The Second Machine Age

I’ve just finished reading The Second Machine Age for a third time. [Note: It’s how I tend to read nonfiction. Skim-read the first time around, bookmarking stuff for a more leisurely read at a later stage. Read cover-to-cover the second time around, no notes, no interruptions, no nothing, just absorbing the whole argument. The third reading then concentrates on following up the bookmarks and notes. And yes, I do this all “analogue”].

You should read it. At least once anyway, depending on your reading style.

Before I explain why you should read the book, some disclosures:

  • I know both the authors reasonably well. I knew of Erik Brynjolfsson through his work on Incomplete Contracts in the early 1990s, and met him for the first time in 1997 after joining Dresdner Kleinwort (DrKW). I met Andy McAfee first off in late 2004/early 2005, while I was Global CIO at DrKW and while he was still at Harvard, when he looked closely at the workings of our department as part of his research. I’ve kept in touch with both of them since.
  • I trained as an economist, with considerable emphasis on mathematics and statistics, and in all probability was influenced by those anchors and frames when reading the book.
  • I believed, and continue to believe, that ubiquitous affordable connectivity (with the requisite compute power and storage) can transform lives and help us solve hitherto intractable problems in health, welfare and even government, principally through education and training. Which meant the core message of the book resonated strongly with me. You can see this reflected in the Kernel I wrote for my blog around a decade ago (and even in my belief that Jerry Garcia and the Grateful Dead taught me more about modern business models than anyone else did).

Most of you know all this already. It still behooves me to declare this, to show my biases openly, even though I have no financial interest in any of this.

And so on to the book.


I read a lot of reviews, usually in magazines I subscribe to. I tend to read Amazon Reviews only “after the fact”, having already decided to investigate a particular book or film or piece of music. When it comes to reviews, the ones I’ve liked the most have come from Kirkus; I’ve grown to rely on them over the years, despite changes in style and ownership.

When it comes to writing reviews, it’s something I do rarely. And when I do, I try and focus on reviewing rather than summarising. My intent is to encourage readers to buy/rent the particular item, rather than to provide a substitute. So let me move on. Why should you read the book?

The book is about something that matters to all of us

The basic thesis of the book is this: We’re creating, adapting and using digital technologies at an incredible pace today. As a result, we’re seeing many beneficial outcomes, and some that aren’t beneficial. How can we retain and grow the benefits while minimising the harms? The authors spend time on each aspect of this: the current state of digital technologies; how we got here, the history and background; where we appear to be headed.  The benefits that have accrued as a result, the benefits we’re accruing, the benefits we have yet to gain. And the harms: what hasn’t worked well, what’s disturbing in current trends, what the “default” prognosis is. What we should do about it, as individuals, as corporations, as governments, now and in time to come.

None of us is in a position to do much about the harms unless we understand the root causes and the benefits as well. Particularly when it comes to policy and regulation in the context of anything “digital”, the current track records of governments worldwide leaves much to be desired. If you add “jobs” and “inequality” to the mix, then the undesirability increases exponentially. So it is important for us to understand just what’s happening and why. This is not the first book in the world to look at this, nor is it the only one. But it frames it well, and focuses on the right issues.

Why the right issues? Because it spends time looking at the nature of the jobs that will be created as well as the nature of the jobs that will disappear. The inequalities that have grown as a result, the inequalities that will continue to grow if we don’t do something about it. What that something could be, how and why we should intervene. Why all this is happening now.

Yes, these things matter. And we need to be discussing them dispassionately and apolitically, with a true global perspective. Which is why this book is important.

You may feel you’ve read it all. Perhaps you have. What matters is not just the ground covered but the context and sequence in which it’s being covered. So even if you think you know it all, do read it.

It’s accessible, approachable, easy to read

You don’t have to be an economist or a statistician to read the book. Adopting a flowing conversational style, the authors take you through a journey in three parts: the origins and characteristics of the Second Machine Age, what we gain or lose as a result, what we can do about it. That may be enough for most readers. For the more curious, there are references and quotations peppered throughout the book. There’s also a 32-page Notes section at the end of the book, well worth delving into.

It comes in many formats: not just print and digital but audio as well (and I’m sure there will be video versions even if only in TED-talk format). To each his own. The key point here is that you can choose how you want to engage with the book. As far as I can make out, it is currently only available in English. I’m sure that will change, given time.

What I liked about it

The authors spend time looking into a number of issues of interest to all of us. Here’s a sample:

  • reasons for exponential growth: the chessboard’s “second half” construct
  • the importance of “recombination” in innovation
  • differences between economic growth and GDP growth: where that matters, where it shouldn’t
  • the things that computers are good/bad at, what’s changing in this respect
  • why, in robotics, “hard problems are easy, easy problems are hard”: the implications thereof
  • what happens in industries where the “first copy” is expensive, all others are cheap
  • why innovation could be “used up” in the past, and why this is no longer the case
  • reasons why inequality has increased, why winner-take-all models are common
  • the share of economic return captured by capital rather than labour, how it’s changing

I was particularly taken with a quote from Garry Kasparov (incidentally, one of at least two misspellings I noted during my three reads, it’s the old proofreading journo in me). Kasparov was talking about a chess match with Veselin Topalov, in which both players were allowed to “consult” computers whenever they liked:

Since we both had equal access to the same database, the advantage came down to creating a new idea at some point

In some ways, rightly or wrongly, I see this as the seminal idea of the book — The onset of digital technologies has some hard-to-bear consequences in terms of the commoditisation of skills that were historically valuable and in terms of sharply lowering prices of goods that were historically valuable. This onset is happening faster, more widely and more deeply than people expected, and inequality is growing as a result. To counter this, we have to teach people skills we’ve not paid attention to: creativity, ideation, invention and innovation, recombination; with the social, political and economic policies and infrastructure in place to support all this.

