the new license requirement will address the risk that the covered products may be used in, or diverted to, a ‘military end use’ or ‘military end user’ in China and Russia. The Company does not sell products to customers in Russia.
The regulations are revised on a regular basis so twenty years later maybe it's more like anything above a 3090 Ti. Also I suspect that "so powerful it's almost illegal" marketing back in the day was just marketing.
It's possible that the US got wind of some military/dual-use supercomputers based on A100/H100 and decided to block them but that's just speculation.
Even if you had your GPU-packed Kilo-Google per person multiplanetary Gigagalactic computer guessing numbers for 37 times the age of the universe, it would still have a 1 in 32 billion chance of cracking. https://youtu.be/S9JGmA5_unY?t=32
Your video is talking about 256-bit symmetric encryption. RSA is asymmetric encryption. RSA 2048 is ~112 bits of protection when comparing to symmetric keys, kinda. We're still talking a long time with current computing hardware though.
> For example, if SHA-256 is used with RSA and a 2048-bit key, the
combination can provide no more than 112 bits of security because a 2048-bit RSA key cannot provide more than 112 bits of security strength
- NIST SP 800-57 PART 1, Page 58
As far as I can tell, at least for components, it's never gotten less restrictive.
Any processor considered fast in the late 1980s is still under EAR despite being inadequate for a modern toaster.
It's truly ridiculous trying to manage the US Government's bureaucratic bullshit sometimes. I mean, really, a toaster.
Jokes on them. China should already have IP ownerships of these products just because of the fact that Nvidia has their manufacturing hubs based there. They will get the IP and then copy the product and sell them to the rest of the world in 1/10th of price. SEC is not stupid to not know this. May be they just want Nvidia stocks to go down so their already bought puts print now? Paul Pelosi recently bought those puts.
Semiconductors involve the most advanced manufacturing in human history, and even with blueprints of chips, China is decades behind other countries/companies who have fabs and would have no way to manufacture the chips.
If it was as simple as "steal blueprints", they would've built their own homegrown fabs that rival TSMCs, but they haven't. This is the same reason they haven't managed to build a modern jetliner that can even remotely compete with Boeing and AirBus. For certain types of products, having the blueprint is only half the battle, you need the manufacturing know-how to actually produce something that advanced, at scale, and at high quality.
And similarly, I can say that rather than put up an argument of value that stands on its own, you had to add barely related political commentary at the end of your ramble. This shows someone not arguing in good faith or trying to better understand a topic, but who just wants people to know their views on something.
So the only thing "classic" is your classic conservative rambling talking points that I've heard a thousand times. Try to have a conversation about a topic without injecting commentary on the political figures your type seems to be hopelessly obsessed with.
This is an SEC filing by Nvidia to disclose that the US Government has placed a restriction on these exports. This is not an SEC-issued or SEC-enforced export ban (which would be stupid and not make any sense anyway).
What a strange fantasy. Nvidia has no such exposure. Their products are produced using leading edge fabs which are not located in China. That board parters located in China integrate Nvidia chips into cards which are then sold to the Chinese market doesn't mean that those board partners have the IP to build the chips. They just have spec sheets for technical integration.
It's no different than China's COMAC who's domestically build airliner largely importing various Western components and integrating them. Just because they integrate the GE / Safran LEAP engine in their airliner, doesn't give them the IP or the capability to manufacturer that said engine.
One would assume that this would force the highest end GPUs to not be assembled in China, probably just another (good) strategy to ensure these strategically important resources aren’t being manufactured by an adversary.
I don’t know about the current generation, but past high end pre-“data center” branded Quadro and Tesla cards were manufactured by PNY in the United States and Taiwan (which is not the PRC).
Companies are supposed to watch for diversion of controlled tech. There are various red flags like buyers who aren't qualified to use what they're buying.
Charging companies for evading sanctions and export controls is a pretty regular thing. It’s not about single units crossing the borders but large numbers which are much harder to hide and investigation of such things isn’t that hard.
What do you do with 100,000 central asian guest workers each carrying one GPU in their cabin baggage when they go to Russia for their construction gig?
Billion dollar operations like that would be noticed by someone up the chain of suppliers. That just isn’t how these things usually happen. It’s usually a couple of hops of resellers in increasingly distant from American jurisdiction areas.
They are going to buy those GPUs in retail and offload for a 20% markup on Russian market. One or two GPUs and they paid for the plane ticket. They can even sell it on Avito on their own. Look, ma, no operation.
So you have noticed a lot of retail demand for those consumer goods, what would you do?
Could you describe how this works? What if a company in Kazakhstan buys them and then ships them to Russia? How would the US "charge" a Kazakhstan company? Is it AMD that would be charged? Would AMD be liable for every card they sold? How many levels of misdirection until they're liable?
There is a weird extraterritoriality concept that essentially says that if you use USD to commit a crime (they have the buy cards in USD at the beginning), that you can be prosecuted in the US, even if you are not from the US.
"any company or individual can be investigated, charged and sentenced because of a remote link with US jurisdiction, such as a transaction in US dollars, the existence of a subsidiary onto US territory, or an exchange with a US citizen."
I may be wrong (I'm just guessing) but I suppose that this may be one of the non-technical and legal reason crypto platforms prefer USDC or USDT rather than pure USD.
People in other countries don't use USD to buy Nvidia GPUs, do they? Does Nvidia only accept USD? I would assume they would have somewhat independent foreign markets set up.
ZTE, a Chinese company, was blacklisted for years and paid something like a billion dollars for evading US sanctions.
It is really quite hard to not have connections to the US economy. If you really don't want to do business with us, you probably want to bank with banks that have connections with us. If you don't play ball with our legal system, we'll go after your assets at your bank, and if your bank doesn't want to play ball we'll sanction them and completely cut them off from any bank connected to us. There is no foreign bank that is "too big to fail" if they don't comply with our courts, and really there aren't any that are going to stand behind somebody else's fraud.
> It is really quite hard to not have connections to the US economy.
I'm not sure I understand. AI startups are everywhere. It seems it would be trivial to set up a Kazakhstan AI "startup", buy a bunch, then ship them to Russia/China. Or are you saying that any non-governmental entity that would be interested in these would also be interested in the US market? And, someone friendly to the US would eventually peek in their server room and rat them out?
You would need a lot of capital to do that, where's it coming from and do the people who have it really want to be sanctioned, lose privileges to fly basically anywhere, have bank accounts, or travel anywhere at all?
Like everybody involved has to be fine with being airgapped entirely from the western banking system and any allied country or risk quite a long time in prison.
Who's got tens or hundreds of million dollars and wants to risk it evading export controls for a percentage profit?
Perhaps you don't understand, your Kazakstan bank would freeze your funds for the Americans because they didn't want to lose the ability to transact with any bank attached to the US or any bank attached to any bank attached to the US. The financial reach of US courts is long and not many people with lots of resources want to be blacklisted from ever traveling, spending money, or doing any transactions with the west.
And do you think the CIA isn't watching some random ex-soviet satellite state startup with sketchy funding sources buying enough A100s to build a supercomputer and then producing nothing with them? It's their whole job to look for interesting things, and they find them more than is imagined. If you hear about it, they're doing a bad job.
Everybody's entire phone records and so much more are part of an NSA dragnet. Do you really think large transactions for high potency computer equipment is being ignored?
Gazprombank seems to be doing business fine with western companies and countries, and they have branches in the CIS. So your post seems factually incorrect.
Yes, but chairman Xi is stupid enough to launch an invasion even involuntarily due to military provocations (misfires) to quell unrest at home by projecting anger outside.
Not that there's anything coming up right now, but Pakistan is looking extremely fragile right now with a dead economy and an unprecedented natural disaster. That, along with a more right wing India can't be good news.
There’s a long history of the top tier of computing hardware being export controlled. This is more of a recognition (a bit late eh?) of the utility of these graphics cards, though really it’s about time we started calling them something else.
An AI arms race will be limited by the silicon it's built on. The arms race will be in the fabrication of the chips. You can only scale old nodes so far.
The cheapest solution out of all the possibilities is to just use a back channel to get them, of course.
Allowing US companies to serve people relevant ads and allowing US companies to aid our number one adversary in developing the most potentially dangerous technology humans will every create are two very different things.
They can, but don't. Applying specific restrictions and switching from small fines to business-continuity-affecting rulings would solve the issue very quickly.
Since we have Cold War II (with the hot war still not fought by NATO), we should expect similarities to Cold War I: bans on high-tech exports, classifying more stuff, even public lies in an attempt to misdirect the adversary (see [SDI]). Good thing if encryption is not declared munition as it was back then [W].
Depends on what you're really trying to do. Having to pay that putative 30% markup on hardware is going to be a crippling disadvantage for a China-based AI product, for example.
