The trade shocks should not be handled with tariffs or restricting trade. It's domestic policy issue.
So of course the worker who over-invested in job A and was compensated based on that over-investment is going to be worse off in job B when he/she has under-invested in the skills needed for job B.
The fields that are best for workers today are those that have built in required continuing education, or fields in which rapid change is widely acknowledged and understood by the workforce (software engineering, etc.)
This is an important point, and I think plays into the question about tariffs (which, to be clear, I think are probably net negative).
The conventional argument is that free trade is a Kaldor Hicks [0] improvement, not a Pareto improvement. KH means that the sum of benefits and costs is positive, and it would be possible (in principle) to tax the beneficiaries and pay off the people harmed and create a synthetic Pareto outcome.
You could view tariffs as a clumsy attempt to apply the Kaldor-Hicks concept. I personally think there are much better ways, but I think the argument is at least plausible.
[0] https://en.wikipedia.org/wiki/Kaldor%E2%80%93Hicks_efficienc...