Some of the comments here are along the lines of "Well, they should go get a better job!"
Hint, folks: not everyone can get a professional job paying six figures a year. And even if they could, the moment they did, your own wages would drop substantially.
Having sympathy isn't going to fix things, though.
How can we get people in this country to be capable of doing jobs that people willing to work for next to nothing in other countries can't do just as capably? If we can retrain a bunch of them to be great software engineers, our own wages might drop a bit from their absurdly high peaks, but currently the demand for people able to write good software is underserved, and it could be much larger, as software enters other facets of life and needs to be maintained and rewritten to deal with changing demands.
Having more engineers would enable a greater range of software to be created.
Establishing sympathy is a good first step, otherwise you end up discussing with people whether there's anything to be concerned about =)
The thing is, we'll always need warehouse workers in one form or another. One form draws from the relatively less formally educated and puts them to work; if we go that route, it's morally imperative that we do our best to make sure they work in a workplace with dignity and a voice. Education of any particular worker won't help that, as s/he'll just be replaced by another.
The other form, of course, is to move from human to machine labor. In the long term this is the best route. (The question is then, though, what of the people who are deskilled and unemployable?)
I'm pretty sure machine labor is the inevitable endgame for warehouse labor, and I think it's likely coming sooner than most other manual labor that's reasonable to automate, since it's a controlled, structured environment and there's no need to interface with humans.
For the unemployable, I think we'll need a combination of super-cheap scalable teaching methods and a more socialist system. Hopefully we'll be able to transition relatively smoothly, if that's the case.
You sure about that? I've seen a lot of supermarkets going self-checkout lately. Sure, it's not out-sourcing in the usual sense; but it is replacing workers with something cheaper (equipment which was probably manufactured in another country).
Someone has to update our nation's crumbling infrastructure.
True, but the number of workers you need depends on the construction techniques you use. If manpower is expensive, you buy prefabricated components and only do the final assembly on site.
Someone has to change the IV bags and push around the stretchers.
True, but technology and a willingness to spend more on equipment can reduce the number of people you need to do this. (e.g., "smart" IV bags which alert staff when they need to be changed, rather than having nurses walk around checking the bags.)
Someone has to teach future generations about math and science and history and puberty.
Khan Academy.
Electricians and plumbers, they're definitely going to be in demand.
True, but the more expensive they are, the greater the pressure will be to create hot water heaters which don't need to be replaced every five years.
Reports of the death of the American blue collar job are greatly exaggerated.
There are a lot of job categories which will never be eliminated entirely, sure -- but most of them can still be dramatically downsized via the application of technology and sufficient quantities of money.
Nobody is so essential to the continuation of civilization that they can set their own wages. There's always going to be a point where people will say "you know, we've got a cheaper option".
"If manpower is expensive, you buy prefabricated components and only do the final assembly on site."
Not infrastructure, but that's exactly what Boeing has done for years, assemble airplanes out of prefabbed subassemblies.
When Boeing gets a contract to sell planes to, say, China or Japan, part of the deal is often that the subassemblies (wings, body sections, whatever) are made in that country, shipped to the US and assembled by Boeing.
Airplanes are a more controlled and regular construction environment than highways, but I don't see why it won't become more common. Residential and commercial buildings too, I suppose. Suburban houses are already cookie cutter these days.
2. Construction - There is a glut of unskilled construction labor right now, though, coming off the housing boom. Could conceivably be automated, but it's probably a lot farther off.
3. Healthcare - Hard to automate, but the things you mention seem less so than construction.
4. Teaching - Scalable teaching methods - the few best professors can teach a vast number of students relative to what used to be possible. Khan Academy and Stanford's upcoming online classes seem like early stabs at what I think will eventually become the norm.
5. Maintenance work - you're probably right, this is very case-by-case, and requires a lot of training to be able to deal with all contingencies.
Overall, I don't think the future of American blue collar labor's supply vs. demand is very bright at all, unless there's a big uptick in demand for hard-to-automate fields.
EDIT: And when I say automation, I usually mean single-purpose robots or devices that handle the busywork parts of things, like self-checkout machines. cperciva covers this pretty well in his healthcare section. General AI is obviously very hard, and good construction bots would probably need something like that if they were to be able to deal with problems and exceptions on their own.
Hint, folks: not everyone can get a professional job paying six figures a year. And even if they could, the moment they did, your own wages would drop substantially.
Have some sympathy.