Sometimes there can be sharp differences in a Democratic presidential debate, like when Sen. Elizabeth Warren turned a question about the threat of jobs lost to automation by insisting that bad trade rules are really to blame.

Nonsense, responded lawyer and entrepreneur Andrew Yang, who noted that driving trucks is the top job in 29 states. He's got friends in California testing self-driving trucks right now. What will truck drivers and truck-stop staff do for work when the trucks drive themselves?

Any media that stepped in to fact-check this last debate could have left this one at "they're both at least a little right," if by "trade" Warren meant domestic job losses due to lower cost imports.

Unfortunately, what we know for sure is that politics remains the art of trying to win elections based not on what lies ahead but what should be done to fix the problems of the past. Voters have already experienced those problems.

That's what makes Yang worth listening to. Pressure on American jobs due to Chinese imports seems to have peaked over a decade ago. The adoption of technologies to let machines take over much of the work people do, on the other hand, is a transition that's just getting going.

What is usually called the China Shock to the American economy was certainly a real thing. It developed about when China was admitted to the World Trade Organization in 2001, although this basically meant China had to liberalize its economy. These changes had the effect of making its businesses more efficient and tougher global competitors.

What followed was a period that researchers later called "the great employment sag," as the number of working-age men in the United States without a job swelled and the growth in the employment rate of women stagnated. What really got hurt were so-called middle-skill jobs, like skilled production work. This was not the kind of economic effect that got evenly distributed throughout the country, as some industries were far more vulnerable to offshore competition than others.

Here in Minnesota, the people working in manufacturing, as tracked by the Bureau of Labor Statistics, slipped by about 25%, or 100,000 jobs, from the boom year of 1999 through the end of 2010. That was a decline from nearly 15% of the state's jobs in 1999 to less than 11% by 2010.

In a way, what happened with China was unusual, as generally Americans don't really want low-value products from low-wage countries. Americans wanted Nikon cameras from Japan and Peter Wolters grinding machines for their shop floors from Germany.

Better technology and know-how for managing longer supply chains and faster global shipping meant the story with China turned out differently. China could supply much of the labor that went into products with American brands on them. From 2001 through the start of the Great Recession, the value of imports from China nearly quintupled.

In the decade of the 2000s, when the transition was in high gear, the Star Tribune was running stories about how Americans were trying to compete.

One approach was doing more here for American customers, maybe by becoming good at making custom parts on tight deadlines or taking over a bigger chunk of the supply chain by delivering completed subassemblies. Other companies found their own Chinese suppliers or even set up shop there.

Moving production to a low-cost supplier could be a rocky transition. One problem, as a Bloomington plant manager once explained to me, was outsourcing key parts to a manufacturer on the other side of the world and then one day discovering in the latest shipment that one part seemed to no longer quite fit. The local manufacturer could hope that it was just a bad batch and that the supplier promptly caught and fixed its process error. But given that there were parts in transit from West Coast ports, on ships on the Pacific and in shipping containers still in China, bad parts might keep arriving for weeks.

By the time the Great Recession was in full swing in 2008 and 2009, the growth in Chinese imports of consumer goods as a share of the U.S. economy was largely over. Eventually growth in imports of capital goods and equipment slowed as well. One way to interpret this is that American producers that survived had actually figured out how to stay in the game.

Of course, another strategy to stay competitive was to invest in new equipment that could take out labor costs and help remaining workers produce more. That meant deploying robots and automation.

Estimates about how many American jobs are currently vulnerable to being taken over by machines are all over the map, and of course it's not going to be all losses since lots of other jobs will be created. Researchers at McKinsey & Co. have worked on this problem, and in a report from last summer they wrote that in one scenario more than one in four jobs could be lost in 512 mostly rural U.S. counties by 2030.

But the important thing to understand is that automation is already, and increasingly, all around us.

Earlier this fall, on a tour at a Twin Cities steel fabrication firm, the owners matter-of-factly pointed out a 4,000-watt laser cutter making a steel part that was hard to identify, other than that it had to be 20 feet long with a lot of angles and holes. A finished piece got taken off the machine and stacked as the machine started making the next one, all of it running by itself. The machine gets set up to run for up to 72 hours.

In China, there could be hundreds of firms with the capability to make that part. Yet this work was being done here, to precise specifications, at a price that obviously represents good value for the customer.

But no worker ever touches it. • 612-673-4302