Reading: Satya Nadella, “Hit Refresh”

Satya Nadella became Microsoft’s third chief executive in February 2014, inheriting a firm whose principal product franchises were under pressure on every front: Windows from Apple’s mobile platform, Office from Google’s web suite, the server business from Amazon’s cloud. Hit Refresh, which he published in 2017, is the book he wrote three years into the rebuilding — early enough that the outcome of the rebuilding was not yet certain, late enough that the operational principles he had been working from could be set down. The book is largely a memoir; the chapter on automation, artificial intelligence, and the future of work is the one that bears most directly on the practice’s clients and is the chapter the practice has carried into client conversations.

Nadella addresses a question almost every operations practitioner is asked at some point and most decline to answer well: if the work this firm does can be increasingly done by machines, what becomes of the people who do it now? His position is unusually plain for a chief executive of a firm that sells the machines:

> Finally, questions are being asked about whether this next industrial revolution will be a jobless one. To help us investigate this question, MIT economist Daron Acemoğlu visited our campus to report on his research into the effects of technology automation on labour. He found that new intelligent machines, particularly industrial robots, could have very consequential effects on the labour market. His estimates suggest that, on average, each additional industrial robot reduces employment by about three workers. > > Nevertheless, Acemoğlu argues that other powerful changes triggered by this onslaught could at least partly reverse these consequences. As machines replace labour in some tasks, firms will be incentivised to create new tasks in which humans have a comparative advantage. Acemoğlu sums it up this way: “Although automation tends to reduce employment and the share of labour in national income, the creation of more complex tasks has the opposite effect.” > > We need a balanced growth path. We need to invest in a new social contract for this age of AI and automation that fosters the equilibrium between individual labour — one’s agency, wages, sense of purpose and fulfilment — and return on capital.

Two things in this passage make it worth re-reading. The first is the willingness to put a number on the disemployment effect — each additional industrial robot reduces employment by about three workers — without softening it. Most chief executives writing about automation in 2017 declined to do so; Nadella does so and proceeds. The second is the explicit recognition that the question is one of new tasks rather than the absence of tasks, and that the firm has work to do in creating those new tasks rather than waiting for the labour market to do it.

A second passage, on the same theme:

> Some argue that robots will take all our jobs, but this so-called “lump of labour” argument — the notion that there is a limited amount of work available — has always been disproved. It’s just that different types of labour will be needed. And humans will add value where machines cannot. As we encounter more and more artificial intelligence, real intelligence, real empathy, and real common sense will be scarce. The new jobs will be predicated on knowing how to work with machines, but also on these uniquely human attributes.

The Lean implication is direct. The discipline of Lean has always been, in its central commitment, a discipline of respect for people — Toyota’s own term, jinkado, the human-systems-and-machines integration the company organised its operations around. A firm whose Lean transformation is essentially a head-count reduction has misunderstood the discipline; a firm whose transformation creates more complex work that the operators previously displaced are then trained into is doing it correctly. Nadella’s chapter is the clearest contemporary statement of why the second version is the only one worth pursuing, written by an executive whose own firm is in the business of building the machines that make the question pressing.

The book is published by Microsoft’s own house, and reads, in places, like a corporate manifesto. The chapter on automation reads like something else — a careful operational position taken by a chief executive who has thought about the question harder than the question is usually thought about, and who has the standing to be heard.

· · ·