolution onward, referred to at the beginning of this article, was driven by innovations based on electricity, the internal combustion engine and indoor plumbing. However, some economists and cultural observers see evidence that innovation has slowed since about the 1970’s. The low-hanging fruits have been scooped. Apple’s App Store is not adding a lot of really exciting new tools or entertainments and the iPhone comes out each iteration just tweaked from the previous.  But Brynjolfsson and McAfee don’t see the slow down this way.

Recombinant innovation or the mix and remix of ideas have always been part of humankind’s search for ‘a better way’ to do things and in the inflection period the mix and remix is amplified by the two other forces. After citing several examples, Brynjolfsson and McAfee state optimistically “Progress doesn’t run out, it accumulates. And the digital world doesn’t respect any boundaries.” I’m not so sure about this. The idea of limits to growth has been put forward in the conceptualization of Planetary Boundaries which could curb at least the idea that infinite progress comes from growth. Other ideas like Sustainability and the Steady State Economy seem to suggest that boundaries do exist. Growth or Progress before the take-off two hundred years ago is often attributed to these four forces: “By finally discovering the four main prerequisites for sustained economic growth—property rights, scientific rationalism, capital markets, and fast and efficient communication and transportation—humanity was jarred awake from this two hundred thousand year slumber.” ( Growth and Progress seem to be corollary terms.

This book is about a theory of how the economy will evolve. Brynjolfsson and McAfee describe it as two choices, the ‘innovation as fruit’ view or the ‘innovation as building block’ view of the world. The ‘innovation as fruit’ view contends that all the low-hanging fruit has been picked and so growth by innovation has slowed. The only constraint to the ‘innovation as building block’ world they see is the question of value or which eggs do we put in our basket when there are so many eggs to choose from.

All of this, the exponential improvement, the easy and cheap replication through digitization, the innovation upon innovation, results in a more abundant world or what the authors call ‘the Bounty’. Like a Yin/Yang configuration, the authors also describe the alarming divergence of wealth that is so pervasive throughout the world today. They call this ‘the Spread’ and they ask ‘which is bigger?’. In Chapter Eleven the authors elaborate on the Bounty which is more than cheap consumer goods although one could argue that more variety or choice and different levels of quality are still ultimately just more consumer goods. But lets stay with the argument that ‘rising tides lift all boats’ and who cares if some people get really rich as long as the bottom get something. Is this really a ‘happy phenomenon’ as the authors call it? They do acknowledge the danger of unequal access to the Bounty and recall the history of one classic technological disruption in the early Industrial Revolutiion known as the “Luddite Fallacy”. Technology did put some people out of work but eventually created many more jobs.

Economic commenter, John Mauldin, sees the implications of rapid change this way:

“We are rapidly entering the Age of Transformation, a period in which change will continue to accelerate until it comes at us blazingly fast. And I’m not talking about just the introduction of new technologies; employment and job creation are also changing extraordinarily quickly.

And any reasonable analysis suggests that in the future the rate at which jobs are being lost to new technologies is only going to double and triple. This is one of the central problems facing society today, not just in the US but all across the developed world.

But hold on a moment, there might be a solution, says Mr. Mauldin. “So now I am here to tell you that technology is not the problem. Technology is the solution. Well, actually I agree it’s the problem if it’s your job. But the solution is to figure out how to get in front of the technology curve or figure out who is in front of it and get involved with them.”

In another post on his Outside the Box website, Mauldin writes:

“There is an interesting historical precedent for our situation, an era during which the technological firmament shifted just as abruptly as it is here and now. In the United Kingdom in the year 1800, the textile industry dominated economic life, particularly in Northern England and Scotland. Cotton-spinners, weavers (mostly of stockings), and croppers (who trimmed large sheets of woven wool) worked from home, were well compensated, and enjoyed ample leisure time.”

And this: (Originally published in Smithsonian magazine, January 2017)

“At heart, the fight was not really about technology. The Luddites were happy to use machinery—indeed, weavers had used smaller frames for decades. What galled them was the new logic of industrial capitalism, where the productivity gains from new technology enriched only the machines’ owners and weren’t shared with the workers.”