> Just buy less GPUs and run them for more time. It's not like GPU clusters are running at 100% anyway
"Just get way better at managing compute resources. You know, that area of research every giant, multi-national cloud provider is pouring tons of money into. Just do that."
It's also going to be a 30% additional discount on Chinese TPU equivalent chips, especially since they have 7nm production now. I foresee it backfiring.
That depends on how SMIC's ramp goes. Right now they're shipping what amount to samples only. And even then there's no chips to fab absent more development work. Making a ML accelerator is hard. If AMD and Intel (actually Google and Tesla seem closer) can't beat NVIDIA, you really think Goodix is going to get it done?
High tech is still high tech for a reason. Sure, in the long term China is going to get there eventually. But in the intermediate decade or two, there's plenty of time for trade sanctions to work to shape markets.
Are they? Do you know why China stopped submitting any supercomputers to Top500 despite widespread reports of two exascale supercomputers coming online? And why would they prioritize commercial chip sales over national security projects?
AMD and Google have already beat Nvidia at the ML accelerator hardware game in the past. Google TPUs have had the edge over Nvidia cards for a few years, and AMD spent almost a decade with an edge in compute over Nvidia at every level. It's just that Google TPUs aren't commercially available and AMD GPUs lacked the software stack and critical mass.
Actually building a fast ML accelerator is far from a monumental task. It's difficult, but well within the reach of a country which fully designed a Top10 processor by itself. And now has the means to manufacture super high end chips.
It's not a case of intermediate decade or two. The Chinese are pretty much already there. They already have at least one indigenous exascale supercomputer and their own 7nm process. They can make an ML accelerator comparable to the A100 if they want to. Believe it or not, there are only so many ways of doing matrix multiplication in hardware. The hard part is lithography, and they did it.
EDIT: I forgot to mention, but SMIC wanted to keep their 7nm production secret to reduce incentives for the US to force ASML to stop 28nm lithography machines, keeping the impression that banning the sale of EUV machinery was enough to stop their advances. That's why they mix everything under 28nm in their financial reports, to show their very strong profitability without tipping their hand.
Your contention is that SMIC is successfully operating a high volume 7nm process... in secret? What potential incentive would they have to do that? This isn't a spy novel; if a company has a money factory they tell their investors about it!
Why they do this can be only guessed, but it is a fact that SMIC has made chips using a 7 nm FinFET CMOS process very similar to the TSMC 7 nm process.
This has been discovered by studying with an electron microscope some samples of ASICs made at SMIC for crypto mining.
This story has been reported in a large number of publications and it is possible to buy a detailed report with the result of the investigations about the secret 7 nm process of SMIC.
Because of the similarities with the TSMC process, and because it is known that SMIC has hired former TSMC employees, it is believed that SMIC might have copied various parts of the TSMC process, which could be a reason to not bragg about it.
However such copying is not as simple as non-specialists may believe. I have worked in semiconductor manufacturing, and even if you have all the details about how a technological process is done in another plant, when using other equipment you will not be able to reproduce the results without a lot of tweaking and changes.
Having stolen information about how the competition does it is helpful and it can shorten a lot the development time, but it cannot replace the need for a team of competent engineers, able to develop such a process by themselves, even if in a longer time.
They have an incredibly strong incentive which is to buy as many 28nm lithography machines as possible before the US forces ASML not to sell those as well! So they wanted to keep 7nm production secret for as long as possible, because the US operated on the incorrect assumption that blocking EUV sales was enough to stop SMIC from moving forwards.
If you look at the SMIC earnings report, you'll notice they added a new category called "28nm or less". That's their way of telling investors they have a money factory without telling the US to sanction them even more! Also, a very large part of SMIC >32% is owned by the Chinese state, and their profits have been skyrocketing in recent years.
Alibaba has already the fastest per socket server CPU for non-floating-point workloads, which has 128 Armv9 cores and DDR5 and PCIe5 before the launch of Intel Sapphire Rapids and AMD Genoa.
At this time, China is at least at parity with USA in chip design, so no export bans of US chips can hurt them.
Their only vulnerability is the dependence on the superior manufacturing capability of TSMC, so only banning exports from TSMC to China can hurt them (like Huawei was hurt when they were no longer able to make chips at TSMC).
It would take them years to acquire the GPUs at the scale that they need them. By the time they do that they're 2 generations behind. It's about adding a barrier.
You are right that the export ban for GPUs is much more likely to hurt Russia, who are nowhere near of being able to have their own GPU architecture.
China on the other hand, has already designed faster GPUs than those that are banned now for export to them, so they will not be affected. NVIDIA will be more affected by the ban than China, as they have already announced that they will have to delay the deployment of H100 until they will succeed to migrate the assembly of cards to locations outside China.
Other comments have already provided links to benchmarks comparing A100 with the Chinese Biren GPU.
This doesn't sound right to me. As a back-of-the-envelop calculation how much realistically would say the Russian government need? Even if they need in the league of 10,000s annually I think it would be easy for them to source this from a handful of friendly countries in a couple of months.
For example they could set-up a couple of "mining startups" in Dubai, Egypt, Serbia or India. Then buy 1000s of GPUs under the companies name and then put them in a yacht and ship to Moscow.
I literally know people who fly into Russia with entire disassembled gaming rigs in their suitcases (and do their best to avoid customs) from Turkey, etc. I'm sure if the Russian government needed these, they'd set up an operation to let them through. Russia is not as closed off as people seem to imagine.
And then Kazakhstan buys them from the Saudis or from Turkey, who are de facto US allies (Turkey is even a NATO member, a big one). The US can try and do something about it, as supposedly the CIA has just tried to do a couple of weeks ago in connection to the recent sanctions against Russia, but what do you when the Turks tell you to shove it? (which is what Turkey's response to the CIA demands was). This is one of the "best" sources I could find in English [1] because, obviously, the Western media has not reported on it.
Bingo. The current work-around to get electronics to Russia is via Dubai. It's harder with Turkey because their airlines don't allow lithium batteries for safety reasons.
any vendor has an owner who very likely enjoys traveling to the west, banking there etc. risking being on some black list for a marginal gain is not very attractive prospect
Very worrying to see the restrictions against China escalating;
Especially since China is incredibly advanced in hardware manufacturing and is not at war with the West.
At best, this is just pushing for escalation so they can restrict exports on their side too.
Even, for them. In one of their speech (in France, on a YouTube channel called Thinkerview for those who know), one of their representative quoted a Chinese proverb saying: "Nothing is more favourable than stability, nothing is more harmful than chaos".
Social stability, free markets and long-term peace are in the interest of both parties.
I'm not sure I understand (litterally, without being sarcastic or something).
You mean that the point of no return is already passed and that there is no reason to try to calm down things using diplomacy and cooperation, that it's purely on a power level that there is the solution ?
Are the China-embarrassing state visits to Taiwan an example of diplomacy and cooperation?
That kind of cooperation and diplomacy is called coercion and incitement.
The problem is, China and India have grown up now and they are going to behave like adults. And how do adult role models behave in the century XXI? Afghanistan, Lybia, Iraq, that's how.
No other country has such fits of rage to a diplomatic visit. Pretending that having leaders of one country visit another is coercion or incitement is gaslighting.
Probably the same kind of appeasement China does for the US by not selling weapons to Alaska and conducting freedom of navigation patrols between Alaska and Washington.
Sorry what does this comment mean? I genuinely don't understand (not a native speaker).
EDIT: Oh I think I get it, you want more cooperation, peace, among humans/countries etc.
Yeah, most people want it, but in today's world, it is increasingly unlikely.
On one hand, I don't like the fact that Western world emphasizes diversity and multiculturalism too much. We are too different, and forcing everyone to accept everyone is just pipe dream. At one point my values will clash with someone else's. I think fences, borders, are good. Deglobalization is good, slower economy is good. Just make sure we don't invade one another (I know, also pipe dream).
I also quite admire Xi Jinping. He is a dictator, and knows that he serves his country, an ethnic nation. He doesn't pretend to care about other ethnic minorities or other nations. He made China great. Hopefully China doesn't invade other countries, and prefer to be defensive.
In short: fences make good neighbors, let's all slow down on (forced) diversity and multiculturalism, let's all go back to our own fences, and cooperate from within the fences. We can visit our neighbors sometimes, have a good laugh, and after that we go home to our own fences.
Cooperation is overrated. I'd settle with peaceful coexistence. It is heaven on Earth.
I don't seek downvote, but I know I will be downvoted for this opinion, just make sure you squint to read the light colored comments.
Yeah you found the right word, peaceful coexistence :)
I'm maybe too dreamy, but I feel we are all prisoners on Earth, until we reach another planet.
And I believe game theory shows that prisoners are better collaborating ;)
As an entrepreneur, if I'd be running NVidia (or just investor in $NVDA), I'd be extremely annoyed that I am forbidden to sell my products to a large market, just because of political pressure.
