I appreciate the problem Western economies have with Chinese subsidies but, back 70 years ago, we had the same problem with Japanese technology. The Japanese had no problem setting up factories to produce goods for export with Made in Britain on them because they had named the location of the factories as Britain. We solved this problem by becoming as efficient as the Japanese by allowing them to open factories in Britain to produce the goods we needed. The French had eleven companies producing washing machines when the Common Market opened the borders and import tariffs were abolished. Most of the French factories closed down due to the Italian washing machines being better and cheaper. The same happened in Britain when the restrictions on Japanese motor cars were removed.
A worse problem arose at the start of the 20th century when Germany started a race to build a bigger warship fleet than the Royal Navy. This eventually led to WW1. Given the nature of politics and that China is a dictatorship bent on taking revenge for all the humiliations inflicted on them before 1949, we need to find a way to accommodate their annoyance. No good quoting Machiavelli about making them our friends instead of treating them as our enemies. China looks on weakness as an opportunity to exploit. My advice is to continue to outcompete them in technology using government subsidies to the same extent that China does. That will require cooperation between the countries of the West, in particular the Western members of the G7. It has to start with the Americans who are richer and better placed than Europeans to exploit the advantages of mass production.
Isn't one reason why East Asia has become the world's factory because it has so much of the world's population both to work in said industry, and to be a market for its products?
That and the fact that they: a) Guaranteed cheap & compliant labor; b) Refused to add externalized costs (like environmental degradation) into their pricing.
Western, especially American, manufacturers fell utterly for it--even at the cost of giving 50% ownership to their Chinese subsidiaries, and total technology transfer. They then pretended to be shocked when their own Chinese partners matured to the point where they were able to outcompete and bankrupt them on American soil.
Wall Street got us into this mess: the urge to profit tends to trump patriotism. Which is why China *still* enjoys Most Favored Nation trading status here in the U.S. But Wall Street will be needed to delay or even prevent future destruction at China's hands, by helping to implement the microprocessor applications detailed in this absolutely excellent Noahpinion article.
1. Wasn't the electronics industry increasingly concentrated in East Asian countries (first Japan and then Taiwan and South Korea) even prior to the rise of China itself?
2. Isn't the equivalent of MFN status for China now entrenched via the WTO, meaning it is no longer something that the US could unilaterally revoke?
I will admit that I started sweating when after I submitted my proposal, I got around to listening to the china shock Hexapodia and Noah said to Brad "we do not need a strategic chips reserve". The die was already cast. Thanks again for running this, it was a genuine delight
I'm curious about how your concept of a Strategic Chip Pool could be implemented. Who would pay for it? And isn't even foundation chip design evolving so rapidly as to make a reserve of existing chips obsolescent?
So the goal of the standard wouldn't be to enforce stagnation or choke off innovation, but to ensure substitutability. Lots of foundational chips are in practice commodity style devices; the capabilities of low-performance microcontrollers don't vary much with time, and aren't differentiators of performance in the end application.
The Open Foundational body would identify design features (pincout, I/O mapping, physical form factor, packaging material, etc.) which aren't determinants of performance, and set a standard for those features. The goal would be to make these devices substitutable across suppliers. To quote my proposal: "while there are numerous semiconductor manufacturing firms, in practice there may only be a single supplier for a specified foundational component".
There would have to be iterative updates of these standards, such as in the ASHRAE building performance standards. The goal isn't to stifle innovation, so there's no technological mandate in production or performance; however you arrive at an Open Foundational device is fine (with the caveat that standardized architectural or interconnect protocols may be required).
If such a standard is in place, a government stockpile to backstop its own operations is feasible. In the long run this would probably include a mix of other devices, most likely FPGAs, AI accelerators, and memory, but to start out would fill it with foundational devices. The goal of this is to provide sufficient year-to-year demand to actually motivate investment in a buildout of open foundational manufacturing base. Even legacy fabs are incredibly capital intensive and complicated to construct (~$6 billion and 18 months). The industry is also super cyclical, so it's an article of faith among producers that they should ignore short run demand signals.
If you want more depth my proposal covers all of this in much greater detail, including the technical basis for the standard being feasible.
