Hacker News Reader

Comment by throwawaygh | Hacker News Reader

The most important paragraph in this piece is last one, which plugs a $22.8 billion bill to support domestic semiconductor manufacturing in the USA and suggests the need for new entrants.

Semiconductor manufacturing is truly one of the few remaining manufacturing industries in which the USA has a competitive advantage (albeit quite eroded by TSMC over the past few years).

Especially with US aviation circling the drain, this is an industry we can't afford to lose. Removing semiconductors from US manufacturing statistics makes US manufacturing statistics go from "stagnating" to "extraordinary decline".

And it's not just the manufacturing sector. Semiconductors power some of America's most valuable companies (including but not limited to the FAANGs).

US economic and actual security absolutely depends on a robust supply of world-class semi-conductors manufactured on-shore.

Congress should take the advice in this piece and avoid making the same mistake with Intel that it made with Boeing (piling the bulk of all federal and state support for an entire sector into a single clearly dysfunctional company)

View on HN

US manufacturing has been growing steadily for decades. The US concentrates this in sectors that are high tech, expensive, and profitable. Low margin, low profit manufacturing has been offshored, but manufacturing hasn't been stagnant at all.

As a percentage of GDP, manufacturing has been on a downward trend, but that's because of the growth of service industry.

https://www.macrotrends.net/countries/USA/united-states/manu...

Does that data set factor in inflation? I could't tell from write up.

If not, The 1997 number would be ~$2,200B, after adjusting. That would still only be a slight decline.

"Data are in current U.S. dollars."

It's true, US manufacturing output is not in decline. But it is stagnating, in the sense that growth rates are much slower than they used to be. Look at the average from 2000 to 2010 or from 2010 to 2020 and compare to any decade in the 20th century.

But that's not really the point. The point is, look at those numbers sans semiconductors and you see a pretty scary decline in manufacturing output.

It is totally fair to point out that this has been a strategic choice to focus on profitable industries, and that that's even a good thing. That's sort of the point here: the choice has been made, semiconductors are now a key industry for the US manufacturing sector, and AFAICT we don't have any plans for what would replace semiconductors.

IDK if we have a healthy amount of manufacturing output, but it's pretty clear that losing leadership in semiconductor manufacturing would be quite bad for the USA.

> The point is, look at those numbers sans semiconductors and you see a pretty scary decline in manufacturing output.

Not at all: https://www.nam.org/state-manufacturing-data/2019-united-sta...

Semiconductor manufacturing is a component of the second-largest US manufacturing bucket in 2017, and was only #8 of the top 10 manufacturing buckets for job growth in 2018. It is certainly an important sector for economic and strategic reasons, but you are significantly overstating its importance.

> manufacturing buckets for job growth in 2018... It is certainly an important sector for economic and strategic reasons, but you are significantly overstating its importance.

But I'm not talking about jobs, at all. The US needs a robust manufacturing sector so that it remains economically competitive. Figuring out how to translate that economic competitiveness into material benefits for individuals is a separate problem. Confusing the two is itself a risk to our economic competitiveness. (To make this point is a somewhat hyperbolic way: if the goal is jobs, build a canal through the center of the country. Using spoons.)

Also, I think the manufacturing output numbers alone miss something important from my original post. Losing semiconductor manufacturing as a core competency also threatens the most profitable and important components of the US services sector.

> But I'm not talking about jobs, at all. The US needs a robust manufacturing sector so that it remains economically competitive. Figuring out how to translate that economic competitiveness into material benefits for individuals is a separate problem. Confusing the two is itself a risk to our economic competitiveness. (To make this point is a somewhat hyperbolic way: if the goal is jobs, build a canal through the center of the country. Using spoons.)

Why, then, did you ignore the other chart I pointed out that specifically dealt with the economic value?

> Also, I think the manufacturing output numbers alone miss something important from my original post. Losing semiconductor manufacturing as a core competency also threatens the most profitable and important components of the US services sector.

I agree that losing this competency would be damaging. I disagree that it represents a threat. The threat is the continued consolidation of certain types of semiconductor manufacturing into an ever-shrinking pool of companies. While it would be slightly better if that pool were partly concentrated in the US, the size and diversity of that pool is a much bigger issue than where its members are located.

> Why, then, did you ignore the other chart I pointed out that specifically dealt with the economic value?

The only other chart lists "Computer and electronic products" as the second-largest manufacturing sector, which seems to confirm my assertion that semiconductors have been an important source of growth. It's the second largest category and has had annualized real growth rates 2x+ the rest US manufacturing for decades.

