First in an occasional series on the future of work.

The pro-mining coalition that is pushing the proposed PolyMet ­Mining project in northeastern Minnesota decided to get right to the point — by giving itself the name "Jobs for ­Minnesotans."

It's backing PolyMet's promise to put at least 300 people to work mining ­copper and other metals, with 60 more likely down the road. These jobs will be welcome in a part of Minnesota that has generally had higher unemployment and lower household incomes than the state as a whole.

There's a great resurrection aspect to the story, too, because the company plans to restart operations at the site of an Erie Mining plant in Hoyt Lakes that was shut down in 2001.

But even if PolyMet eventually hires those 300 people, it won't be anything like the Erie plant's heyday. Back when that facility was booming, the payroll peaked at more than 10 times the number PolyMet promises.

The projects are different: iron mining at Erie vs. a new copper and nickel operation. Erie's operation was more self-contained, while PolyMet would draw on local suppliers who will also employ people.

Still, the numbers are ­stunning. Advocates are ­celebrating a controversial project that promises to employ one person for every 10 who once worked at the site.

This is now reality in large areas of the American economy, as new machines and more-efficient ways of making things reduce the need for people. In mining, the trend took hold three decades ago, when the industry responded to a threat of Brazilian competition by changing a lot of work rules and effectively doubling the output of each worker.

At PolyMet, there would be everything from processing machines that run and monitor themselves to massive 240-ton haul trucks that are easily twice the size of the trucks back when the Erie mine payroll peaked.

This shift in the economy — better capital equipment and automation leading to dramatic productivity gains — has long been predicted. Economist John ­Maynard Keynes famously looked decades ahead in a Great Depression-era essay and suggested "we may be able to perform all the operations of agriculture, mining, and manufacture with a quarter of the human effort to which we have been accustomed."

He nailed the trend but was off on the details. These operations don't take nearly a quarter of the labor they did back then.

Running an ore-processing machine is a classic example of a middle-skill job, a kind of job that's being threatened by better and better technology. They are called middle skill because training and experience really matter but they don't usually require university degrees or professional credentials, like a lawyer.

These jobs pay well. The more than 4,000 jobs in the iron-mining industry in northeastern Minnesota have an average wage and benefits package worth roughly $100,000, according to the Iron Mining Association of Minnesota.

As these kinds of jobs come under increasing pressure, it has created a strange paradox in the labor market, with people worried about jobs and job security at a time when the state unemployment rate bumps along at less than 4 percent.

It's possible to hear big employers complain of a labor shortage and a "war for talent" the same week state and local officials talk up an unproven plan to require more workers.

The issue is that, as middle-skill jobs go away, comparable new jobs aren't emerging. The occupations with the greatest number of projected openings in Minnesota are, in order: retail sales, cashiers, food preparation and ­restaurant server.

The good news is that these are jobs that don't sound all that easy to replace with a machine. The bad news is they likely won't pay well enough for a comfortable middle-class life.

Trend likely to accelerate

Nonetheless, the impact of automation and machines on work is likely to accelerate. A new McKinsey Global Institute report said about half of the tasks we do for work could be done with "currently demonstrated technologies." You maybe hadn't heard, for example, that Amazon.com almost single-handedly kicked off a warehouse robot arms race by buying a leading robot maker and deciding not to sell the machines to ­anybody else.

That's not to say there's no hope for distribution-center employees who get fired to make way for a robot. Economists have long cautioned not to worry too much about ­losing jobs to machines.

For one thing, it's misleading to think that there's just one big lump of work to do. If a machine does work it won't necessarily mean we'll have less to do, as new wants and needs emerge and people get paid to meet some of them.

It's also probably pointless to complain about the adoption of new technology. The technologists will patiently explain to people my age that the pace of change is accelerating, as the gadget era comes to a close and artificial ­intelligence takes off.

Not that long ago it was easy to read assurances that even artificial intelligence would be no match for humans in a lot of the tasks we do, ones that demand improvisation and judgment. Driving a car — safely reacting to the unpredictable nature of the road — used to be considered one of those tasks. Just last week, the ride-hailing service Uber struck a deal to work with the parent company of Mercedes-Benz on self-driving cars.

Just taking people home after the bars close isn't the vision of self-driving auto enthusiasts. Why not self-driving trucks? At last count there were more than 3.5 million big trucks on the road, and to get anywhere they each need an expensive human driver.

It's impossible to say when, if ever, those drivers will become obsolete. What is clear is that there would be even fewer good-paying jobs at places like PolyMet if those 240-ton mining trucks start driving themselves.

lee.schafer@startribune.com 612-673-4302