What I’m thinking about as a result (and what I will probably continue to think about for some time): examples

  • The continuing shift towards our engaging with digital information using our senses, as the keyboard does its dinosaur dance: Siri and Kinect and Glass are just manifestations of technologies where we’re removing frictions like QWERTY from the process, and replacing them with our senses.
  • Reconciling the Carlota Perez “Paradigms”, the Tyler Cowen “Stagnations”, the W Brian Arthur “Increasing Returns” and the Hagel/Seely Brown “Big Shift” models and bringing them all into this context. They’re not conflicting, I’ve learnt from all of them, but the time has come to look at the confluence with care.
  • Looking more deeply into developing “filtering” as a new skill set in its own right, and thinking about the implications for jobs

But they’re not what’s important. What really really intrigues me is something altogether different, and that is this:

Today, I read that facebook has offered to buy Whatsapp for up to $19 billion. A classic example of a Winner-Takes-All market action, with the risk that inequalities will grow as a result….. unless the winners do something radically different. Which is where I think Marc Benioff (who is chairman and CEO of the company I work for) has the right answer in his 1+1+1 model. The Salesforce Foundation is part of the reason why I choose to work where I work, and I think this will increasingly be the case for job choices everywhere. We have to build into our chosen modes of employment, sustainable and scalable ways of reducing inequality.

Update: I have just seen this, a formal case study of the 1+1+1 model. Definitely worth reading in this context.

What I feel is left unanswered, or at least under-answered

  • The role of ubiquitous affordable connectivity in all this: why access to the internet needs to become a universal human right … and how this can become policy.
  • The role of platforms in all this: how the general purpose computing infrastructures will manifest themselves in future, why the concepts of APIs matter, how the democratisation of development will influence the redistribution of returns to labour (versus capital).
  • The role of governments in all this (as opposed to government: this is a global issue requiring cooperation between governments). As an Indian living in the UK working for a global company headquartered out of the US, I face a regular problem. Many of the debates I’m really interested in become, more by accident than design, completely US-centric. It’s not surprising given the dominance of the US economy and given the predominance of Silicon Valley. But it’s a problem. Particularly when it comes to creating the right infrastructure, policies and regulations to allow people to think anew, to dream anew and to convert their dreams into reality. The institutions that come in the way are largely governmental in origin.
  • The need to overhaul intellectual property regimes in order to reduce inequality. I won’t say any more about this other than to note that the previous sentence probably cost me a significant portion of my readership.

Please read the book. Let me know what you think.  And if you find that this review helped you, let me know. Who knows, I might even review another book.

3 thoughts on “The Second Machine Age”

  1. JP, 1+1+1 model, terrific. $19bn Winner-Takes-All market action, terrifying.

    On governments, both UK and US have policy creating an off-the-charts 1% by way of “biggest server effects”.

    Chiefs in Financial Services can buy assets, leverage these assets up with debt multiple times, then avoid paying income tax on “the carry” because it’s unearned based on debt.
    Concurrently the risks of doing so are very efficiently transferred away from the top to many end nodes using network effects to holders of notes in houses, education, car, etc.
    Lehman Brothers, Facebook, Blackrock, Autonomy –all use biggest server effects.

    Ladle on the Second Machine Age effects of robots picking up repetitive jobs from many people, we’ve now created a new class of folks unable to stay ahead of the robots?

    How do we make the incomes flatter when “affordable connectivity (with the requisite compute power and storage)” is creating such Winner-Takes-All market actions?

    Coda. Your review very helpful to shape my question :-)

  2. Thanks. On your recommendation I purchased and read the book. I agree, it’s an excellent presentation about how computers are changing the world. However, that is not what’s motivating me to reply.

    Your description of how you read books interested me. I’d like to recommend a slightly different approach. Amazon’s whispersync technology is fantastic for books with clear ideas you wish to go through and think think about clearly, individually and more deeply.

    I have to make a confession in describing this. For personal reasons, twice a week I make a 90 minute drive on a fairly straight and not heavily traveled road. At least when I’m driving it. Whispersync does an audio reading of the book and highlights that phrase that is just about to be read. Kindle books allow individual words to be easily highlighted and marked. Meaning you can highlight an area you want to return to and then go back an read it more throughly (preferable, while you’re not driving).

    This method allows me to skim a large number of books while I’m cooking, watching football and yes, driving. I can then go back and focus on the sections that are particularly interesting.

    While this approach might not be as thorough as your 3-time reading, it is very beneficial.


Let me know what you think

This site uses Akismet to reduce spam. Learn how your comment data is processed.