The US military can't even keep Chinese chips out of our weapons, which seems like the most important place to start. I think whether the chips in cars or phones are made in US, Europe, Japan, or China is much less important. Having a detailed breakdown of the supply chains for electronics is probably the most important of the suggestions here, since without that we don't even know what we should worry about.
As for weaponizing EDA, I think that is not really feasible. It is well known that the license managers for EDA are broken by the Chinese (and others) so they can run EDA tools anyway. The joke in the EDA industry is that you can sell any tool to a Chinese company, but only one copy. The Chinese already have all the EDA tools for non-leading-edge nodes, and the EDA companies don't have any way to stop them running them on their own servers or clouds.
So, the U.S. and other countries are serious in efforts to stop the importing of illegal drugs (e.g. fentanyl). If sales to China of certain technologies is so critical to the Chips War, why not criminalize these activities? If the west decides a particular set of technologies would give China a leg-up in the Chips War, a perp walk of a few executives of corporations illegally selling these technologies to China is more meaningful than fines.
Based on no technical knowledge whatsoever, it seems to me we (i.e., the world) face choices similar to those we made in our history as new technologies have developed. For example, atomic energy--in the 40's there was the debate between sharing technology with the Soviets or walling them off from it. The outcome has been separate developments among different nations or blocs of nations. A more recent example is space technology, where China and US/EU/Japan/Russia have separate space stations, and the miliarization of space seems well underway as the world abandons the idealism of the 1960s.
A different example might be Morse code. which ended by becoming the universal language of telegraphy. The development of the international postal system and the internet also offer examples.
The different between the cases seems to be that atomic energy is seen as increasing vulnerability, while communication networks offer more benefits and don't seem to create vulnerability.
My suggestion is that national security isn't the only parameter by which to evaluate these proposals.
If there's any system which could provide incentives for China to participate, that should be considered.
I appreciate the problem Western economies have with Chinese subsidies but, back 70 years ago, we had the same problem with Japanese technology. The Japanese had no problem setting up factories to produce goods for export with Made in Britain on them because they had named the location of the factories as Britain. We solved this problem by becoming as efficient as the Japanese by allowing them to open factories in Britain to produce the goods we needed. The French had eleven companies producing washing machines when the Common Market opened the borders and import tariffs were abolished. Most of the French factories closed down due to the Italian washing machines being better and cheaper. The same happened in Britain when the restrictions on Japanese motor cars were removed.
A worse problem arose at the start of the 20th century when Germany started a race to build a bigger warship fleet than the Royal Navy. This eventually led to WW1. Given the nature of politics and that China is a dictatorship bent on taking revenge for all the humiliations inflicted on them before 1949, we need to find a way to accommodate their annoyance. No good quoting Machiavelli about making them our friends instead of treating them as our enemies. China looks on weakness as an opportunity to exploit. My advice is to continue to outcompete them in technology using government subsidies to the same extent that China does. That will require cooperation between the countries of the West, in particular the Western members of the G7. It has to start with the Americans who are richer and better placed than Europeans to exploit the advantages of mass production.
Isn't one reason why East Asia has become the world's factory because it has so much of the world's population both to work in said industry, and to be a market for its products?
https://en.wikipedia.org/wiki/Valeriepieris_circle
That and the fact that they: a) Guaranteed cheap & compliant labor; b) Refused to add externalized costs (like environmental degradation) into their pricing.
Western, especially American, manufacturers fell utterly for it--even at the cost of giving 50% ownership to their Chinese subsidiaries, and total technology transfer. They then pretended to be shocked when their own Chinese partners matured to the point where they were able to outcompete and bankrupt them on American soil.
Wall Street got us into this mess: the urge to profit tends to trump patriotism. Which is why China *still* enjoys Most Favored Nation trading status here in the U.S. But Wall Street will be needed to delay or even prevent future destruction at China's hands, by helping to implement the microprocessor applications detailed in this absolutely excellent Noahpinion article.
1. Wasn't the electronics industry increasingly concentrated in East Asian countries (first Japan and then Taiwan and South Korea) even prior to the rise of China itself?