I don't see any breakdown of semiconductors/not-semiconductors for that category, and this is a 2017 snapshot as opposed to longitudinal data, so it's hard to say anything more specific using the data you linked to.

> The threat is the continued consolidation of certain types of semiconductor manufacturing into an ever-shrinking pool of companies. While it would be slightly better if that pool were partly concentrated in the US, the size and diversity of that pool is a much bigger issue than where its members are located.

Sure, I agree. (Caveat of course that geopolitical diversity matters too. N separate SOEs would increase diversity in a way that wouldn't be particularly helpful to the USA, for example.)

In any case, this is actually something I more-or-less agree with, and hence the last sentence of my original post.

The US doesn’t need to competency in commodity manufacturing that’s not going to increase our standard of living. Things get offshored when we have better investments to make with our people and capital here.

The US will not be in the dominant position where it can have such incredibly advantageous trade relationships that it can increase standard of living using only services forever.

It also causes a huge vulnerability.

Free trade has only worked for 240 years, let me know when managed trade works once.

it worked because the cards were heavily stacked toward the US and other first world countries.

as the developing world improves their standard of living, so does the same decrease for people in the developed world, because these developing nations are no longer dependent on the developed one's.

It was always just a question of time with globalization in effect.

personally, i think that this is a good thing though. we in the developed world have been living unsustainable lives for generations, living off of the developing nations.

though thats most definitely not the only reason why everything has been getting harder for the working population. automation, higher efficiency, extreme concentration of wealth and more do their part in the current situation as well.

Our standard of living continues to increase along with the developing world, and free trade deserves much of the credit.

And eventually many countries in the developing world will have a higher power and higher standard of living than us, and our service industries will not have much of a comparative advantage or really much added value, and power relations will change which will completely upend the favourability of trade relations.

Hasn't worked for most third world countries, actually. Most that succeeded was with massive state involvement including trade management.

Somehow, China has managed to create a huge expansion of their middle class with outsourced manufacturing.

>According to a study by consulting firm McKinsey & Company, 76 percent of China’s urban population will be considered middle class by 2022. In 2000, just 4 percent of the urban population was considered middle class.

https://www.businessinsider.com/chinas-middle-class-is-explo...

Meanwhile, in America:

>Fewer millennials in 2019 are middle class compared to baby boomers when they were in their 20s

https://www.marketwatch.com/story/why-the-middle-class-is-sh...

Meh. Middle class Chinese means something very different from middle class American. There’s something true here, but the exact comparison you’re making isn’t quite it.

Also, valuearb is kind of right. Except that the right solution is more efficient onshore manufacturing, not offshoring.

>Middle class Chinese means something very different from middle class American

Of course it does. They do have that extra trillion people to bring the average down.

However, moving the manufacturing jobs from here to there has certainly raised their average standard of living while reducing our own.

No it’s raised both our standards of living. Google the law of comparative advantage.

No, it's raised profit margins for manufacturers who no longer have to pay a living wage, healthcare, benefits, and obey stricter environmental and worker safety regulations.

The unemployed here are not better off.

What unemployed? Unemployment was at record low levels before the pandemic.

Ever been to the rust belt? There is plenty of unemployment in our nation's former manufacturing region.

In addition, a Walmart cashier, Starbucks barista or Uber driver does not have the same standard of living that a UAW member once did.

People are, on average, worse off than the preceding generations, all in the name of higher profit margins for the few.

>Across Western Europe and North America, the middle class is shrinking, according to a new report by the OECD. With each new generation, a smaller share of the population find themselves earning middle incomes.

When the Baby Boomers, who were born between 1943 and 1964, were in their twenties, some 68% were in middle-income households. Only 60% of Millennials, who were born between 1983 and 2002, could say the same at a similar time in their lives.

https://qz.com/1592826/the-middle-class-is-shrinking-generat...

Unemployment statistics and median incomes prove your anecdotes and cherry picked statistics wrong.

Genuinely curious then: where do we actually excel?

The only examples I can think of right now where we are genuinely number one is high priced low volume boutique aerospace (SpaceX and other new space companies, satellites, experimental military aircraft, electric planes, etc.) and of course weapons.

We can manufacture decent cars, but at least half of this is under the direction and management of overseas companies like Toyota and Nissan. US car companies have been on the ropes for decades.

We seem to be losing the bulk of aerospace, generation turbines, electronics, chips, medical devices, power equipment, materials, ... ?

Seeing graphs like this makes me wonder if I am living in the late days of the USSR when I'm sure many optimistic graphs were published by Pravda.

You excel in cultural exports still.

