Incidentally, talking of McKinsey, here is a link to their latest study of productivity trends. They argue that labor-saving productivity improvements do not destroy jobs over time. Generally reductions in the price of the goods produced mean comsumers now have additional resources to spend on new and different goods. Time to shop!
Historically this is true. We eliminated most jobs in subsistence agriculture in the west, which of course employed the majority of the population in medieval times. We are not worse off for it.
But the nature of the new goods and services which appear matter as well, and the nature of the demand and preferences themselves. The study does not really get to grips with this. If people can afford a car for the first time, two or three generations ago, that boosts an auto industry that employes hundreds of thousands in everything from steel production to car loans to final assembly to retail. If people now choose to spend more time on Facebook, how many extra jobs does it create? Not many.
Too much economic commentary has in essence a faith that "something will turn up" to employ people providing new goods and services, because it always has before. But if the nature of needs and technology is changing, you can't simply assume that.
There is plenty of scope for improvements in productivity in existing goods and services, too, expecially in sectors like healthcare. This line from the report really struck me: "Today, nurses still spend less than 40 percent of their time with patients and the rest on paperwork." Healthcare has been one of the biggest sources of job growth in the last ten years. But it is extremely inefficient.