Regarding your solutions and analysis, I don't know much about China politics other than gifs or memes that I saw on reddit, so I cannot say :D
The problem with being in the same cell is eventually one wants to smoke and one hates smoke. Everyone should just be in their own cell, and we can do whatever we want. We can still play card game and talk.
China is the only old civilization that still strive today. As a country, they are really big. It takes tremendous amount of leadership to bring China to what it is today.
Yes the memes about China are true, like the Uyghur Muslims treatment, the police state etc. But despite all of these, despite the lack of diversity, despite the lack of Western values, they still made great progress. It shows that a nation doesn't need diversity and multiculturalism to be strong.
I'm fine with diversity and multiculturalism, I just prefer the fenced version of it.
To tell the truth, I am scared of China too (and I am scared of their influence and power over my home country, I currently reside in the US). But I respect their strength.
> China is the only old civilization that still strive today.
i doubt that is actually true - the old chinese civilization is long gone. It's like saying that the american indian civilization is still here today.
> they still made great progress
they did - and to obtain that progress, they had to open up to world trade, to become part of the international community, and to cooperate.
It's proof that free market and international trade is not zero sum, and that all parties benefit.
But now that progress has been made, the chinese gov't seems to want to usurp more - both politically as well as economically. They desire the advances that their western trade partners enjoy, but don't desire the political freedom, and self-determination that would also come with such advances.
I think the US is not in a good strategic position to overcome china economically, unfortunately. The fact that the US has resorted to politically motivated bans like that is a clear sign. And china has progressed enough that i dont believe such bans will be effective at all - at most it will slow down the progress by about a couple years.
> i doubt that is actually true - the old chinese civilization is long gone. It's like saying that the american indian civilization is still here today.
Native American Indian didn't speak English, got obliterated by European settlers, their culture and their language destroyed. The American Indians have zero continuity today. They are a lost people. The Chinese was and still is the same, the Han. Despite Mongol and Japan's invasion, they still retained their culture, their buildings, their language, and their genetic make up.
> they did - and to obtain that progress, they had to open up to world trade, to become part of the international community, and to cooperate.
> It's proof that free market and international trade is not zero sum, and that all parties benefit.
I agree. Hence why they showed that a nation can become strong without forcing diversity. They just stick to their own self, their own culture, and generally absorb world's technology, research, and information. In today's free flow of information, we don't need diversity and multiculturalism.
Although, I argue that diversity and multiculturalism has zero correlation with why US is a strong nation. US is a strong nation simply because this land is abundant, rich, has a lot of river, geographically advantageous, and largely left alone when the whole world was fighting.
> But now that progress has been made, the chinese gov't seems to want to usurp more - both politically as well as economically. They desire the advances that their western trade partners enjoy, but don't desire the political freedom, and self-determination that would also come with such advances.
Yeah, I can understand why the Chinese govt want that. If I were the Chinese govt, I would think the same too. If I were an average Chinese citizen, I'd want the opposite.
> I think the US is not in a good strategic position to overcome china economically, unfortunately. The fact that the US has resorted to politically motivated bans like that is a clear sign. And china has progressed enough that i dont believe such bans will be effective at all - at most it will slow down the progress by about a couple years.
Native American cultures and languages are still very much alive. Have you not heard of the Cherokee and Navajo nations, and the many others? They are definitely not a lost people. Also, not all Chinese people are Han.
I’m glad you’re willing to take the mask off and show your true feelings. It’s really great for everybody in this forum who may have been persuaded by your hateful nativist screed to see what kind of person you truly are.
Could very well be, I sincerely hope not. If you remove the facade of petty wars of insert-your-current-thing, I was surprised to see that Biden is even more pro-US and anti-China/Russia than Trump from a strategic standpoint. US is becoming the IBM of sorts of nation. Slow, sluggish, inefficient and it's gonna continue to peddle on as far as no one fucks with the food supplies (or guns).
i dont think it's wrong to compare the US to IBM. The GDP per capita is high still because the returns from investments made many decades ago (in fact, almost a century ago!) are bearing fruit. I'm talking about highways, infrastructure and research/development.
But those investments will eventually be completely consumed. So what investments are being made today that would replace it?
IBM was still very profitable when their downfall began.
> "This low-volume production product may be the steppingstone for a true 7nm process that incorporates scaled logic and memory bitcells. Since bitcoin miners have limited RAM requirements, they likely do not feature the typical bitcell memory that the true 7nm technology definition requires (both scaled logic and bitcell adoption). This chipset likely demonstrates the logic part but not the bitcell aspect," said TechInsights.
Biren is made at TSMC, exactly as NVIDIA A100 and H100.
Biren is designed in China, like A100 & H100 are designed in USA.
As long as TSMC is not forbidden to export to China, forbidding only the export of chips designed in USA can no longer affect China, when they have pretty much identical chip design capabilities with USA.
you'd hope people who make these decisions also include some mid and long term thinking. On that time scale any politicization of trade is going to accelerate competitors and incentivize everyone to look towards alternatives. That's basically what's happened in software already.
A major reason the US has had sustained influence in the world without using much force is because it has not weaponized trade. It's the reason for the strength of the dollar, the dominance of US tech and finance, and so on. Throwing that away to merely stagger a country that's destined to catch up eventually doesn't seem wise.
I'm not a foreign policy expert, but the PRC seems to have very aggressive territorial ambitions. Its nine dash line incorporates vast swathes of land and water that international law recognizes as belonging to other countries.
To strengthen its claims in the West Pacific, the PRC has embarked on an unprecedented island building program in the region, and turned those islands into military outposts.
Here's a good breakdown of the history of PRC claims in the West Philippine Sea:
This is what the West tried to do with Russia. Germany especially was heavy on economic ties. "Surely if we're economically tied together, Russia won't risk their prosperity by full-on invading another country!" How well did that turn out for them?
A100 are current-gen Ampere NVIDIA cards on ~~Samsung's~~ TSMC's N7 processes for the datacenter, whereas H100 are next-gen Hopper NVIDIA cards on TSMC's N4 process for the datacenter.
Pascal as well! GTX1050Ti and below were at Samsung, while TSMC handled GTX1060 and above. I remember choosing a GTX1050Ti over the GTX1060 simply for the novelty, back before the first mining boom made vanity choices completely silly.
Nvidia has been with 5 fabs so far, with at least three generations of GPUs split between fabs.
Real Question is what yields have they been producing those chips at. Mining chips pay for themselves, so it's not as much of a big deal if your process is bad and only 20% of your wafer is usable. If you want to make a super computer out of chips from a process with 20% yields, then you're spending absurd amounts.
I have a question for you. Do you know that, or why, China stopped submitting new entries to the Top500, while numerous sources are claiming China brought online exascale computers with custom architectures?
Mining chips don't just simply pay for themselves. If it was cheaper to get the chips fabbed at Samsung or TSMC at any other lithography they'd be fabbed there. The company that made the miners isn't even Chinese. That it was fabbed there means that not only was there yield leftover from whatever state project was required, it was also commercially competitive.
I think it's generally accepted that this was a copy paste of the TSMC process. I think we'll have to wait to see if they can advance this, on their own. As the other person pointed out, a good indicator would be yield.
It's a copy past of the TSMC process because they were able to offer far higher higher salaries and attract TSMC talent. Given the reality that they have a lot of that talent and more each year and given the excellent track record of SMIC leadership, they will probably be able to advance on this, limited by the lithography machines they can procure.
If their yield was abysmal, no one would be willing to buy the chips since they aren't selling them at a loss and have plenty of public sector demand. The fact that they were able to sell them commercially and thus competitively means that their yields are okay at least.
Only people without experience in semiconductor manufacturing imagine that it is possible to copy and paste a CMOS manufacturing process.
Even with perfect knowledge about how a process is done at TSMC, implementing it in another fab with different equipment requires a very large amount of work done by very competent engineers, to substitute some process steps and tweak the technological parameters for the others.
Someone who succeeds to reproduce a stolen manufacturing process is very likely able to further develop such a process on their own.
Even within a single company, i.e. when having all the technical documentation and when having continuous access to all the people working there, it is extremely challenging to build a second semiconductor plant that will be able to make the same chips as a previous plant, with the same manufacturing process.
It took Intel many years of huge efforts until they have perfected a methodology to "copy exactly", to be able to build additional fabs for expanding the production of an already existing processor.
I think it's more nuanced. Left to develop it on their own, they might not succeed with a catch up effort in 5 years. Or 10. But in 20 years probably. 30 years certainly.
exactly - i think making commercially viable versions available for import into the country do more to disrupt their indigenous industry than banning the tech!
