Over at Equitable Growth: The Four Must-Reads You Need to Understand Thomas Piketty and His "Capital in the Twenty-First Century": Focus
Liveblogging World War II: February 13, 1945: Firebombing of Dresden

Over at Equitable Growth: Has David Autor Gotten Himself Hopelessly Confused with Respect to Average and Marginal Products?

Over at Equitable Growth: Something has bothered me ever since I read the highly-eminent and highly-esteemed David Autor's "Polanyi's Paradox and the Shape of Employment Growth":

David Autor (2014): Polanyi’s Paradox and the Shape of Employment Growth: "[The] human tasks that have proved most amenable to computerization...

...are those that follow explicit, codifiable procedures.... Tasks that have proved most vexing to automate are those that demand... skills that we understand only tacitly.... The interplay between machine and human comparative advantage allows computers to substitute for workers in performing routine, codifiable tasks while amplifying the comparative advantage of workers in supplying problem solving skills, adaptability, and creativity. Understanding this interplay is central to interpreting and forecasting the changing structure of employment in the U.S. and other industrialized countries.... READ MOAR

As the price of computing power has fallen, computers have increasingly displaced workers in... middle-skilled cognitive and manual activities, such as bookkeeping, clerical work, and repetitive production.... But the scope for substitution is bounded.... There are many tasks that we understand tacitly and accomplish effortlessly for which we do not know the explicit “rules” or procedures... break an egg over the edge of a mixing bowl, identify a distinct species of birds based only on a fleeting glimpse, write a persuasive paragraph, or develop a hypothesis to explain a poorly understood phenomenon....

Let me interrupt to say that I would be serious money that soon computerized robots will be breaking eggs and identifying omelets, and scrutinizing satellite photographs to conduct bird censuses, so it seems to me that at least two of Autor's four examples are badly-chosen from the viewpoint of a generation from now.

Autor continues:

At an economic level... tasks that cannot be substituted by computerization are generally complemented by it. This point is as fundamental as it is overlooked.... Productivity improvements in one set of tasks almost necessarily increase the economic value of the remaining tasks.... By historical standards, contemporary construction workers are akin to cyborgs. Augmented by cranes, excavators, arc welders, and pneumatic nail guns, the quantity of physical work that a skilled construction worker can accomplish in an eight-hour workday is staggering.... Construction workers have not been devalued by this substitution. Despite the array of capital equipment available, a construction site without construction workers produces nothing.... A worker wielding a single shovel can do a fairly limited amount of good or harm in an eight-hour day. A worker operating a front-end loader can accomplish far more. To a first approximation, automation has therefore complemented construction workers--and it has done so in part by substituting for a subset of their job tasks...

But I was always taught that factors of production were not complements if additional quantities of the one increased the average product of the other.

I was always taught that factors of production were complements if and only if additional quantities of the one increased the marginal product of the other.

If additional quantities of the one decreased the marginal product of the other than the factors were not complements but substitutes--even though the average product of the other factor went up.

Is anything Autor has said relevant to the effect of computerization on the marginal product of construction workers, or indeed of production workers, or wage- and salary-workers, or indeed of labor more generally? Why don't graphs like this create a prima facie case for Autor and others that while labor (in the form of human brains serving as cybernetic control mechanisms for blue-collar machine-managing production and white-collar information-managing paper-shuffling tasks) was a complement to industrialization up until the coming of the Information Age, that now labor is not now a complement to computer-embodying machines but a substitute for them?

Graph Nonfarm Business Sector Labor Share FRED St Louis Fed

I mean: am I missing something very basic here?