The issue is that labor productivity (level of tech) in American mfg hasn't broadly increased at the rate we'd need to manufacture many things at reasonable prices for the American consumer. This makes Baumol's cost disease a huge issue: https://en.wikipedia.org/wiki/Baumol_effect You can see this manifest in healthcare as one of the most egregious examples; the top cause of margin pressure for hospitals is labor: https://www.hfma.org/press-releases/health-systems-near-thei....
While we can still manufacture things that require comparatively high levels of skill, technology, and capex, it's never again (absent a depression greatly outstripping the 1930s) going to be profitable to pay American workers to make t-shirts rather than Bangladeshis.
There's a good argument to be made that a combination of outsourcing and illegal labor caused problems by suppressing investment in tech and automation for thirty years plus, and there are certain things we probably should make here. But ultimately the stuff we actually need to manufacture are things core to sustaining life and the military. Medical supplies, weapons, food, oil, metals, chemicals, etc.
We can, with time and good industrial policy, bring back some manufacturing. That would be a case of short-term pain for long-term benefit. But even then, that's true only insofar as we give people a shot to actually buy American. Moonshot investments in roboticization and industrial automation for a few years would really make this easier, along with using the huge amount of post-HS education dollars we spend to focus on training skilled engineers to implement this sort of thing, along with things like skilled machinists. But these tariffs don't really give American companies a shot.
We cannot, with any reasonably-good outcome, bring back manufacturing jobs. That midwest factory worker is never going to be paid $30/hour plus pension/retirement contributions, good medical, etc. to make regular, el cheapo consumer goods.