I don't think so. It's a little too convenient for the West to think this way, so there's a duty of skepticism. I think you can look at military stuff (fighter jets, spacecrafts, aircraft carriers), or civilian stuff like EVs, 5G network equipment, and see a clear ability to develop competitive and innovative indigenous technology when the sustained political will is there. It's bootstrapped off of imported platforms for sure but goes beyond what most other countries are able to produce.
Eh, this, and probably others in your list (Nortel breach with 5G is a candidate), weren't really independently developed [1]:
> When the Soviets refused to part with their Su-33 design secrets, China purchased an Su-33 prototype aircraft from Ukraine, dubbed the T-10K-3, and quickly set about reverse engineering it.
> The J-20, ... America’s F-22 Raptor. Plans ... were stolen by a Chinese national named Su Bin, who was sentenced to 46 months in federal prison for his crime.
> Lockheed Martin’s F-35 Joint Strike Fighter was also compromised by Su Bin, leading to China’s J-31 program
I think it will take some time to see if they can become independent. Or, maybe independence doesn't matter! Copy and improve may end up being a successful strategy.
Everything in competition is about time and effort. No one cares if they figure it out 20 years down the road. It literally won't matter at that point, because the US will have much better by then. The world may be a different place, etc. It's all about the now.
>The world may be a different place, etc. It's all about the now.
And that new world could be dominated by a new technology. Truth is that yes, with enough resources they can catch up technologically as many nations have in the past. That's no big surprise really, being able to produce ones own chips is so useful for defense you'd expect China and India to have that on a priority list somewhere. It wouldn't give them any advantage over other world powers
Yes, but once NVIDIA cannot sell to them, all the millions of dollars spent on NVIDIA products will now go to a domestic Chinese alternative. They will catch up fast.
Seems rather naive to assume that it for some reason would remain out of reach for a totalitarian state with a 1B+ population. And especially considering that there appear to be a few Chinese names on pretty much every major research paper released these days.
That's not how tech works. They will master and improve on 7nm, and probably future nodes. But the "reach" isn't 7nm, it's the continuous advancement. Rumors put Nvidia and AMD on 5nm, and Apple on 3nm.
As is, they've shown they can get something very close to the TSMC process working [1], partly by buying talent [2]. We'll have to see how that translates to long term improvements. But, they definitely can't rely on that strategy, long term.
Well, I think Chinese providers will not entrust in US technology moving forward. Who knows when next ban would come.
I think this ban will result in worse off situtation for both countries. China is a huge market for Nvidia after all. While China will become ever more isolated in the landscape of AI/Graphics community, which is a big bad news for it apparently.
China will create its own GPUs eventually, but probably with an incompatible standard that suits only Chinese market. Like Japan in 80s/90s.
12nm, about 5 years behind; it would be a very bad deal against current AMD Radeon RX offerings though the Chinese Economic Planning body would very likely throw a lot of support to secure domestic IP and production capacity.
FWIW, Pascal GPUs are completely usable for image classification, model training and even inferencing the full Stable Diffusion model. China doesn't need a dense node, just big server-grade GPU dies and the same software we have.
Realistically speaking the datacenters / enterprise would have to buy the next best thing in the short term, though you are right in that China (being a planned economy) will take this as economic warfare, designate GPUs as strategic resource and invest in creating domestic production capacity.
If high speed trains are anything to go by (China sends regards to the California High Speed Rail Project) then the US will be left behind by the end of the decade.
Most of the world trades more with China than with the USA, if it is contingent on that then not only does China have access they are most often the biggest trading partner.
> it will become more like cold wars, where market access will become limited to allies only.
Yes, it seems like the USA has admitted marketplace defeat and will proceed to isolate themselves as much as possible.
> Aug 31 (Reuters) - Advanced Micro Devices Inc (AMD.O) on Wednesday said U.S. officials have told it to stop exporting its top artificial intelligence chip to China, according to a spokesman.
> The company said new license requirements prevent shipment of its MI250 chips to China but it believes its MI100 chips are not affected by the new requirement.
It seems ridiculous that we don't officially acknowledge Taiwan. But to say it's not China, might be misleading. Does the Taiwanese government still make claim on China?
I think it's only ridiculous to the recipients of US MSM (which don't get me wrong, is hegemonic, but still insular). There's no confusion to it from the perspective of the RoC government (which claims a much wider legacy of the Qing government than the PRC does), the US government's official stance and 180 of 193 UN member states.
Not if both governments consider the people living on the other government’s landmass to be rightfully their citizens as well.
A close analogy might be to East and West Germany: both agreed that there should be just one place called “Germany”; they just didn’t agree on who should be in charge of it.
RoC only claims that because any deviation from their existing policy is "declaring independence" and what China wants as an excuse to invade and kill millions of people.
Which is only ridiculous and unknown to consumers of China and Russia MSM.
To say "to say it's not China, might be misleading" feels like it's time to (re)look up the history of modern China, starting with the Guangzhou uprisings and following that all the way to Mao literally wiping the slate and resetting what "China" meant (in the most literal sense: modern China isn't even 50 years old, the previous culture was actively murdered out of existence), while the original Chinese government fled to Taiwan in the expectation to eventually go back home and retake their seats =)
Its roots are related to how China became a republic in first place, when the Emperor was booted from power, China quickly became the land of warlordism, with generals treating their subordinates as personal army, and each warlord promoting a different ideology.
Although officially in history books people separate the period as many smaller civil wars, the fact is that China is having a single civil war that started back when the monarchy ended.
Taiwan itself originated when one of the groups in the civil war (The right Kuomingtang, beacuse by the way, the "Left Kuomingtang" is a thing and one of the few legal parties in China) was losing the war on mainland China, and had a golden opportunity in Taiwan:
Taiwan had been conquered some time before by Japan, and after WW2 ended, it was returned to China, the governor of Taiwan was very corrupt and ended pissing off the population that rioted, so Kuomingtang sent soldiers there to massacre the rioters (alongside a lot of random people). After a while they ended just sending everyone over and basically conquering the island by force.
So in a way, communist party "won" the civil war, they had almost the entirety of Mainland China, and started their communist dictatorship.
Meanwhile, Kuomingtang as they arrived on Taiwan, concluded the Taiwanese were too "japanified" and weren't "chinese enough", and decided that the right thing to do is murder people and terrorize them until they become Chinese, similar to what Mao was doing to "not communist enough" people. People call this period of time the "white terror"
So currently there are three positions about this:
1. Mainland government and a lot of people living there, support the idea Communist party must finish winning the civil war and erase Kuomingtang from the face of the earth, and take Taiwan.
2. Kuomingtang still exists, and their supporters among Taiwan's population believe they should prepare and then fight back, and win the civil war, wipe communist party out of the face of the earth, and re-take mainland China.
3. A variety of people, including a lot of native Taiwanese and pro-japan Taiwanese, want Taiwan to be a real country, that is not China, those that are more dictatorial in views want both the Communist party and the Kuomingtang gone.
Views 1 and 2 are both called "one china policy" and is how you manage to have strategic ambiguity. View 3 is what communist party is accusing Pelosi of spousing.
As for Taiwan current views: officially the government mostly claims there is one China, to avoid resuming the hot war. But pro-independence parties been making gains in elections. It might be just a matter of time until someone officially declares independence and the temporary (even if long lasting) truce ends.
I know a little some Taiwanese people, and can say they have a variety of views, one of the most obvious ones is a guy making indie games that put flag of Taiwan on his games splash screen with the phrase: "If you are not in favor of Taiwan independence you aren't human, and fuck Kuomingtang". I won't say name of games don't ask.
> Kuomingtang still exists, and their supporters among Taiwan's population believe they should prepare and then fight back, and win the civil war, wipe communist party out of the face of the earth, and re-take mainland China.
This is a rather outdated caricature. If anything, Kuomintang is now seen as being too soft on the PRC. Of course, the KMT is not a homogeneous blob and there are diverse viewpoints vis-a-vis Cross-Strait relations within the party. But there is zero support for taking the mainland by force.
Whit the the new delegation headed to Taiwan, they probably do not want to upset the Logan Act:
The Logan Act (1 Stat. 613, 18 U.S.C. § 953, enacted January 30, 1799) is a United States federal law that criminalizes negotiation by unauthorized American citizens with foreign governments having a dispute with the United States. The intent behind the Act is to prevent unauthorized negotiations from undermining the government’s position.[2] The Act was passed following George Logan’s unauthorized negotiations with France in 1798, and was signed into law by President John Adams on January 30, 1799. The Act was amended in 1994, changing the penalty for violation from “fined $5,000” to “fined under this title”; this appears to be the only amendment to the Act.[2] Violation of the Logan Act is a felony.
---
If we officially recognize Taiwan - then we are in dispute with China - and that fucks up a lot of things.