But he doesn’t see Bryn’s book this way. He says:

“Erik Brynjolfsson is less pessimistic. An MIT economist who co-authored The Second Machine Age, he thinks automation won’t necessarily be so bad. The Luddites thought machines destroyed jobs, but they were only half right: They can also, eventually, create new ones. “A lot of skilled artisans did lose their jobs,” Brynjolfsson says, but several decades later demand for labor rose as new job categories emerged, like office work. “Average wages have been increasing for the past 200 years,” he notes. “The machines were creating wealth!”

Here’s another version of this idea (from Singularity University’s Lisa K Solomon):

“Nearly two decades ago, military planners coined an acronym to capture the nature of an increasingly unpredictable and dynamic world. They called it VUCA—an environment of nonstop volatility, uncertainty, complexity and ambiguity.”

Furthermore, says Solomon,

“Technology disruption is quickly outpacing existing regulations, laws, and societal norms. There are already on-going tax and labor feuds between industry disruptors such as Airbnb and Uber and the communities they serve.

But those legal battles pale in comparison to the ethical battles we may soon face when workers in large industries such as food or transportation are replaced by autonomous systems. And we’ve hardly begun to explore the implications of a future in which genetic modifications have become significantly more accessible and widespread.”

The word ‘disruption’ is important here. The idea of Disruptive Innovation comes from Clayton Christensen and is used as a kind of mantra to signal a significant change akin to ‘this is not business as usual’. It’s not quite a paradigm shift as described by Thomas Kuhn (1992). It is more ‘significant societal impact’ and seems to be reserved for entrepreneurs doing something in the marketplace.

Artificial Intelligence (AI) and a planetary digital network are the two most significant ‘one-time events’ that will transform our world say the authors. ‘Thinking machines’ that can become expert in pattern recognition and complex communication already exist. You can follow the development of one strand of this developing technology in the ongoing work to make ‘Chatbots’ more human-like. Here is the title of one article I found on Medium: “Chatbots Will Change Your Life But They Won’t Be Your Friend – Welcome to the post-app world.” Here is another about robots or bots: ‘What you really need to know’. Both these articles point to issues like ethical problems or not-yet-fully-human type problems.

Another article, this time from Slate, describes the perils of early, or perhaps we should say ‘too early’ release of AI. The particulars of these two cases are similar in the sense that they resulted in either untruthful or illegal actions from the AI in use. Here is what Slate concludes:

“Machine learning is expected to power advances in fields as diverse as finance, journalism, and medical diagnosis. As these technologies proliferate, the same dynamic that encouraged Uber to put flawed self-driving cars on city streets will push other companies to release products with potentially dangerous flaws that we haven’t yet anticipated. The moral here is not that we should ban artificial intelligence or demonize the companies that develop it. But there’s a strong need for regulators and consumer protection agencies to better understand these technologies, so that they can apply the proper scrutiny and enforce standards where necessary. Artificial intelligence from the likes of Google and Uber might reshape our lives in wonderful ways. But in the meantime, we should be wary of their early efforts.”

Elon Musk created OpenAI to try and steer the artificial intelligence movement.

“Because of AI’s surprising history, it’s hard to predict when human-level AI might come within reach. When it does, it’ll be important to have a leading research institution which can prioritize a good outcome for all over its own self-interest.”

Discussing Productivity Growth in US manufacturing the authors argue in Chapter 7 that late twentieth century GDP growth was a result of increased IT investment, ignoring the disastrous ‘Tech Bubble’ from 1995-2001 and burying a contrary perspective by Susan Housman in the endnotes. The argument is based on a correlation of productivity growth and IT investment which seems to be observable in the evidence the authors cite, however, they then add “a closer look at recent numbers tells a more nuanced story”. Productivity growth slowed after 2005, yet the authors say that “part of the recent slowdown simply reflects the Great Recession and its aftermath” although the GR didn’t start until 2008, so this is confusing. There is throughout the book a stance that seeks to portray a positivist attitude opposing the nay-sayers and ‘glass half-empty’ attitude.

Shifting gears, the authors then explore what they call the ‘Spread’ or the vast difference in wealth between the ‘winners and losers’ of the new economy brought on by the Second Machine Age. The ‘winner-take-all’ economy has spawned new terms, the ‘One-Percent’ and the ‘Ninety-nine Percent’ of the Occupy Wall Street movement. Spectacular wealth is created for winners in this economy through the Network Effect, where more users make a product more useful or valuable. A different analysis of the ‘Spread’ however, is found in the work of Thomas Picketty’s Capital in the Twenty-First Century. Picketty’s central idea is that the cause of the disproportionate wealth observed in our time is due to a slower rate of growth in the economy relative to the higher rate of return on investments, especially where those investments are ‘protected’ such as inherited wealth.