What? Turbines? Materials? Only the US and Europe can actually build functional jet engines and most of the bleeding edge materials. Even semi functional jet engines are still a distant dream for China. Same thing goes for chips, where only Taiwan and the US can manufacture current gen nodes. If anything, the US offloaded commodity manufacturing that is easily reproductible to build a very very high moat around key technologies instead of fighting to the bottom for scraps of old industries.

As for medical devices and tech, the US probably creates more new drugs and treatments than the rest of the world combined. when's the last time china, india or russia made a break through in medecine? When's the last time the US did? Last year?

> Same thing goes for chips, where only Taiwan and the US can manufacture current gen nodes.

This view is rampant throughout this thread and I keep wondering why South Korea is omitted. I thought SK semiconductor tech was comparable the US and Taiwan?

Sorry, yes south korea has an amazing industry. Japan too even if it's not the leader it used to be. but even then, last gen chip design is overwhelmingly coming from the US. Apple, AMD, Qualcomm and Intel have the most performant designs by far, so the moat is still definitely present. Taiwan and SK are arguably better in manufacturing the chips right now though I agree.

I remember that Fujitsu fabricated the SPARC processor, but what other semiconductors is Japan known for?

Pezy SC was kicking around top500 for a while, Sony domminates imaging sensors, likewise for sending niche medical and electrical stuff.

Kioxia (formerly Toshiba Memory) is large semiconductor for NAND chips, its fabs are in Japan.

Samsung's transistors are only slightly behind TSMC, SS8nm probably trades blows with intel 10nm and solidly beats 14+++. And their DRAM and NAND flash is the probably the best, so it is at least comparable.

But beyond SK I cannot think of any other countries with competitive fabs.

Also only a tangent but I really believe that deal where TSMC is going to put a N7 fab in the USA is either going to get restructured into something way smaller or never ever going to happen.

Sony dominates imaging sensors.

> when [was] the last time china, india or russia made a break through in [medicine]?

2015.

https://www.nobelprize.org/prizes/medicine/2015/tu/facts/

(PS. This took less than 30 seconds to find and confirm).

Exactly my point. The USA has breakthroughs every years and very bleeding edge meds coming out every few months. 2015 is not recently at all, relatively speaking.

The prize was awarded for work done in the 1970s, fwiw.

(PS. This took me most of a minute of reading the link you provided).

Sure, any to answer the more general question in the OP:

Russia invented and built and then sold to the USA the RD-180 rocket engines that are used for the first stage of the American Atlas V launch vehicle.

Which the USA will no doubt "reverse engineer" soon :)

The OP needs to get out more often.

I feel weird playing the jingoism game, because it doesn't even vaguely reflect my feelings on the subject, but here we go:

SpaceX, a privately-held American company, is the only source of innovation in rocketry at the present time. Their reusable launch platform plays in its own league; everything else is steam power, Falcon is internal combustion.

That's just a fact. Unlike insisting that China, Russia, India are incapable of innovation more generally, which is risible.

My vague impression, in fact, is that jet turbines are quite a bit more difficult to manufacture than rocket engines, even if the latter are thought of as 'higher tech'.

I'm fairly certain we'll discover, in the next ten years, that China is quite capable of making cutting-edge computer chips, and just hadn't gotten around to it yet.

I've never shared my countrymen's obsession with insisting on being 'the greatest'. This country is great in its own right, along many dimensions; it is, in fact, right at the cutting edge for medicine and CPUs; denigrating the work of other nations only detracts from this.

> My vague impression, in fact, is that jet turbines are quite a bit more difficult to manufacture than rocket engines, even if the latter are thought of as 'higher tech'.

My impression is that “real” rocket engines include a jet engine just to pump fuel. They have many other pieces as well.

Not that a modern high-bypass turbofan is simple, but don’t discount the complexity of rocket science.

Jet engines will not be a distant dream for China for long, neither will chips. China has functional 14nm now and has been building their own EUV machines in expectation of the recent US moves.

It will be the same story for jet engines, Chinese improvement is very rapid. It was the same story for every rising power so far, there is no real reason that China will never be able to build a jet engine, and no reason that they will never be able to run a fab (and indeed, they are accelerating quite fast).

Chinese engine improvement has not been rapid by any measure. China still has to import the majority of high performance fighter engines from Russia (primarily the AL-31). China has been trying to develop the WS-15 as a replacement for almost 30 years, and still relies upon the AL-31 for their premier stealth fighter (J-20).

They have started to deploy the WS-10 Taihang engine for their J-10 fighter, but the performance of this engine is markedly worse than comparable Western engines. Military grade engines are hard!