Its also why certain investments are funneled through Proxy Countries, such has been done recently through Ukraine...
> We don't need to "officially" recognize Taiwan when many congress folks invest heavily there.
Fun fact about laws: yeah you do. You can only litigate the letter of the law, not the spirit. If the law says "you cannot export to China" and the fabs are in Taiwan, then either NVidia's breaking the law by default, or they're not and the US officially recognizes Taiwan as not being part of China.
Whether that's what they intentionally went for when they wrote this: preeeetty sure it wasn't, but they just created a potentially massive legal situation.
You can absolutely write law that says "you can only sell to the Taiwan region of china". Not recognizing independence but still restricting sales to mainland china.
Laws and regulations can be written to work with such situations, e.g., to treat different regions differently. It's not even terribly unusual. I haven't checked into this particular case, but this situation is so long-standing that I suspect things have had plenty of time to get worked out.
No. The only countries which officially recognize Taiwan as independent are: Belize, Guatemala, Haiti, Holy See (Vatican City), Honduras, Marshall Islands, Nauru, Palau, Paraguay, Saint Kitts and Nevis, Saint Lucia, Saint Vincent and the Grenadines, and Tuvalu.
Slightly missed the point: NVidia uses TSMC to make their GPUs. TSMC is in (as the name suggests) Taiwan, so either NVidia -by default- is violating this order by exporting to China because the fabs are in China, or they're not exporting to China because the US now officially recognizes Taiwan as "not China".
This seems like an attempt to play lawyerball and find larger significance than exists in a fact. For instance, the US recognizes Taiwan passports for visitors, those are not China passports... and they treat visitors differently, see how Taiwan is on this list and China is not: https://www.cbp.gov/sites/default/files/assets/documents/201...
So "by default" the US already "officially" recognizes Taiwan as "not China." I think the mistake here is confusing de facto with "officially."
(To put it another way: even assuming your point that "technically" the US could try to prosecute Nvidia for this and a court would get the question of "is this China" - but do you actually think the US will? Or will they continue to walk the same line they've walked for decades in the same situation?)
In so many ways: it really wasn't. China made a bluff, because that's what you do when you're in power in China, and the US went "we're not Chinese" and then China went "k" and nothing happened and no one cared. China didn't back down, they never stepped up. Blustering is literally part of "that just what you do". You don't call someone's bluff in China, you never believed them in the first place, you only nodded initially because that's just what you do.
PRC set new military norm across the strait completely uncontested, and now post Pelosi policy wank in US is how to reset that new balance... and the answer was... 2 DDGs transiting through the strait, aka nothing. Not even a carrier. PRC chipped more away at defacto TW soverienty than US gained trying to boost TW dejure status.
>China's final warning
It's hilarious when people daftly throw this Soviet meme from the 60 which ended in Sino-Soviet border war in 69. PRC has gone to war with US+UN, USSR, India... threatend UK over HK handover, and supplied Vietnam against France over security concerns much smaller than TW. That makes PRC the only country who has no qualms with going hot with EVERY NPT nuclear state. China's final warning has consistently preluded to war.
For a long time the US has more or less treated Taiwan and mainland China as separate countries for all practical purposes, including trade regulations, while employing careful diplomatic language to avoid officially stating as much. It is a situation that is much more complicated in theory than it is in practice.
Banning the export of sensitive technology is not exactly the equivalent to a MacArthurian full throated embrace of the nuclear arts to defend an ally.
China's repeat statements about how it vows to bring Taiwan into the fold are like the gif of a truck about to crash from multiple angles that never actually crashes: https://c.tenor.com/cD8WqQ-ZGXcAAAAC/truck-crash.gif
At least to me this order shows the US govt is finally taking AI global implications seriously. And if they're taking those implications seriously, they will never let China have Taiwan.
The USA would not merely be defending an ally; it would be defending a part of its supply chain that somehow, somewhere along the line, became essential to pretty much every moving part of the country.
Moving enough fab to the continental United States to make a meaningful difference could be done long-term as a matter of strategic policy, but you can't make a change like that fast enough to address the USA's _present day_ geopolitical concerns. For now, Taiwan is kinda important.
RoC only claims that because any deviation from their existing policy is "declaring independence" and what China wants as an excuse to invade and kill millions of people.
No they don’t. They have a totally obsolete, totally ignored (by the government and everybody else) article in their constitution claiming China, which they can’t remove because the actual China has threatened to immediately invade them if they do.
That's only true if you take at face value an official lie that absolutely EVERYONE understands to be a lie, as well as understands that everyone else knows it's a lie.
That's like saying North Korea is a democracy because everyone calls it the Democratic People's Republic of Korea.
Territorial disputes are a problem for NATO membership, especially wrt article 5. Even if Taiwan gave up any claims to the Chinese mainland or other disputed territories, China's claims of sovereignty of the island could justifiably be enough to deny membership. On top of that, NATO is explicitly a European organization, with no member countries outside the European area besides USA and Canada.
https://www.nato.int/cps/en/natohq/official_texts_24733.htm
As a Chinese, what I want to say is that populism here has risen rapidly in recent years, the Thucydides trap has been fulfilled again, and a war is basically inevitable. is the United States determined to deal with this earth war? How much are you willing to pay to protect Taiwan?
>is the United States determined to deal with this earth war? How much are you willing to pay to protect Taiwan?
Yes the US is very much determined to win such a war. A free and independent Taiwan is the linchpin of American foreign policy in Asia. America is willing to spend great fortunes in treasure and blood to keep it that way.
Counting the track records … Vietnam, Afghan … apart from wasting lots of tax payer money and making the people in the area miserable, what else did we achieve?
Those are your go to for examples of the US lacking determination and resolve to stand by allies? Sure the US abandoned the South Vietnamese and the Afghan Governments, but only after two decades of military conflict in support of each. A dragged out conflict over Taiwan hurts the CCP the most.
Please remember the CCP has no feeling … hard to make it feel hurt. However, a dragged out conflict would hurt the people at TW and waste our money … let’s focus on building this country strong and take care the ppl here …
What a bizarre take. The people of Taiwan do not want to share the fate of Hong Kongers and are willing to fight to keep it that way. China imports 10 million barrels of oil a day. Any protracted conflict against the USN will see the CENTCOM stopping that flow of oil, which mostly comes from the Middle East. China is neither food nor energy secure and every additional month of conflict over Taiwan would be absolutely devastating to them.
Even without a war, China will be down to 650,000,000 people in the next 30-40 years. Throw a war into that mix and China's demographic numbers become even more grim.
In both Vietnam and Afghanistan, the US was participating (on the losing side) in an asymmetric civil war between local factions. In Taiwan they would be allied with a modern state, with a modern military and very high support among the population, attempting to repel an invasion from a (de facto) foreign state.
Politicians with donors and special interests to please keep pushing our foreign interventionism. The actual people here in the US don't really want to fight in any wars or instigate conflicts. But a representative democracy is not what we really have here any more.
Wars happen because each of the two sides think they can win.
In this case, the US is obviously confident it can win. This one is easy to see in this scenario.
China on the other hand questions the commitment of the US: “really, you’re going to go to war to protect Taiwan? It’s China! What do you care?”, and this could be true if geographic and geopolitical realities were different.
But what the CCP somewhat and I think many Chinese citizens misunderstand is that Taiwan isn’t just some island. Taiwan is a crucial linchpin to all of America’s presence and legitimacy in Asia.
So for China (more so Chinese citizens) it’s just some island that belongs to China anyway. Of course China would win and take over because Americans don’t really want to fight or think they can win right?
There lies the miscalculation that leads to the war. It’s not about Taiwan the island, it’s about existential foreign policy for America. The US is extremely determined to maintain the status quo. China misunderstands this and thinks that an invasion over Taiwan won’t lead to an actual war because they don’t understand what the war is about.
Note: When I say China, I mean people who question whether the U.S. is determined to fight a war over Taiwan. Certainly main Chinese including leadership in the CCP understand what’s at stake, which is why they haven’t undertaken an invasion as of yet.
The US will win in a non-war scenario, by using its influence over the other countries and isolating China from the US's western allies. It's an easy choice and seemingly an easy win.
A real military conflict with China near China's front door is another thing. I doubt anyone with some common sense should call that an easy win.
> The US will win in a non-war scenario, by using its influence over the other countries and isolating China from the US's western allies.
The US has lost a lot of their soft power, beginning with the Iraq war, but exploding during the 45th Presidency. Their failure to convince Germany to drop North Stream 2 is the best example of that loss - particularly the gang of senators which thought it would be a good idea to threaten a harbor with sanctions [1] was perceived as extremely rude here. And a lot of nominal allies are highly skeptical of how reliable an ally the US themselves are, given the very high likelihood of the Republicans getting obstructionist power again either this November or in two years and the lack of decent candidates on the Democrat side.