Here’s John Mauldin again:

“We tend to talk as though growth, taxes, currencies, central banks, politics, debt, jobs, and technology are independent topics; but drawing hard lines between them is impossible. The more comprehensive and systemic our analysis can be, the better.” (from Thoughts from the Front Line).

Brynjolfsson and McAfee turn to economic theory to suggest we are not out of luck with the ‘inflection point’. We know what to do.

Another important response in the new paradigm to the rise of robot labour is the idea of a Universal Basic Income (UBI). (See: Develop this!

We are in a time of change and although changes seem to occur at speeds much faster than we feel we can handle, we can act in some very important ways to make the changes that are coming less disruptive. In a chapter entitled Policy Recommendations the authors discuss six areas we, as a society (remember, the ‘we’ here is American society in particular) can influence change. The theme here is that we do know what to do, so lets just do it.

Education is the first area discussed because better educated populations have better job prospects and earn more money than poorer educated populations. Technologies, like MOOCs and the best use of analytics, plus recognition that Teachers matter more than anything in successful learning are the core ideas. Education Reform in the US is a huge topic and the brief nod to its importance is the main point here. The authors cite Charter schools as the solution but seem unaware of the controversy around this topic. Next, through Startups or entrepreneurship job creation is boosted. The biggest hurdle to the creation of Startups though is the proliferation of ‘regulations’ which slow or hinder the growth of new companies. The subtext here is that big government or too much government is the culprit. Another area we can leverage is matching human supply and demand described in terms of talents and abilities matched with employers needs. Government should, on the other hand, support innovation by funding research and protecting patents and copyrights, although that is also a tricky field to navigate properly. Finally, something should be done about the decaying infrastructure, again, probably a Government thing to do but the authors are comfortable with this Keynsian idea because of the ‘positive externalities’.

Longer term the authors see a growing issue around the loss of work. The chapter Avoiding the Three Evils introduction begins with a quote from Voltaire “Work saves a man from three great evils, boredom, vice, and need”. Solutions proposed include a Basic Income but this, the authors fear, would lead to Voltair’s evils. What then?

What is the role of Government in this debate? Here is Robert Schiller on this issue:

“When many people are no longer able to find work to support a family, troubling consequences ensue, and, as Phelps stresses, ‘the functioning of the entire community may be impaired.’ In other words, there are externalities to robotization that justify some government intervention.”

The suggestion in the book is to put a tax on the Robots that replace the Humans. Although robots will contribute to increased productivity in the long run, a tax in the short run or transition to the new economy may be a good idea.

Singularity University offers this:

“As technology enables teams, big and small, to make an impact as never before, leaders and organizations need to reimagine who they are serving, what they are serving, and how they are serving them in viable, sustainable and profitable ways. Businesses no longer need to choose between maximizing profit and helping society. They can choose to do both.”

I am interested in education, so what are the implications of the Second Machine Age for this discipline? Here is Audrey Waters, a well known technology skeptic:

“It was then and there on that trip that I had a revelation about how many entrepreneurs and engineers in Silicon Valley conceive of education and the role of technology in reshaping it: that is, if you collect enough data – lots and lots and lots of data – you can build a map. This is their conceptual framework for visualizing how “learners” (and that word is used to describe various, imagined students, workers, and consumers) get from here to there, whether it’s through a course or through a degree program or towards a job. With enough data and some machine learning, you can identify – statistically – the most common obstacles. You can plot the most frequently traveled path and the one that folks traverse most quickly. You can optimize. And once you’ve trained those algorithms, you can apply them everywhere. You can scale.

It’s a model. It’s a metaphor.

It’s an aspiration – a human aspiration, to be clear. This isn’t what machines “want.” (Machines have no wants.)”


“Robots don’t apply for jobs. Robots don’t “come for jobs.” Rather, business owners opt to automate rather than employ people. In other words, this refrain that “robots are coming for your job” is not so much a reflection of some tremendous breakthrough (or potential breakthrough) in automation, let alone artificial intelligence. Rather, it’s a proclamation about profits and politics. It’s a proclamation about labor and capital.”

