> "What's gone isn't coming back until it's cheaper to build it in the US." Yes and no. Yes, at a high level that's true. But you seem to be making an assump

> "What's gone isn't coming back until it's cheaper to build it in the US." Yes and no. Yes, at a high level that's true. But you seem to be making an assumption that human labor costs will remain an important component and that the "it" being manufactured will remain constant. Consider that automation may not even be cheaper on a 'net screws turned per hour' basis than cheap humans. But quality and consistency matters. Speed of line changes matters. Speed of retraining matters. And, further, current designs are limited to what can be assembled cheaply (that is: reasonably quickly/safely/accurately) by two human hands and ten human fingers, with a human attached (who hopefully doesn't require slow or expensive tools/safety equipment/etc). Automation removes those concerns entirely. Caustic compounds could be mixed on the spot, fixation technologies that are infeasible for human dexterity are possible, multiple steps could be completed in parallel, etc. So keep in mind that while it may be sufficiently cheaper to have, say, Brazilians turning screws than Chinese, such that iPads could some day be cheaper to build in Brazil - it's entirely likely that before that happens, robots will be building newer iPad designs, more cheaply than could ever be feasibly constructed by humans even at a zero wage.