I agree that the US lost a lot of soft power capabilities over the years, but the Russian invasion of Ukraine really reversed most of those loses in short order. What a blessing in disguise for the United States. For example, you mention that a lot of allies are highly skeptical of how reliable of an ally the US is, but you can see actually that when it came down to it the US was extremely reliable in defending Ukraine and bolstering NATO forces, and the issue of Taiwan is a linchpin of US hegemony: it's existential, there's no doubt the US will go to war to defend allies in Asia including Taiwan because failure to do so basically kicks the US out of Asia and destroys those alliances. I think this goes back to my original comment about China's (or at least many of their more hawkish citizens, less so the CCP I believe) misunderstanding about what's at stake regarding Taiwan.
Though I think your point about obstructionist Republicans is well-founded and accurate and extremely accurate when it comes to domestic policies. With that being said I don't see a significant threat there in terms of US vs Russia or US vs China given that even when their vote was mostly inconsequential (i.e. they could try and virtue signal) most members of Congress, either Republican or Democrat, tended to vote to support US efforts against Russian aggression. The Senate was voting something like 99-1 to support Ukraine (don't quote me on the exact bill or bills) when they didn't really need to and could have tried to stick it to the Democrats as war mongering and followed in Trump's footsteps there.
> And the issue of Taiwan is a linchpin of US hegemony: it's existential, there's no doubt the US will go to war to defend allies in Asia including Taiwan because failure to do so basically kicks the US out of Asia and destroys those alliances.
For now it is, the interesting question is if the US voter class is ready to go to war again after the disasters of Iraq and Afghanistan - at least from an European POV it looks like wide swaths of the population do not want further overseas engagements and wish to reduce military spending in order to have more money available to fix pressing issues (crumbling infrastructure, healthcare, poverty, high energy prices). Additionally, while Trump himself is a very vocal opponent of China and I think Taiwan will be secured even if the Republicans win either of the next elections, it is nowhere near certain that Trump as a President will apply the same towards Russia or other enemies of the Western sphere - particularly not given strong indications of both himself and his close environment being influenced by Russia, both financially and via "kompromat" rumors (e.g. the "pee tape" allegation).
Polling, from even before the Russian invasion of Ukraine, says that a majority of Americans support defending Taiwan from Chinese aggression. After Ukraine, support is probably even higher. Even more, there is strong bipartisan support for it.
I think many outside of the US misunderstand the impact of the wars in Iraq and Afghanistan. Outside of political talking points and nightly news, the wars basically didn't exist. I mean if you just turned off the news, you'd have no idea that the US was fighting two wars. The military was at war, but not America. There weren't really any day-to-day effects and to the extent that there were, those are mostly long gone. I don't think that these wars have any bearing on additional US engagement, especially against China or Russia. In other words, I don't think the US is war weary at all, though I do think that out of the issues you mentioned some measure of the population doesn't think they're problems, except high energy prices which I think the US can mostly address domestically. I'd also add from a more realistic standpoint the military isn't preventing fixes for healthcare, poverty, or crumbling infrastructure. What's preventing addressing those issues is that some people don't think they are issues and those that do disagree on how to solve them.
The main threat to the US is dissolution of the democracy, but I think even in that case there would still be strong (maybe even stronger) support for defense of Taiwan because it's an existential issue for the US. For China to invade Taiwan with impunity would dissolve all American credibility in the Pacific and the ramifications of that are significant. It also would kick the US out in terms of trade in favor of China, which would do damage to the US economy.
There certainly is some concern that China may try to engage in a fast, hard engagement with the US and that could cause the US to withdraw, but given the significance of maintaining economic and military interests it's likely that it would just lead to a swift, hard response and from there we're probably in a pretty destructive war. I don't think even Trump or someone similar could prevent a US response here.
The US doesn't think they can "win" in the sense that someone 'wins' chess or checkers.
What the US can do however, it make the conquest of Taiwan very expensive. Sort of what they're doing in the Ukraine but probably with a shade more direct involvement. China will experience a abrupt cutoff from the US lead international order in addition to some limited skirmishes with western forces. But they also won't be physically prevented by western forces from occupying Taiwan. Western arms and intel will help fuel the Taiwan government until it's defeated. The goal isn't to win. It's simply to make the invasion very expensive.
Now, what I think western leaders are missing in this calculus is that China will not stop at Taiwan. The expense of exclusion from the western system is already priced in. It'll be paid for with additional expansion. So, after a short rest, China will ramp up with the annexation of additional territory within and around the nine dash-line setting up further conflicts with those countries. Taiwan's violent subjugation will give those countries pause and likely result in some combination of territorial concessions or vassalization; similar to the recent integration of the Solomon Islands into China's new extra-national order.
In parallel, Russia will solidify their holdings in western and southern Ukraine. I think at some point they'll test European and Nato resolve with a handful of attacks / incursions. But don't expect Nato to invade Ukraine or Russia anytime soon, because, you know, nukes. Ultimately, I predict Russia will move to finish their conquest of Georgia or accept their capitulation. Armenia will welcome Russian forces and Russia will begin moving further south into the middle east where they'll support Iran in conflict with some of the wealthy but poorly defended Arab states in the late 2020's or early 2030's.
I'd agree that describing this potential conflict as "win" or "lose" oversimplifies the engagement. But I would definitely disagree here:
> But they also won't be physically prevented by western forces from occupying Taiwan. Western arms and intel will help fuel the Taiwan government until it's defeated.
I'm a little perplexed because you seem to understand what's at stake, but you don't think the US would fight? All evidence points to the contrary. I mean not only is the US military present, the actual stakes are existential physical and economic security. If you aren't going to fight for those stakes I don't know what you do.
> In parallel, Russia will solidify their holdings in western and southern Ukraine.
I think this remains to be seen. I'm curious about how well Ukraine will do under current operations. It's bewildering to me that Ukraine can just strike and blow up Russian targets with impunity, including in Belgorod which presumably had some sort of air defense capabilities. I don't think Russia wants a fight with NATO unless the goal is to make defeat easier to swallow, because engagement with NATO necessitates a nuclear response from Russia due to what would clearly appear to be a decisive victory for NATO based on Russian performance in Ukraine and I'm not sure Russia wants to use nuclear weapons either. Engagement with NATO seems to be a lose/lose.
> I'm a little perplexed because you seem to understand what's at stake, but you don't think the US would fight? All evidence points to the contrary. I mean not only is the US military present, the actual stakes are existential physical and economic security. If you aren't going to fight for those stakes I don't know what you do.
The US is talking the talk but won't walk the walk. There remains virtually no political will to fight other peoples battles. This is why politicians from both major parties have argued for leaving NATO. Americans are tired of endless wars and the prospect of another is unbearable. Moreover, a protracted war with China will bring the sort of economic discomfort to the american populace not experienced since world war 2. American's don't have the will for it.
> I think this remains to be seen. I'm curious about how well Ukraine will do under current operations. It's bewildering to me that Ukraine can just strike and blow up Russian targets with impunity, including in Belgorod which presumably had some sort of air defense capabilities. I don't think Russia wants a fight with NATO unless the goal is to make defeat easier to swallow, because engagement with NATO necessitates a nuclear response from Russia due to what would clearly appear to be a decisive victory for NATO based on Russian performance in Ukraine and I'm not sure Russia wants to use nuclear weapons either. Engagement with NATO seems to be a lose/lose.
Russia's war strategy is very different from the west. Where the west see's troop and equipment loses as a negative, Russia views them as a mixed bag. Sure, it sucks to loose hardware but armed peasants will fight harder with the clarity of what loosing means. It's the pattern that won them world war 2 and brought them success in Chechnya.
Now more strategically think about it this way: Russia has no worries about loosing territory in the current conflict with Ukraine. Even if tomorrow they, say, provoke a Nato member by invading it... Nato will respond only to push Russian forces from their territory. Nato will not use nukes fist. And Nato lacks the political will to invade Russia or impose regime change. Since that's the case, Russia can really do whatever they want. Russia's warning to the west remains. Touch our land and we'll nuke you. And given no credible counter to this threat, Western powers will remain very wary of provoking a 'miss-calculation' that could lead to the use of nuclear weapons.
The same rational exists in Taiwan. So, if China invades Taiwan, Western forces might have a few exchanges with them, but do you really see American boots-on-the-ground fighting and killing hundreds-of-thousands or millions of Chinese troops? When would American troops arrive even?
After China starts the war? Because that's what it will take to 'win' for Taiwan. Okay, maybe you think American would do that. That's just defense. Can you imagine American troops landing in China for the purpose of threat neutralization? It'll never happen. Since the west won't violate China's borders China is also secure.
Territorial security is a partial basis of external belligerence. This is the lesson America has been teaching the world for the last 70 years.