Jet engines are indeed hard. But with the upcoming tech transfer from Ukraine as well as the slow but steady improvement of the WS-15, I would be incredibly surprised if it took more than 6 years for the WS-15 program to yield a serviceable engine.

Even the US relied on foreign jet engines for a time, although you won't hear about it a lot - for example, an absolutely huge amount of engineers were hired from AVRO in Canada after the development of the superior Iroquois engine (which had by quite a margin the highest nominal thrust as well as the highest) as well as other Orenda engines, and their work was absolutely instrumental to the US becoming superior to the USSR in jet engine technology. I don't know if this is publicly available information, so you might have to trust me on it, but I guarantee you even the US had a 20+ year period where it had to steal engineers and tech in order to jumpstart it's lead, which is exactly what we are seeing with China.

If the WS-15 ends up being serviceable in 6 more years, it will have taken SARI and Xi'an almost 35 years to re-create something marginally better than the current Russian engines, and no where near what the US is delivered routinely by PW and GE. Chinese engines can't approach the thrust of Western engines, nor the reliability.

So say they iron out the kinks in the WS-15 (though odds are they'll settle for the WS-10). They then have a roughly 1990s era engine, when the US will be deploying Adaptive Cycle Engines etc.

Sure, but this always how it works. The first high-quality engine will take 35 years, the next 15. The important thing here is to develop the capability to manufacture high-quality single crystal components.

As far as adaptive cycle engines, China is also already researching them. They also have a lead in supersonic combustion as of today.

Discounting the competition is sure path to defeat. I am not in the semiconductor industry, while everyone is chasing a node sizes down to zero, I think the real breakthroughs will be in reducing the production costs to zero.

Music, movies, software, high speed pizza delivery

Underrated response.

As true today as when it was written!

Computer hacker, pizza deliverator, and greatest swordsman in the world, at your service.

There's Tesla as an American car company, which is arguably the most fearsome in the world right now. Kind of interesting how US manufacturing lead in several industries is tent-poled by one immigrant, Elon Musk. SpaceX doesn't just manufacture rockets, they are also now the largest satellite manufacturer in the world.

We're also really good at manufacturing houses and food. Besides cars, houses, and food actually representing a good chunk of physical things we buy, more of our spending is moving towards software/information in which we have a crazy lead.

I see it as manufacturing was displaced by an even more value adding industry. Though certainly it isn't good for our national security to be relying so much on imports for physical goods.

I think Tesla is a tech company and not a car company. The amount of technologies that goes into a Tesla car is unreal and Tesla abilities to use the sensors and chip on car and update a car via software is rather fearsome.

Why not both?

Tesla does, after all, manufacture cars — it doesn’t just write software that goes into cars made by others.

I’m not really seeing that - a Tesla Model S seems to me like a much simpler device than a Mercedes S Class. Talking about overall complexity, not just software.

Why Mercedes electric cars are not Tesla quality then? I mean the drive, the interior is better in Mercedes.

Actually, we're really not that great at manufacturing houses. We have too few where they are needed, the quality standards both in terms of energy efficiency, design and construction tend to be low, and consumer satisfaction is reportedly dropping.

Too few where needed is a product of laws. Most people have no idea what makes a quality house and so pick the wrong things to focus on. Codes don't allow junk houses, and energy efficiency is pretty high - what counts as a passive house in some areas is too poorly insulated for the northern US code.

> The US concentrates this in sectors that are high tech, expensive, and profitable. Low margin, low profit manufacturing has been offshored, but manufacturing hasn't been stagnant at all.

Slight refinement: it’s a barbell: production of low value goods that are not worth shipping has remained, e.g. paper. But that’s just damning.

> US manufacturing has been growing steadily for decades.

You need to research a bit more what is behind these digits.

A lot of funny accounting where companies which haven't manufactured literally anything on US soil for decades somehow manage to accrue billions in "manufacturing output"

>As a percentage of GDP, manufacturing has been on a downward trend, but that's because of the growth of service industry.

Or is it the other way around? Manufacturing is declining, and services are growing because they're the only jobs left for former manufacturing employees?

Interesting manufacturing output comparison list, but why is China not included in it?

IMO your last statement is kind of ridiculous. Intel is literally the only company that does high volume wafer manufacturing in the US at anything approaching leading edge geometries. Are there any other American companies the government could even realistically support? (The only other US company that has any modern wafer fabs in the US is Micron and the only really have R&D and legacy long-life products being manufactured in the US.) Maybe if this program occurred 20 years ago you might have had another option. (And you would have basically looked like an idiot as for the next 10 years Intel pretty much made all the right decisions in manufacturing while everyone else in the US went down dead ends.)

Are there any other American companies the government could even realistically support?

