Rather than deskilling work, we need to upskill employees. Rather than outsourcing low-value jobs, we need to increase the creative content of every role. Instead of assuming that middle-class jobs must ultimately fall to globalization and automation, we need to redesign work environments so they elicit the everyday genius of every human being.

By taking labor ever more out of the equation, automation removes any advantage countries with lower wage demands might have, because the costs of technology, unlike labor, are pretty much the same everywhere.

However, automation is not only likely to entrench further structural inequality between countries. Without a fundamental shift in the way economies are organized, it will dramatically exacerbate inequality within many countries as well. It will do this firstly by diminishing opportunities for unskilled and semi-skilled people to find decent employment, while simultaneously inflating the incomes of those few who continue to manage what are largely automated businesses. As importantly, it will increase returns on capital rather than labor, so expanding the wealth of those who have cash invested in businesses, rather than those who depend on taking cash from them in exchange for labor. This means straightforwardly that automation will generate further wealth for the already wealthy, while further disadvantaging those who do not have the means to purchase stakes in companies and so free-ride off the work done by automata. Of course, this would not be as much of a challenge were it not the case that since the Great Decoupling, the wealthiest 1 percent of people globally has captured twice as much of the new wealth generated by economic growth as the rest of us. The richest 10 percent of people on earth now own an estimated 85 percent of all global assets, and the richest 1 percent own 45 percent of all global assets.

In a recent article called “Are Ideas Getting Harder to Find?,” economists from Stanford and LSE analysed this phenomenon quantitatively. Across a range of industries, across firms, and in the aggregate economic data they found the same thing: progress becomes harder and harder. Based on their numbers, in order to double our overall level of technological advancement, we need to put in, conservatively, four times as much research effort as we did for the previous doubling. To illustrate, suppose (simplistically) that initially it took 10 person-years of “research” to double the world’s level of technological advancement: to move from knowing only how to make a stone axe to knowing how to make both an axe and a spear. In order to get the next doubling of technological progress, it would take 40 person-years of research. The next doubling would take 160 person-years, then 640 person-years, then 2,560 person-years, and so on.

My critique of GDP measurement in practice relates to its inability to report sufficiently accurately what it is intended to measure: the value of economic output. And pronouncements based on such data about long-run trends in income or the rate of increase of productivity should be taken with a grain of salt – or several. When pundits fret over whether the latest figure for GDP growth is an annual rate of 1.8 per cent or 1.9 per cent, they are fussing over differences that are insignificant in relation to the fundamental and inescapable uncertainties in the data they are citing.

In the abstract, economic theory held: swift redundancies would lead to quicker recovery and hence new, better jobs. But he stopped believing in it. Or rather, he gradually became convinced that the axe wasn’t falling in the places that theory called for. The bosses weren’t losing their jobs or homes. They weren’t leading from the front and sacrificing themselves for the greater good. How many ordinary jobs could you save with the CEO’s salary? It was simple arithmetic. McKinsey, too, was making a healthy surplus that could have kept most of these people employed, at least for another year. There had to be a better way to fit this puzzle together.