> The US is talking the talk but won't walk the walk. There remains virtually no political will to fight other peoples battles. This is why politicians from both major parties have argued for leaving NATO. Americans are tired of endless wars and the prospect of another is unbearable.
I'm not sure this issue over Taiwan, certainly at the State level, is viewed as other people's battles. I think this issue is viewed squarely as what it is, which is global supremacy. And to your point about both parties arguing for leaving NATO, that tends to occur on both the far right and far left, but you see in practice massive support (99-1 kind of votes in the Senate which are unheard of) to support Ukraine even at the risk of escalation and in spite of the energy issues. Gas in the Great Lakes region was pushing over $5.00/gallon and not a single thing changed w.r.t US support for Ukraine. Europe is far more likely to bow out than the US here. But insomuch as both parties may have people who are arguing to leave NATO (which, no surprise is primarily due to Russian disinformation campaigns), virtually nobody is arguing to leave Taiwan to China. You can see continued US + ally action that runs contrary here. I don't believe that there is any evidence to suggest that at least the political leadership and State apparatus of the United States has anything in mind but maintaining dominance.
> Moreover, a protracted war with China will bring the sort of economic discomfort to the american populace not experienced since world war 2. American's don't have the will for it.
I agree that this is a risk, but the US is in a far better position here than China. While Americans may have a hard time obtaining iPhones, Chinese would have a hard time importing food or energy, especially as the US completely cuts off all shipping into the mainland. You're also guessing that this would be a protracted war, when that may not be the case. I'll dig it up if you are interested but the RAND Corporation did a pretty good study about the likely trajectory of the war and surmised that both sides would be incentivized to make the war as brutal and destructive for the other side as possible to get them to capitulate early, but that runs the risk of redoubling efforts to win the war versus capitulation. That is to say, it's possible (perhaps likely) that in the outbreak of a war we eventually wind up in a protracted conflict, but we may also wind up in a position where one side completely dominates the other and the war is short and intense. The scary thing here would be a long and intense war which I believe is most likely unless the US wins decisively and early.
> Russia's war strategy is very different from the west. Where the west see's troop and equipment loses as a negative, Russia views them as a mixed bag. Sure, it sucks to loose hardware but armed peasants will fight harder with the clarity of what loosing means. It's the pattern that won them world war 2 and brought them success in Chechnya.
I agree that Russia views this very differently than the west and it's why so many were surprised by Russia launching the invasion. They'd say things like "but why don't they just stop and sell things and coexist peacefully?" which is to misunderstand what the aim of Russia is here. However, that being said Russia's strategy, as you argue, of just losing troops until they win only works to the extent that they don't lose too much military equipment. This isn't the Red Army. You can also see this in the lack of Russia actually fully engaging in the war in terms of conscription, calling it a "special military operation", etc. Putin rightly fears that public opinion would turn against him and the war should Moscow and St. Petersburg and other more ethnically White Russian areas start to feel the effects in terms of a draft or something like that. So the point there is that Russia does have a limit on what it is willing to lose. P...
We'd be seeing significantly less concern in the West if the "official" Chinese worldview were more democratic, less authoritarian, more humanistic, and less illiberal.
How long will it take for the emerging Chinese middle class to force changes internally, move up Maslow's hierarchy, and demand freedom of conscience, speech, religion, assembly, press, petition, and privacy? IMO it's a bit of a race between the Chinese middle class' moderating influence and whatever spark (Taiwan?) sends us tumbling to war. Maybe if we can make it 60 more years (2 generations) without a war, the Chinese middle class will pick up some momentum in demanding their rights. That's the larger, longer-term tragedy of Tiananmen Square, that it set this process back decades (again, IMO).
Things are going to get worse in China, not better. In 60 years, they'll have half their current population (One Child Policy). Not only that, but Western companies have seen the writing on the wall. Production is moving out of China to Laos, Vietnam, Cambodia, etc.. Semiconductor facilities are being built in America.
We're not going to be able to avoid a war with China, because Pooh Bear has to have a scapegoat and the U.S. will be it, but make no mistake, China is on the way down, not up...
How so? The US never had a one child policy and replenishes population by allowing migrants to become Americans (try becoming a Chinese--this basically never happens). Manufacturing is moving to the US. Regarding scapegoating it's China where you see government propaganda adverts showing terrible and ugly capitalist western countries, nothing even close the other way around.
> How long will it take for the emerging Chinese middle class to force changes internally, move up Maslow's hierarchy, and demand freedom of conscience, speech, religion, assembly, press, petition, and privacy?
Standards of living in China were low very recently, for now middle class remembers this, and not having access to a parallel timeline (plus being under influence of ever-present propaganda) they think they should thank CCP for that. I suspect many of them are dissatisfied Taiwan is not attacked yet, hardly a moderating influence.
> We'd be seeing significantly less concern in the West if the "official" Chinese worldview were more democratic, less authoritarian, more humanistic, and less illiberal. |
As a Chinese have a mostly negative feeling about CCP government, I'm not sure if this is true.
Imagine hypothetically, Chinese government got magically replaced by a democratic regime. Is the intensifying competition between US and China going to stop? I think the same thing like US and Japan competition in the 90's, fuelled by purely economical reasons, will still be there.
While I deeply hope China can get a more liberal government in the future, I don't believe the current concern of US is driven by "belief in democratics".
Also look at non-democratic non-liberal at all countries like Saudi Arab, why US sell F15s to them but have concerns selling chips to China?
Theoretically it's supposed to be now (>$10K GDPPc). Yet, 4 years of Trump really wrecked everybody minds. Now those values aren't seen as ideal but surrender to the US. I saw Chinese reaction to the death of Gorby recently, wasnt pleasant.
Anyway, Chinese determination to recover Taiwan will not change even if her GDPPc is 50k or more. The quest of national restoration will not be abated by a mere richer country.
As you can see in Ukraine, Americans love an underdog. If Russia’s nuclear deterrent were any weaker, American soldiers would be fighting in Donetsk and the Harry S. Truman would be in the Black Sea.
Americans don’t necessarily love an underdog, they love anyone opposed to their geopolitical enemies. The US doesn’t support Ukraine because it’s invaded, they support Ukraine because it’s invaded by Russia.
"Protect Taiwan" isn't the core issue is it? It's the production of chips which is the rate of technological growth and mantel of "technological supremacy."
The importance of TSMC has been vastly exaggerated. In case of war their factories would be destroyed anyway. Remove TSMC and the importance of Taiwan remains exactly the same.
If the USA doesn't protect Taiwan it'll be their Suez canal moment. The moment the British and the French realized they were not great powers anymore.
>is the United States determined to deal with this earth war? How much are you willing to pay to protect Taiwan?
A majority of Americans support defending Taiwan from Chinese invasion. How much are the Chinese people willing to pay to attempt the largest amphibious invasion since D-Day while being starved of oil and international trade?
Did Japan ever think that it would be settled by an atomic bomb when it invaded the whole of Asia in World War II? Confrontation at any time is a game, who knows who is the final winner?
China under communist rule has a long history of changing their tune on a dime and the population going along with it without difficulty. If the Party decides tomorrow stoking nationalism is a bad idea it'll take a matter of weeks to have people saying, "The Taiwanese are our brothers, why would we fight them?" followed weeks later by "We never said we'd invade Taiwan, you just dont understand the way China works."
In the last 80 years the Chinese have become masters of holding diametrically opposed ideas. I believe they can back down from this without breaking a sweat.
Serious Question: What about the Apple M[x] series things?
The M2 integrates an Apple designed ten-core (or eight-core) graphics processing unit (GPU). Each GPU core is split into 32 execution units, which each contain eight arithmetic logic units (ALUs). In total, the M2 GPU contains up to 320 execution units or 2,560 ALUs, which have a maximum floating point (FP32) performance of 3.6 TFLOPs.
I'm starting to think it's a good idea to stock up on all the electronics you think you might need for the next decade. Because when China finally invades Taiwan, the supply of electronic devices worldwide is going to instantly disappear. How many electronic devices do you own that don't have any part of their supply chain in China or Taiwan or adjacent countries? The COVID shortages we all saw were practically business as usual compared to what may soon come...
I'm wondering what contingency plans US companies have for this? It seems to me like Apple would be screwed. Microsoft's hardware businesses would have problems. Google and Meta and Amazon are probably OK, though it may be tough to build and maintain datacenters, but everyone will be in the same boat.
The supply of TSMC chips won't disappear bc the US will defend Taiwan. It's far too important. The companies with major reliance on mainland China (Apple) are the ones that need to worry.
A Taiwanese war immediately means a naval blockade, the us won’t be able to lift it with any certainty. There is 0 interest in a hot war with China among Americans. I doubt we’ll see any American military action that could lead to a deployment other than naval.