Micron is perhaps a candidate, though not sure it meets all your criteria.

Unlike Intel, Micron has practically outsourced most of their manufacturing to Singapore, Taiwan and Japan. Their only production fabs in the US are in Utah and Virginia both of which are mostly focused on niche or older products. More importantly, they only produce memory. Logic and memory processes are very different.

does it have to high volume ? lots of semi-conductor companies in the US, some of them flying under the radar. the us stimulus will help these companies invest more capital, talent. hell, I will say any country needs to be manufacturing and designing their own silicon. and like said by ben, not all silicon has to be for pc's.

None of them are on cutting edge processes.

Doesn't Samsung have a semiconductor fab in Austin? I guess they don't count as a US company ...

I think they dropped the whole semiconductor team a few months ago iirc.

That is their custom CPU design team, nothing to do whatsoever with this. (and which wasn't very good in practice). Samsung is committed to their semiconductor business.

Yes, I agree with you. It's a huge problem for the US right now, both with Intel and with Boeing.

My last statement is ridiculous, but also TFA is correct in its diagnosis of Intel.

So... where does that leave us?

I don't pretend to have good answers to your question. There are some smaller shops nowhere near the scale of Intel. Some foreign-owned companies have fabs in the USA (including TSMC). Also, Intel is made of people who might be interested in striking out on their own.

But, yeah. Building things is hard and expensive and risky.

There are many other aerospace companies that are in principle capable of designing high performance airplanes in the US. There are literally no other companies in the US with modern semiconductor manufacturing expertise (in logic, not memory, GF doesn't count). The level of capital expenditures and technological expertise and IP required to manufacture leading edge semiconductors is staggering. It's unlikely you could even start a new company that could manufacture leading edge semiconductors for the entire $23 billion that is in that bill. (They'd also probably get sued into oblivion by existing manufacturers with basically zero patent protection.) Just look at what happened to Globalfoundries. It was given effectively free capital for close to a decade, a not totally instate starting point of starting from nothing AMD's old fabs, and in the end gave up.

The "diagnosis" of Intel is that it threw away the chance of higher volume for lower margin which gave TSMC a bunch of money to catch up to them. (I think that's a very hindsight is 20/20 kind of view too...I don't think anyone thought that the iPhone would become larger than the entire PC market when it came out...and at those volumes Apple would eventually realize anyway that it wouldn't make sense to keep paying Intel to make chips for them.) I don't see how a start up is going to capture that volume without some sort of pseudo-captive market to grow with (which is basically how the various companies in China are progressing...).

Global Foundries may be stuck at 14nm, but they'd be the most natural successor to next-generation if their financials could be improved a bit. Its a shame that the New York based semiconductor fab has fallen behind.

They've spun down so much of their R&D, that it'd be a real uphill battle to get going again. There's been so much brain drain.

And the wall at 14nm is them not investing in EUV, so they'd have even more work versus say the 14nm/28nm transition.

Global Foundries is hardly an "American" company...it's owned by the Emirate of Abu Dhabi.

> IMO your last statement is kind of ridiculous. Intel is literally the only company that does high volume wafer manufacturing in the US at anything approaching leading edge geometries. Are there any other American companies the government could even realistically support?

Well yes, but... you could hire a bunch of foreigners to run things, as that's where the current experience mostly is, as you point out. But that experience includes use of semi fab equipment, essentially all of which is built in the USA.

That said, I'm not a fan of industrial policy; not sure what results were generated by the megabucks spent on Sematech (and I worked for Sematech sibling/neighbor MCC in its day too).

What about IBM? They have a few foundries in the US (and Canada), and used to produce the processors for Xbox 360 and PS3, which were both pretty advanced. And these were far from their most advanced chips, the stuff they built for their Z-series is truly impressive.

I haven't followed them much since, but 10 years ago, they were doing good stuff when I worked there.

IBM's fabs are acquired by GlobalFoundries and GF retired to develop leading process.

Thus continuing IBM's very sad tendency to sell every part of itself that's interesting.

I didn't know, but I'm not surprised.

Semiconductor manufacturing is not what average Joe thinks when they talk manufacturing. It’s highly localized and requires not that many, but very highly educated workers.

While I highly support US investing there more, it’s not something that gives US blue collar jobs in the rural America. And push for more manufacturing in USA is about that.

For better or for worse, the American model of working in the same paper mill/steel factory/automobile plant/coal mine that your father and grandfather worked at is over. Automation, commodification, and globalization means that the profit margins in these fields are in free-fall and the ability to have the sorts of wages that these professions used to have where workers could be paid sufficient to support their families on a single salary are long gone and they are not coming back.

