Technology has long displaced jobs and occupations, often with little warning. The good it brings is immense but often overlooked, while the much smaller costs are more concentrated. Though the net effect is overwhelmingly positive, the workers who are unable to adapt usually fare poorly.

The adoption of machines and technology to the workplace displays all the regularity that economic theory suggests. Firms strive to maximize profits with these technologies, which has led them to invest in capital that both replaces workers and complements their skills. Machines and computers have displaced the workers who perform routine but costly tasks. While this has displaced some occupations in all skill levels, it has fallen hardest on middle-skill workers. Notably, these are the workers who benefitted most from early 20th century technologies.

Technology has also been used to complement the skills of workers. In recent decades, complementary technologies tend to go first to the places where skills are most scarce. So, technology investment has first been deployed to very high-skilled workers. At the same time, our overall consumption patterns have shifted away from goods and towards services that are primarily provided by workers at the high-skill and low-skill ends of the spectrum.

A consequence of these economic forces is that, as demand for their services has grown, employment growth has polarized among very high-skilled and very low-skilled workers. Typically, higher demand would lead to wage growth for both categories. However, the bulk of displaced workers moved down the skill ladder, not up. That has created an abundance of workers competing for low-skilled jobs and a scarcity of workers at the top. This, rather than the mysterious forces of globalization or corporate capitalism, accounts for most increasing income inequality.

This might seem altogether a dismal accounting of the past half century, but the same economic forces that brought us to where we are now remain ruthlessly at work. In that we should be most grateful. In future, many of these same economic forces beckon growth in middle-skill occupations where people, not machines dominate. Human skills (defined by MIT’s David Autor as interpersonal interaction, flexibility, adaptability and problem solving) will surely grow in the coming decades.

I think there is good anecdotal reason to believe this is true. Many hotel maids now use iPads in their jobs, while machines replacing pharmacists will be ubiquitous in a decade. It is also true that most new occupations cannot readily be off-shored, so labor market competition will not depress wages. Health and personal services, construction and maintenance operations are good examples of middle-skill occupations where technology will increasingly complement, but likely cannot readily replace, human skills.

We appear likely to enter a lengthy period where middle-skill jobs will be in greater demand than those at the bottom and top of the skill ladder. We need to find ways to make workers better ready for this. Ironically though, this shift will be slowed if we surrender our good judgment to the siren song of anti-globalization zealots.