I am with David Autor here: individual tasks that are components of jobs will be automated, but human thought and judgment will continue to be able to add value throughout the economy. There is, however, nothing to require that a world of abundant capital and sophisticated computers will be a world in which the income distribution will be relatively equal: David Autor: Polanyi’s Paradox: Will It Be Overcome?: "Jobs are made up of many tasks.... While automation and computerization can substitute for some of them, understanding the interaction between technology and employment requires thinking about more than just substitution. It requires thinking about... how human labor can often complement new technology... price and income elasticities for different kinds of output.... The tasks that have proved most vexing to automate are those demanding flexibility, judgment, and common sense—skills that we understand only tacitly. I referred to this constraint above as Polanyi’s paradox. In the past decade, computerization and robotics have progressed into spheres of human activity that were considered off limits only a few years earlier—driving vehicles, parsing legal documents... agricultural field labor. Is Polanyi’s Paradox soon to be at least mostly overcome, in the sense that the vast majority of tasks will soon be automated? My reading of the evidence suggests otherwise...
...Polanyi’s paradox helps to explain what has not yet been accomplished, and further illuminates the paths by which more will ultimately be accomplished.... Two... paths... to traverse to automate tasks for which we “do not know the rules”: environmental control and machine learning. The first... circumvents Polanyi’s paradox by regularizing the environment.... The second... develop[s] machines that attempt to infer tacit rules from context, abundant data, and applied statistics.... Machine learning.... Engineers... unable to program a machine to “simulate” a nonroutine task by following a scripted procedure... program a machine to master the task autonomously by studying successful examples of the task being carried out by others... exposure, training, and reinforcement... to accomplish tasks that have proved dauntingly challenging to codify with explicit procedures....
Consider the task of visually identifying a chair... Pre-specifying requisite features—and more sophisticated variants of this approach—would likely have very high misclassification rates. Yet, any grade-school child could perform this task with high accuracy. What does the child know that the rules-based procedure does not? Unfortunately, we cannot enunciate precisely what the child knows—and this is precisely Polanyi’s paradox. Machine learning potentially circumvents this problem.... [via] large databases of so-called “ground truth”.... Machine learning does not require an explicit physical model of “chairness”... an atheoretical brute force technique.... If the majority of users who recently searched for the terms “degrees bacon” clicked on links for Kevin Bacon rather than links for best bacon cooking temperatures, the search engine would tend to place the Kevin Bacon links higher in the list of results....
The tools are inconsistent: uncannily accurate at times; typically only so-so; and occasionally unfathomable. Moreover, an irony of machine learning algorithms is that they also cannot “tell” programmers why they do what they do.... The underlying technologies—the software, hardware, and training data—are all improving rapidly.... Some researchers expect that as computing power rises and training databases grow, the brute force machine learning approach will approach or exceed human capabilities. Others suspect that machine learning will only ever “get it right” on average, while missing many of the most important and informative exceptions.... One is reminded of Carl Sagan’s (1980, p. 218) remark, “If you wish to make an apple pie from scratch, you must first invent the universe”...
