The group further warned that the tariffs will not result in bringing manufacturing back to the U.S., as Trump has promised, because they erase the certainty that businesses require in order to invest in sourcing changes.

To expand on that, Trump hasn't built any kind of a durable consensus even with the broader Congressional GOP, much less the Democrats, which you'd expect if you wanted the trade environment to be predictable and lasting. He's operating under (questionable) emergency authority from the Executive Branch alone. That might work if his goal is to extract a bunch of taxes out of poorer Americans via tariffs for four years. But it's probably not a very effective way to restructure global supply chains, to convince countries to make major investments in domestic production.

Nobody wants to dump a ton of money into a factory in the US which is going to be globally uncompetitive, only makes sense to serve a protected domestic market, and would take take many years to make back the investment, and then have the protection go away and be left in the hole.

Howard Lutnick, Trump's Secretary of Commerce, had some quote from the other day where it sounded like he was promoting the idea that companies could just heavily automate production, so workers are just programming and maintaining industrial robots, basically.

https://www.businessinsider.com/howard-lutnick-future-jobs-factories-robots-manufacturing-tariffs-trump-2025-5

Howard Lutnick says the 'great jobs of the future' will be fixing robots in factories

So, what would these near-future human workers be doing in factories? Lutnick said in an interview with CNBC on Tuesday that the United States should train people to be technicians for these automated machines.

"It's time to train people not to do the jobs of the past, but to do the great jobs of the future," Lutnick said. "You know, this is the new model, where you work in these kind of plants for the rest of your life, and your kids work here, and your grandkids work here."

In a separate CNBC interview on April 3, Lutnick said US factories are "going to see the greatest surge in training for what we call tradecraft — teaching people how to be robotics, mechanics, engineers, and electricians for high-tech factories."

If he's actually serious about that and it's not just political-speak, there are a couple of problems.

First, the people who voted for Trump over manufacturing probably aren't going to be interested in that. If they had the skillset to do industrial automation, they probably would be a lot less concerned about finding a job.

Here's an NPR Planet Money podcast from a decade back.

https://www.npr.org/sections/money/2012/01/13/145039131/the-transformation-of-american-factory-jobs-in-one-company

Young is the perfect model of the new factory worker. He has an encyclopedic knowledge of metals and microscopes, gauges and plugs. He works on the team that makes fuel injectors, which require precision engineering. At the heart of the assembly process is an automated machine run by a computer process known as CNC.

"When I came here 20 years ago, we didn't have CNC equipment," he says. "It was more of the hammer and screwdriver fix, to where now it's all finesse."

"Now it's all finesse" could be the motto of American manufacturing today. In factories around the country, manufacturing is becoming a high-tech, high-precision business. And not everyone has the finesse to run a CNC machine.

I can read, I've had some computer classes, and I have a Bachelor of Arts degree. But when I asked Ralph's boss, Tony Scalzitti, if he would hire me and train me on the job, his answer surprised me.

"No," he said. "The risk of having you being able to come up to speed with training would be a risk I wouldn't be willing to take."

To become like Ralph, I'd have to learn the machine's computer language. I'd have to learn the strengths of various metals and their resistance to various blades. And then there's something I don't believe I'd ever be able to achieve: the ability to picture dozens of moving parts in my head. Half the people Tony has trained over the years just never were able to get that skill.

And if you don't get that skill, a mistake on this machine can be catastrophic. All the work that's done here happens on a scale of microns. One micron is four-hundred-thousandths of an inch. A human hair, for example, is 70 microns thick. Here, you cannot be off by one-tenth the thickness of a hair.

"A 7- or 8-micron wrong adjustment in this machine cost us a $25,000 workhead spindle," Young says. "Two seconds, we could lose $25,000."

Madelyn "Maddie" Parlier is more like the old style of worker. She does have a high school diploma, but no further education. She works on a simple machine that seals the the cap of a fuel injector onto the body. All she does is insert two parts and push a button. It requires no discretion, no judgment. There's only one way to run it: the right way.

"It does it for you," Maddie says. "All you do is put the piece in, push the clamps down, and push your finger."

There are a lot of things Ralph knows that Maddie wishes she knew. She wants to know how many microns thick the different parts are. She wants to know the computer language used on the machine she runs. She wants to know all the things that make Ralph's job prospects so much brighter than her own. And until she knows those things, her future is far less certain.

Maddie has a job, I learned, because of some simple math. A machine could easily replace her — a robotic arm could put the parts in and take them out — but it would cost around $100,000. Maddie makes a lot less than that, and, for now, the math is in her favor.

But if the price of a robotic arm goes down, or a factory in China learns how to make that part for a lot less, Maddie's job is at risk. Simple calculations like that have cost around 5 million factory workers their jobs over the past decade.

If you voted for Trump because you were thinking that you'd get a lot of low-skill assembly line jobs that don't require a ton of education and training, and you wanted them to be high-pay, you're probably not getting what you want out of this arrangement.

Second, there are some other issues, like environmental regulations, that don't help the US be competitive even if a process is automated.

Third, my understanding is that with tariffs, at least some, if not all goods, it's the final stage in processing that matters for determining origin. So all one really needs to do is move legally-minimum transformative final stage out of China to some other country that has lower US tariffs, leave the inputs in China, and do the last step in that other country. Do no more than is necessary to keep this from being a re-export.

Fourth, even if you (could) change the international system of trade and tracked all the inputs, then there's still little reason to do labor-intensive manufacturing in the US. You might take it out of China, but you'd do it in Vietnam or Indonesia or whatever, somewhere where you aren't going to be left with an un-economically-viable factory if the tariffs on China go away or are relaxed. Avoiding that would require taking the US to autarky, which would really dick up its economy a la North Korea.