This book is getting older now, and exponentially speaking what was once one year old is really two years old, and what is two years old is really four years old, and what is four years old is really eight years old, and what is eight years old is really sixteen years old. Let’s see, if this is April 2017 then sixteen years ago we had just escaped Y2000 and 9/11 was still to come. World changing events, both.


To explore what machine learning could mean in education, EdSurge convened a meetup this past week in San Francisco with Adam Blum (CEO of OpenEd), Armen Pischdotchian, (an academic technology mentor at IBM Watson), Kathy Benemann (CEO of EruditeAI), and Kirill Kireyev (founder of instaGrok and technology head at TextGenome and GYANT). EdSurge’s Tony Wan moderated the session.

Douglas Copeland has this to say on the ‘Nine-to-five is barbaric’ in 2017:

“In such a rapidly evolving society, possessing actual skills – including those which have nothing to do with the internet – is vital, says Coupland. “The winners in this labour force will be the people who have an actual skill,” he says. “Always have an actual skill as a back-up, that’s very good advice.”

You can see The Second Wave of the Second Machine Age – presentation at Oxford Martin Institute, April 2017.



What does this mean for business and the economy? First, a bigger pie is created because productivity can be better. There are objections to automation of course, because it leads to growing inequality among other things.



There is an increase in the differential in earnings depending on education and skill.

There is an increase in the differential in share of wealth with Capital getting more than Labour.

There is a concentration of wealth going to superstar success (the .01 percent).


Implications for the future of work or labour are not all dire.

Technological improvement is occurring at a faster rate than social responses (skills and policies). “The state of implementation is so far behind the frontier of technological change”.



The world of rationality may accept this thesis but the world is also made up of blind forces, belief and ignorance are two of them.

Another problem arises with ‘algorithmic bias’ when the writers of the program build in bias. Bias can creep in from machine-learning as well where machines learn from a great number of repetitions of a case, but the case contains flaws often based in human error.

Here is an example from Science: Self-taught artificial intelligence beats doctors at predicting heart attacks (

“Kontopantelis notes one limitation to the work: Machine-learning algorithms are like black boxes, in that you can see the data that go in and the decision that comes out, but you can’t grasp what happens in between. That makes it difficult for humans to tweak the algorithm, and it thwarts predictions of what it will do in a new scenario.

Will physicians soon adopt similar machine-learning methods in their practices? Doctors really pride themselves on their expertise, Ross says. “But I, being part of a newer generation, see that we can be assisted by the computer.”

From Maciej Cegłowski‘s Idle Words: “Because machine learning tracks human performance so well in some domains (like machine translation or object recognition), there is a temptation to anthropomorphize it. We assume that the machine’s mistakes will be like human mistakes. But this is a dangerous fallacy.”

And this: “Whatever bright line we imagine separating commerce from politics is not something the software that runs these sites can see. All the algorithms know is what they measure, which is the same for advertising as it is in politics: engagement, time on site, who shared what, who clicked what, and who is likely to come back for more.

The persuasion works, and it works the same way in politics as it does in commerce—by getting a rise out of people.”

Try this: “If you want an interesting surprise, go to your Google Maps history. Unless you’ve explicitly turned it off (or ‘paused’ it, in Google’s parlance), you’ll see your physical trail as reported by your mobile phone for the past several years (if you have an Android phone, or have Google maps installed with the necessary permissions). Not only does Google keep it indefinitely, but it’s available for the asking to anyone who gains access to your account.”


David Otter and Michael Polanyi – ‘Polanyi’s Paradox’

John Maynard Keynes – Economic Possibilities for Our Grandchildren

Two contemporary technology superstars, Elon Musk and Mark Zuckerberg, engaged in a public dispute over AI and it’s meaning or place in human society. On the one hand, machine intelligence is seen as a positive addition to the human tool kit and on the other a lurking danger to human life, characterized by the commonly understood Frankenstein story. Who controls the future? Contemporary breakthroughs in AI aren’t life altering but they are potentially useful. Consider face recognition and speech recognition or language translation. These are positive outcomes, useful to society. The concern that Elon Musk has expressed about AI is about the framework of regulation that only government can provide. This is the quandary of the ‘known unknown’ versus the ‘unknown unknown’.

See the Atlantic story here and a similar story in The New Yorker, here and this piece in Medium: ‘Why Elon Musk is Right About AI Regulation’.

Leave a Reply

Your email address will not be published. Required fields are marked *