2. Isn't the equivalent of MFN status for China now entrenched via the WTO, meaning it is no longer something that the US could unilaterally revoke?
1. Yes.
2. Yes and no. MFN status can and has been amended at will by Congress despite WTO rules (e.g., Magnitsky Act). Or by Executive Orders. https://en.wikipedia.org/wiki/Most_favoured_nation#:~:text=China's%20MFN%20status%20was%20made,granted%20MFN%20status%20in%201996.
I will admit that I started sweating when after I submitted my proposal, I got around to listening to the china shock Hexapodia and Noah said to Brad "we do not need a strategic chips reserve". The die was already cast. Thanks again for running this, it was a genuine delight
I'm curious about how your concept of a Strategic Chip Pool could be implemented. Who would pay for it? And isn't even foundation chip design evolving so rapidly as to make a reserve of existing chips obsolescent?
So the goal of the standard wouldn't be to enforce stagnation or choke off innovation, but to ensure substitutability. Lots of foundational chips are in practice commodity style devices; the capabilities of low-performance microcontrollers don't vary much with time, and aren't differentiators of performance in the end application.
The Open Foundational body would identify design features (pincout, I/O mapping, physical form factor, packaging material, etc.) which aren't determinants of performance, and set a standard for those features. The goal would be to make these devices substitutable across suppliers. To quote my proposal: "while there are numerous semiconductor manufacturing firms, in practice there may only be a single supplier for a specified foundational component".
There would have to be iterative updates of these standards, such as in the ASHRAE building performance standards. The goal isn't to stifle innovation, so there's no technological mandate in production or performance; however you arrive at an Open Foundational device is fine (with the caveat that standardized architectural or interconnect protocols may be required).
If such a standard is in place, a government stockpile to backstop its own operations is feasible. In the long run this would probably include a mix of other devices, most likely FPGAs, AI accelerators, and memory, but to start out would fill it with foundational devices. The goal of this is to provide sufficient year-to-year demand to actually motivate investment in a buildout of open foundational manufacturing base. Even legacy fabs are incredibly capital intensive and complicated to construct (~$6 billion and 18 months). The industry is also super cyclical, so it's an article of faith among producers that they should ignore short run demand signals.
If you want more depth my proposal covers all of this in much greater detail, including the technical basis for the standard being feasible.
The US military can't even keep Chinese chips out of our weapons, which seems like the most important place to start. I think whether the chips in cars or phones are made in US, Europe, Japan, or China is much less important. Having a detailed breakdown of the supply chains for electronics is probably the most important of the suggestions here, since without that we don't even know what we should worry about.
As for weaponizing EDA, I think that is not really feasible. It is well known that the license managers for EDA are broken by the Chinese (and others) so they can run EDA tools anyway. The joke in the EDA industry is that you can sell any tool to a Chinese company, but only one copy. The Chinese already have all the EDA tools for non-leading-edge nodes, and the EDA companies don't have any way to stop them running them on their own servers or clouds.
Bravo all!
Great idea to cast a net of inquiry for ideas.
So, the U.S. and other countries are serious in efforts to stop the importing of illegal drugs (e.g. fentanyl). If sales to China of certain technologies is so critical to the Chips War, why not criminalize these activities? If the west decides a particular set of technologies would give China a leg-up in the Chips War, a perp walk of a few executives of corporations illegally selling these technologies to China is more meaningful than fines.
Based on no technical knowledge whatsoever, it seems to me we (i.e., the world) face choices similar to those we made in our history as new technologies have developed. For example, atomic energy--in the 40's there was the debate between sharing technology with the Soviets or walling them off from it. The outcome has been separate developments among different nations or blocs of nations. A more recent example is space technology, where China and US/EU/Japan/Russia have separate space stations, and the miliarization of space seems well underway as the world abandons the idealism of the 1960s.
A different example might be Morse code. which ended by becoming the universal language of telegraphy. The development of the international postal system and the internet also offer examples.
The different between the cases seems to be that atomic energy is seen as increasing vulnerability, while communication networks offer more benefits and don't seem to create vulnerability.
My suggestion is that national security isn't the only parameter by which to evaluate these proposals.
If there's any system which could provide incentives for China to participate, that should be considered.