The US should be focusing on decoupling the need to be able to survive (food, housing, healthcare) from holding a job instead of pushing for bringing back these jobs that are not likely to ever come back.

That's true, though it's interesting that many companies still talk and act like employees should be "loyal". They don't seem to understand how that worked.

I cannot fathom being loyal to a company where they expect two weeks notice from me (at a bare minimum) if I'm leaving but where they don't have to give me two minutes notice before terminating me.

I wouldn't conclude too much about the steady-state price of manufacturing labor based on the present global market.

At the moment regional differences in the price of manufacturing labor are so large that corporations are basically labor market arbitrageurs. But arbitrage opportunities are by their nature kind of ephemeral, ending when the relevant markets stop mis-pricing things.

Your POV seems to be that labor is correctly priced in the current cheapest labor market (the "Race to the Bottom" hypothesis) but I don't know that this is automatically true. It's at least possible that one can only arbitrage the price of labor in developing markets for so long until they develop and drive up the local price of labor, and until they are all developed.

"The Jobs Ain't Comin' Back" is an appealing position for professional automators like us, because it lays the credit for manufacturing labor market dynamics at our own self-important feet, when in actuality it's shaped largely by brutally exploitative and probably transient (on the scale of decades, though) labor-price arbitrage. In particular the current manufacturing labor market artificially rewards centralization because the profit made by labor-price arbitrage is so large compared to shipping costs.

I don't think it's a foregone conclusion that manufacturing is either Sweatshops or Robots in the long run.

The "Race to the bottom" will be ongoing for a long time. The nice thing about manufacturing for a middle-income economy is precisely its role as a stepping point to far higher productivity in services and high-skill sectors - this means that many developing economies will naturally tend to become manufacturing-intensive over time, even beyond the East- and South-East Asian countries that most observers focus on at the moment. The U.S. is supposed to be well beyond that point already.

Problem is most people are past that. China and India do have a few more poor people to take cheap low skill jobs, but overall cost of labor is going up in both countries and manufacturers are starting to look elsewhere. Problem is that elsewhere is countries with much less population so they will run out.

It probably won't be in my lifetime, but my kids need to plan for the day when there isn't a cheap labor source waiting to tap. It will be a good problem to have (think about all the engineers making life better), but it will be a problem.

I think its a bit like a game of hot potato, where the final manufacturing location will basically end up being a global power house for the next couple centuries.

Which at the moment looks to be china, despite textiles trying to move to lower cost countries, the magnetic force of ongoing factory automation in china might pull them back. The US deindustrilized to early for any of these really to come back without some seriously focused "billionaire" class investors putting their money first (aka Musk).

How much can we learn from each other? I work at a factory making things, we have locations around the world because our final products don't fit in a shipping container so local factories out compete global ones. That means that we have manufacturing experts scattered all over the world. We buy parts from all over and share knowledge. (expect China, we are careful what gets shared with China)

Spending on external training is $83bln [1]. Internal training program spend is an order of magnitude higher. Let alone salaries.

That is a lot of spending that could be transitioned to robotics if humans become too costly.

1: https://trainingmag.com/trgmag-article/2019-training-industr...

> It's at least possible that one can only arbitrage the price of labor in developing markets for so long until they develop and drive up the local price of labor, and until they are all developed.

This is a fairly optimistic point of view, IMO. What I think is much more likely to happen is the following cycle:

1. Manufacturing comes to a new place due to low wages (relative to the cost of labor in the previous place), lax environmental regulations, low tax rate for corporate profits, and very low worker safety regulations.

2. This goes well for 20-40 years, but as things typically do, environmental regulations are put into place by the local government, the tax rate rises to pay for the externalities of the low wages, and workers organize, agitate, and win higher wages, more safety regulations, and worker protections. All of this raises the cost of labor.

3. Manufacturers move to a new location because it is much cheaper. This causes people in the old location to lose their jobs, become impoverished, and craters their economy to the point where the government, implored by all of these laid off and angry workers, need to do something to help the situation. This results in a lowering of wages, rolling back of environmental regulations, lowering the corporate tax rate, and repeal of the safety regulations and worker protections.

4. Twenty years later the manufacturers come back to the country because people have become so desperate for jobs that moving back to the country is relatively cheaper compared to the country that manufacturing just moved to.

Just view the policy decisions with respect to manufacturing that the leaders of the United States and Mexico have proposed to "bring back" these blue collar jobs. It all follows this playbook.

If things are better than before after step 4 you still have damped oscillation. I think the distillation of your hypothesis is this: labor (and environmental externality) markets change slowly enough relative to the amortization schedule of a factory that perpetual arbitrage between them is possible.

