My Views on The Second Machine Age

With ‘The Second Machine Age‘, Erik Brynjolfsson and Andrew McAfee have written a fascinating book, which deservedly parks in many’s books of the year lists.  Written by two MIT scientists, it tracks the impact so far of the digital age and looks at the future.  The authors claim the impact of digital technology will be as big, or bigger than that of the first machine age, and will reinvent our lives and economy.

Some key points from the book:

  1. Although the PC was Time’s ‘Person’ of the year in 1982, the full impact of digital technologies has only recently reached full force.  There are different reasons for this:
    • Moore’s Law. First formulated in 1962, it roughly stated that the amount of computer power you can buy for one dollar would double each year.   This ‘law’ has proven remarkably correct for over 40 decades.  It has even proven to be correct for many digital technologies.  That continuous doubling has brought us in the ‘second half of the chessboard’ where the absolute amounts increase phenomenally.
    • Digitization of information.  More information is available in digital format.  With the ‘internet of things’ massive amounts of digital data are created.
    • Network effects.  The more people are online and using an particular platform, the stronger the benefits.
    • Organisational capital.  Companies investing in IT, typically spend about 10 times that amount on changes in the organisation of the company.  Using IT to do existing things faster does not increase productivity much.  It’s using IT to do things fundamentally different what makes the difference (a comment often formulated about using computers in classrooms as well).  Such fundamental rethinking of production processes usually takes several decades, as existing managers retire and are replaced by ‘digital natives’.
  2. Digital goods are fundamentally different from physical goods
    • They have a high cost to create, but are almost free to replicate unlimitedly;
    • They are not exhaustive, you can not ‘use up’ a digital good.
  3. These characteristics have profound effects on our economy
    • Job (and wage) polarisation.  Demand for routine manual and cognitive jobs has been falling, depressing wages.  Demand has been rising for skills that complement digital technologies (pattern recognition, assessing complex situations…).
    • “Winner takes all” markets.  We have been moving from markets with absolute advantages to markets where a small relative advantage leads to absolute dominance.  Being a slightly better writer 50 years ago would give you slightly higher scales, as the extent you could market and sell your writings would be limited.   There was a market for many writers, some slightly better or worse than others.  Today, a writer as JK Rowling can sell her books worldwide, dominating the market for fiction writing.
    • GDP as a measure of economic activity becomes less useful. GDP is based on physical reproduction of items.  When digital items are copied, people enjoy from ‘free’ digital services or produce intangibles such as intellectual property and human capital, these are not captured in GDP.
    • Traditionally, productivity growth has been linked to rising living standards.  The share of labour in GDP growth has fallen in recent years (to the expense of the share of capital).  This, together with the job polarisation, implies that rises in productivity do not longer automatically translate into broadly shared rises in welfare.  This explains the jobless recoveries after recent crises.

Whereas the authors are inherently optimistic about the digital age, they do see risks of high unemployment:

  • Job polarisation and “winner takes all” markets reduce demand for certain jobs.
  • Inelastic demand. Demand for goods may rise slower than prices fall as a result of technological progress.
  • Technological progress may be faster than the ‘adjustment time’ to re-skill workers for jobs that are in demand.

For the solutions, the authors look mainly at education.  This chapter is one of the weakest.  The authors advocate a re-think of the curriculum towards ’21st century skills’ and a stronger focus on entrepreneurship.  The authors put high hopes in MOOCs (only xMOOCs in fact, not the much more interesting cMOOCs), brushing aside the fact that they offer a sub-standard experience of what on-line learning can be.  They have a point that, as on-line learning becomes more mainstream, the characteristics of digital goods markets will impact education, such as job polarisation and ‘winner takes all’ markets. In South Korea, for example, lectures from ‘superstar’ teachers with  astronomical salaries are broadcast on-line, reducing the role of traditional teachers.

But nevertheless a great book on the analysis of how digital technologies have been affecting our lives, and are likely to disrupt it even a lot more.  For the part on how education could (and needs to) respond to our changing society, I would recommend other articles, books and blog posts.