The general trend is toward deskilling (via machines, technology, etc.) to make labor both more flexible (anyone can work anywhere) and a commodity, weakening labor's bargaining power and permitting capital to extort more surplus value, which is no longer a matter of extending working hours but a matter of improving productivity (Marxian jargon for this is "relative surplus value").
This is probably worth remembering in any conventional economic discussion of how boosting productivity is a good thing for the economy. It tends to increase growth (we get more output from less input) but it doesn't mitigate inequalities in distribution of the output -- it tends to leverage the inequities and increase the gap between classes by allowing capitalists to seize more surplus value from smaller pieces of capital. Increased productivity figures also indicate more labor is being squeezed from fewer workers as the "slack" is being wrung out of offices -- i.e. workers who were not laid off are forced to make up for the work of those who were. Productivity gains would be nice if that freed us to do other things with our time, only society is set up to deny us that possibility, stigmatizing unemployment and refusing a social wage independent of work for capitalists. Through productivity adds to GDP, it makes workers more rather less insecure.
Productivity gains, if we accept the Marx/Braverman analysis, are closely allied to deskilling. As Aronwitz explains:
For Marx, the increasing scale of production made possible by machinery, and by capital's intervention in the labor process by its devalorizing labor, becomes identical with the mode of production. The production of relative surplus value...is the sole purpose of production. The subsumption and subordination of science, technology, and human labor are merely facets of this compulsion toward extracting profitsDeskilling removes the limits on capital placed by traditional, artisanal modes of production, permitting a greater division of labor and a more rapid turnover of materials. This has broader ramifications that making work that much more meaningless to the Taylorized worker: "The new machines are forms of the social organization of labor introduced by management," Aronowitz notes. The minute division of labor is institutionalized; workers begin to enter the world of production without prerequisite skills. It becomes understood that one gets a job and is taught to operate a machine (or nowadays a piece of software), which already contains the skills and technology (which now belongs safely to the capitalists who own the machines).
We often hear that education is the key to success in the labor markets. "Skill-biased technical change" is invoked to explain income inequality, an effort to persuade us that it is not merely a matter of the rich getting richer. (Economist David Card disputes the explanatory power of SBTC in this 2002 paper.) And much-remarked-upon charts like this one (via Calculated Risk) seem to prove the power of education.
Does this refute the idea that capitalism thrives by deskilling the labor force? Or does this merely show that education is an extremely valuable signaling device, useful to employers for sorting docile and tractable workers from those less inclined to follow rules and follow through on arbitrary projects. From that point of view, education is about credentialing as opposed to learning. In other words, the skills themselves may not explain the differences in our various fates in the labor market. Education matters insofar as it proves membership to an employable class, but the "skills" one acquires there rarely let people set up as capitalists themselves. The skills are typically not "human capital" in that regard -- they do not allow us to confront capital on a level playing field. Instead the skills we acquire are more about teaching us the proper modes of submission to bosses and the virtues of "go along, get along."