To the extent that this is true, one cause is that governments (with the notable exception of China) handicap themselves by honoring agreements, which they absolutely do not have to do absent a bigger, badder government to make them. A savvy, Machiavellian US government policy would:

1. Promise corporations the world

2. When factories are built, expensive machinery shipped in, and personnel trained, renege and hike taxes, but not so much as to force a write-off

3. When foreign governments make it difficult for US corporations to operate factories abroad and they ask for diplomatic support, shrug and tell them play stupid games, win stupid prizes

4. Do not apply the above policy for US corporations doing natural resource extraction abroad.

(i.e. act more like China)

No, it’s not. The push for maintaining manufacturing capacity is not about jobs though that’s been one way to try to frame it for a myopic populace. I thought more people would figure that out given PPE shortages and the inability to produce enough medications for our citizens but people need to simplify and straw man the issue. It’s more importantly a national security and strategic consideration. But many cannot see that far for some reason.

Yes, this is the real reason. What if a foreign power controls something you desperately need? It's not going to work out too well for your country. In that case, even as there is comparative advantage in global trade, one must still control certain industries domestically.

The US followed the advice of the business schools long ago and outsourced all its core competencies. The US' loss of technical and manufacturing expertise is going to be extremely hard to reverse, and reversing it is going to require a rejection of financial capitalist economic ideology and things like "maximizing shareholder value" as the prime directive.

Unfortunately this election gives us a choice between an utterly unacceptable racist crackpot and a bland candidate who belongs to the old guard who still believes that you can have an economy based on nothing. That doesn't make me optimistic.

Don't know why this is getting downvoted. Tone aside, it's a fairly accurate assessment.

I'd wager it's because he criticizes Biden (too), but come on, people...

I can't say for certain, but I believe that statement is being downvoted because the first paragraph is factually inaccurate in each sentence:

> The US followed the advice of the business schools long ago and outsourced all its core competencies.

Bashing metal and low-tech widget-forming has never really been a core competency of the US. It was useful to keep people employed for a 50 years or so but for the past century the US seems to have aimed for technological synthesis and increasingly high-tech production as a core strength.

> The US' loss of technical and manufacturing expertise is going to be extremely hard to reverse, and reversing it is going to require a rejection of financial capitalist economic ideology and things like "maximizing shareholder value" as the prime directive.

It is unnecessary to reverse something which was not lost in the first place. As others have repeatedly pointed out all over the comments on this article the US has continued to climb up the manufacturing food chain and dominates the top-end. It may employ fewer people, but the productivity and economic benefit of the US manufacturing sector has simply not declined at all and therefore no 'rejection of financial capitalist economic ideology' will be required.

> Bashing metal and low-tech widget-forming has never really been a core competency of the US.

If you think the far east only dominates low-tech widget forming, you should look into where the iPhone and Macbook are manufactured :)

I know where they are assembled. I also know where the displays are actually manufactured, where the RAM comes from, and where the CPU is fabbed, and where the various radio chips are sourced from. I also know where the software is written that makes it all work.

I also know that the iPhone and Macbook are not particularly difficult pieces of engineering and that the tech required to make the shiny bit of tech you see in the store is the bit of the iceberg beneath the waterline that represents a lot of the value in what we are talking about.

> I also know that the iPhone and Macbook are not particularly difficult pieces of engineering

https://en.wikipedia.org/wiki/No_true_Scotsman

It is not a No True Scotsman fallacy, it is simply that the tech and machines/processes used to create the components are far more impressive and represent a more important tech base than the actual output from those processes.

If we want to trade vaguely related aphorisms: “Amateurs talk tactics, professionals talk logistics.” The tip of the spear is less impressive than the logistical tail that makes it possible.

This. I've always been most fascinated by industrial machinery and process engineering more so than the end result of it. It's that piece that represents the Golden egg of industry; and like it or not, the capability to keep the people of your country employed making widget X, receiving paychecks, etc is valuable. Incredibly so. Possibly even more valuable than the money it makes as the know-how it takes to have formulated that process stand-up is nearly impossible to codify, and closer to a luck based emergent outcome than anything you can plan for. Similar, if you'll excuse the long-jump analogy to how all the enzymatic processes our cells must undergo to live and thrive are all based on a fundamentally chaotic process; diffusion.

> It is unnecessary to reverse something which was not lost in the first place. As others have repeatedly pointed out all over the comments on this article the US has continued to climb up the manufacturing food chain and dominates the top-end. It may employ fewer people, but the productivity and economic benefit of the US manufacturing sector has simply not declined at all and therefore no 'rejection of financial capitalist economic ideology' will be required.

