This is the part that makes absolutely no sense to me. The farther you move up this value chain, the fewer and fewer people you can find that can actually do those jobs. They require advanced degrees and highly technical specialization and experience. Not everyone can do that. Unless you have a parallel plan to, perhaps magically, mint more and more PhDs every year what is everyone else supposed to do who was just a "low level unskilled dirty manufacturing" worker?

>American business people were happy to...

You're sure right about that! I think what people are questioning is if, as a state and society, that was a long-term good strategic choice. For the bottomline it absolutely was - for those businessmen and their company's earnings reports.

Okay, odds are good that as a manufacturing layperson, I'm just a dummy for not seeing this, but the "climb up the value ladder" implementation in American business has never made sense to me. Especially with automation.

Here is how I see it as a value investor, so please tell me where I'm wrong.

"Climb up the value ladder, and let someone else occupy all those pesky 'low-value' rungs."

That's what I see many American management subscribe to. The part I didn't understand: those rungs are your "innovation feedstock", the raw material if you will, where you find your next innovation iterations. It's like saying, "meh, I'll let someone else deal with those pesky metallurgy details, and churn out jet turbines with the latest in metallurgy advances". It's a belief that innovation is a "home-run", "big bang" phenomena, and not Edison's "99% perspiration", daily grind-it-out, fight-for-every-inch iteration, that sometimes accretes into an edifice taken as a whole that is only much later celebrated as an "overnight success". If I rely upon general metallurgy advances to inform my design team of what's metallurgic-ally possible, then aren't my competitors availed of the exact same advances at the same time? Whereas, if I have my engineers, materials scientists, and "dirty" forging plant workers come together, asking themselves, "if we need a turbine that has these qualities, it must have a metal with those properties, then what will it take to come up with the materials science and forging techniques to reach that design goal", that's a trade secret that no trivial or even modest amount of reverse engineering is going to ever reveal.

It certainly has been this way for me personally in software and coding, so maybe I'm mistakenly ascribing and generalizing that experience with high-complexity endeavors to manufacturing. I certainly concede that for investors, the numbers show innovation is best modeled as a home-run, big-bang phenomena, and it doesn't make sense to invest in accretion. That's how the entire VC business model is presented, at least. But startups are a tiny fraction of the overall business universe. The successful stories in startups are the highest growth for sure, and the most exciting-shiny-chrome, but for index-investing boring-but-certain-profits level moats of defensible businesses, I see more opportunities in a model where I vertically integrate the value ladder, so the value ladder strategy is instead stated as:

"Climb up the value ladder, and vertically-integrate the entire value chain to capture all accretion benefits to a value equation greater than the sum of all its parts."

The above was my thinking since the 80's when I first heard about the value ladder approach of abandoning "low-value" rungs, when automation wasn't the big deal it is today. With automation now on everyone's mind, now I'm left wondering if the accretion benefits aren't amplified many-fold: exactly who is positioned with the expertise and knowledge base to actually automate, the organizations doing the manufacturing, or the ones who are focused on design (including "high-end" manufacturing), management, sales, and marketing (DMSM)? The growth during the onrushing automation ramp-up is likely in the automation and not the DMSM activities. If I am only junior-grade with my DMSM but I have figured out automation, then my COGS drops to the floor and I have a window of time during which I plow my profits into learning and getting my DMSM story straight to out-compete conventional DMSM shops, and take over their market niches. That window can be a hell of a long time if that automation is nearly perfectly opaque to outsiders, because it isn't "just" code, but data that drives the automation around machine learning, data that you can't get unless you actually were doing the original manufacturing at scale (ruling out replicating the automation by building a pilot plant, which only yields automation-delivering-unacceptable-quality).