For many years, I have observed that folks in the Bay Area (if I might use that as the shorthand for those at the "top-end" of the technical food chain) tend to dismiss and deride the work further down the food chain, whether that work is "bashing metal" or piecing together business applications.

There are two truths I believe about this work that some do not acknowledge:

1. People who bash metal or piece together business applications have expertise and knowledge that folks higher up the food chain do not have.

Most people who make their living high up the food chain did it by proving they could rise up above those further down the food chain, so it makes sense to believe that this work is beneath them. But the domains are different, and the choice to live high up the food chain necessarily means that one chooses to miss out on experiences and knowledge that those further down the food chain will have.

2. The folks higher up the technical food chain are dependent on those lower on the food chain.

If I might pick up an imperfect analogy: Sharks are higher up the food chain than fish. Fish get bossed around by sharks. But sharks are dependent on fish to survive, not the other way around.

> It may employ fewer people, but the productivity and economic benefit of the US manufacturing sector has simply not declined at all...

Unfortunately wanting something to be true, and it actually being true are not the same thing.

Even if a whole bunch of your peers say the same thing, it still doesn’t make it true.

I’m not really sure what to say... please go and do some actual research on this topic.

I don't know about the rest of your comment, but it starts with a false statement.

> the advice of the business schools long ago and outsourced all its core competencies.

Business schools recommend the opposite. Outsource everything OTHER THAN core competencies. Core competencies are your competitive advantage; keep them in house.

> Outsource everything OTHER THAN core competencies.

The core competency of the US today is finance. We've been actively outsourcing everything else we can.

> The core competency of the US today is finance.

Since when? It didn't look very competent in 2008...

We're not even competent anymore at our core competencies…

I've looked and it seems that our actual core competency (the thing we do better than others) is raise soybeans, beef, and chicken. https://money.cnn.com/2018/03/07/news/economy/top-us-exports... for example.

movies, music, and microcode.

and let's not forget disruptive pizza delivery services.

Government picking winners is corrupting and bad for our economy. It speaks to the intellectual bankruptcy of the supporters of these handouts that they can only look backwards and say we need to fund dinosaurs to steal engineers from new industries of the future and put them to work on do overs for Intel and Boeing.

It's only picking winners in the sense that we need to empower players in our market to be able to build here. It's creating the opportunity/playing field for a winner ton exist.

Because "winners" in this sense right now just means "who can build it cheaper" -- and there's no intellectual bankruptcy there. The answer is pretty much always "K, build it cheaper? Build it somewhere else then."

The market is saying the US is better doing other things.

It would be as if in the late 70s deciding we had to subsidize supercomputer and mainframe computer development, to hire all these talented engineers who were wasting time building and programming kit computers and “pc”s that could never be as important to our economy, and our “technological leadership”.

We can’t industrial plan the future.

Right but the market doesn’t have any foresight to say “an adversary is going to take advantage of this dependence the market has created here and strangle us with it.” The “market” like water just flows toward “it’s cheaper to do it this way.”

The market doesn’t care if US democracy triumphs over (or even survives) Chinese state managed capitalism. But I think a lot of us have skin in the game on this one.

China has long recognized the strategic importance of this kind of thinking, see the massive subsidies they provide many of their industries to create the very dependencies you’re crediting as a natural market efficiency.

They aren’t strangling us, they are enriching us. There are no dependencies, just subsidies to give us better deals.

It sure does feel like a good deal for now, comrade.

I don't disagree with this sentiment - we do tend to overly fund existing industry such as air, space, energy.

As far as new investments though, China arguably picks their winners and has learned valuable lessons in boosting manufacturing from garage to factory. Their lead in solar manufacturing was astronomical.

The US have also done this, look at how Boeing, HP, Intel, the Auto industry and others had their start. Lots and lots of loans and grants.

If anything our policies are overly conservative and careful with new investment, investing heavily in the "idea factory" models surrounding higher education and subsidiaries. Fortunately that's also led to a large and mostly competent investment community that is willing to bet the farm on new companies.

Comparing the industrial & investment growth to industry in Europe and the ME has informed my view that we do a pretty decent job. Corruption is there, but it isn't the norm that I've see in other parts of the world.

This only works when you are playing a fair game.

When China plays clear favorites with its companies, the US is forced to compete in tis unfair market.

This is especially relevant for industries that require such an extreme amount of capital to be bootstrapped, that it is essentially impossible for a competitor arise in the market fair and square.

When China picks winners, it benefits the US because those subsidies lower our costs of goods.

Reply on news.ycombinator.com