The US government is more than willing to go to war to keep the world's chip supply out of China's hands. They're also pretty good at ginning up support back home.
The Chinese would be the ones to sink American vessels first. You don't simply declare a blockade, you enforce it. By sinking American ships that don't accept your blockade.
I'm doubt full capacity but I think there's no doubt the US will defend Taiwan with full force of military to protect those chips and this order is more evidence of that. They now seem to get how important AI is to future global order.
Yeah, and when China decides destroy their fabs with missiles. What are are you going to do then. In case you don't know, missile defense doesn't actually work well.
No, but trivial to shutdown fabs by destroying power infra down the chain. Capturing fabs that run on sanctionable supply chains to get look at hardware is secondary to PRC preserving fabs as barginning chip to restart semi production once countdown to western hightech industrial collapse starts ticking.
Well obviously if China wants to cripple the world's chip supply that would suck. They're not going to be allowed to control them though. Those fabs are rigged to be physically bricked if China is close to capturing TSMC anyways.
Protection against a nuclear power is not that easy. If US plans to escalate it with full power against China, chips would be the least of my worries, than earth being fit for life after the war.
Physical production of things during a war is probably quite hard. Also, one of the likely scenarios for the war is a blockade enforced by missiles. [1]
Of course it will. US isn't capable of defending TW anymore. They may spoil an invasion, but there's nothing US can do to prevent PRC from trivially destroying infra the keeps fabs operating. You should expect the entire semi supply chain in East Asian (which supplies the world) to stop functioning if US tries to intervene on TW, especially using JP or SKR infra, nearly everythign critical is within 60m PRC missile range. The reality is PRC has much more leverage over global semi supply chain in even of war vs US who has leverage during peace via sanctions.
> major reliance
Are all in US in terms of high tech industry supported and actual patent fees extracted from production of semi. PRC will lose bunch of manufacturing jobs, but indigenous semi efforts will keep chugging along. Both will be reset to as far back as 1980s tech depending on how PRC escalates, which ultimately cripples US as tech leader more.
My other concern is the collapse of the tech sector in a few months. I'm roughly 80% certain I'll be subject to layoffs in October (though all my coworkers are clueless), and I don't think it will be anywhere close to as certain I can find a new job as during the pandemic (then for every one company doing layoffs there were two with "unprecedented-who-could-see-this-might-end" growth).
Lot harder to justify a new PC when no longer receiving a paycheck, though it will likely be even more important to have one then.
Just got a new System76 machine and I have to say, I'm really loving it.
Hard times ahead, unfortunately after that, it's even harder times.
There are basically two types of companies in tech: a.) companies that don't make a profit, and b.) companies that make a profit from those companies. As an example just consider every startup is basically in category A, and Amazon's most profitable sector, AWS, in category B.
For more than a decade money was more or less free and the only sane choice for investors was to pour cash into anything that might one day get big enough to magically flip the profit switch and make a huge amount of net positive revenue.
I thought the pandemic would be the end of the tech boom, but I was wrong. Not only did more money get pumped into the market but into consumers wallets as well. This led to extremely rapid expansion in the tech industry as investors were willing to pump even more cash into any company (hoping for a rapid IPO and a the ability to cash our there chips before the casino closed) and consumers had a surprising amount of extra cash lying around from being locked up for months.
Unfortunately reality has started to show it's ugly face. Inflation is starting to hit hard, not because of the money supply, but because we are hitting some real resource limits due to a combination of supply chain problems and the Russian war with Ukraine (which, to be fair, is also happening because of resource strain). These are not temporary causes of inflation.
This has caused the central banks to rapidly attempt increase interest to combat this inflation. Ending the era of cheap cash. This is coupled with the fact that things cost more so consumers also can't spend as much (even though they are also taking on more debt then ever, an unsustainable trend).
Sure inflation is cooling a bit but that's largely because of a drop in oil prices (at least in the US) which is a political play from the democrats resulting the record low strategic oil reserves in the US (since 1984) [0]. This cheap oil cannot be sustained and will end after the next election cycle.
The immediate impact of all of this is that direct to consumer companies are seeing an immediate hit (e.g. streaming service subscribers are churning faster than ever). Then non profitable ad revenue companies start to take the hit (see Snap's layoffs today). Next it will be non profitable B2B. But then we'll start to see how much of those profitable companies bottom lines comes form non-profitable companies.
This will have a cascade effect, since as I mentioned, it won't be as easy for us tech workers to find an new job with anywhere near the same TC. As a 200k+ engineer, consider all the crap you pay for without thinking of it because you don't have to. All the ubers and door dash orders that are surprisingly pricy to the average person, but really a brainless decision for any tech workers.
It's just getting started. If you work at a B2B tech company I encourage you to go poke around your internal data and look at how your customers are doing. I promise it's not great.
Then look at the state of venture funding [1]. Sure it's still above 2018 levels but it's rapidly falling. The start up scene is going to dry up quickly.
There's is more I could dive into but this is a good start.
There are so very many companies that exist in category c.) companies that make a profit from non-tech companies. Many of them are not public. Many of them are small. Many of them "only" pay SWEs as CPA-level salary. Many of them are working on "boring" problems. But there are plenty. If you're concerned about categories A and B, perhaps search for a position at companies in category C?
CHIPS bill is a decent start but there is no replacing TSMC. A $50B watered down investment into US chips is small potatoes in grand scheme of things. Consider Apple spent $275B building up Chinese manufacturing to gain access to their market.
I am ready to buy every piece of hardware, smart phone, laptop, GPU in the minutes after an invasion or a blockade is announced, it will be the single best investment of a lifetime.
+ with such goods if you are in the EU, you have up to 14 days to send them back and get a full refund, so basically you have 14 days to resell them double the price with no downside risk.
Online it might, but you can still go to a store. Just have to run very fast and hope it happens when the stores are open otherwise it will be too late.
ideally you know an insider working there to reserve the products for you.
Hence Foxconn building up in Vietnam. I guess if I were a manufacturing company the whole ASEAN region would start to look a lot more attractive. A lot of the China-based mfrs are Taiwan owned and I suspect they are all starting to build for resilience in case of geopolitical trouble.
My understanding is that TSMC is busy building factories in the US. Easy to suspect that this is (1) to keep up production in the event that China takes over Taiwan, (2) to tell China now that even if they do take over Taiwan TSMC production will not stop or be dependent on China but will continue in the US and, thus, lower the interest of China in attacking Taiwan, and (3) once the Chinese military is on the way to Taiwan possibly let TSMC be eager destroy their Taiwan factories on their way out of Taiwan and on the next flight to the US, Belgium, etc., .... Besides, if China does not act nice, they might find that they can't buy any chips to use or copy. The world might regard selling chips to Xi now like selling cordite to Hitler in 1939.
The headlines about Taiwan, I'm guessing, basically add up to a lot of desire by the media for eyeballs for ad revenue and a lot of face saving by Xi, that is, don't let the media and politician talk erode the long existing images and status of Taiwan.
Net, China has plenty to do now without making a big, real issue about Taiwan.
But there is no force guaranteed to keep Xi from making a mistake. E.g., he had China muscle into Hong Kong and violate the "one country, two systems" promise.
I'm guessing that if Xi does to Taiwan what he did to Hong Kong, then TSMC will be producing in the US and on their way out of Taiwan will junk their factories so that China gets nothing.
>On July 27th, Nancy & Paul Pelosi sold all 25,000 of their shares of Nvidia, $NVDA. Today $NVDA was told to restrict chip sales in China & Russia by the US Government on August 26. $NVDA is down 20% since their initial sale, -5% in after hours today.
Cheaters can be incompetent. I find the timing of trades of stocks closely related to what congress was voting more compelling than the under performance. Maybe you are right, but either way, the incentive should not exist at all - lawmakers having access to the markets while in office should be a blight of shame on a nation.
This is kinda ridiculous in the sense of "we've done something!" I am originally from a country, which made a lot of millionaires during the war in Yugoslavia! It makes sense to ban mass products, but niche GPUs can be imported through other countries!
The chances of nothing getting through is zero, but it’s harder to do at scale and much more expensive. Plus the US does go after foreigners that help targeted nations dodge sanctions so can make it very miserable for China’s smugglers further exacerbating the shortage in China.
Normally I do not send a UA header so I configure the localhost forward proxy to add one automatically for www.sec.gov
It does not appear that the SEC verifies email addresses sent in UA headers.
A suprising policy for accessing public information hosted by a taxpayer-funded organisation.
I am not a "bot", I just prefer not to use Chrome/Chormium-based/Safari or other graphical browsers. I use a text-only browser. Generally, SEC filings are just text.
This doesn't just ban A100s. It bans those and "any future NVIDIA integrated circuit achieving both peak performance and chip-to-chip I/O performance equal to or greater than thresholds that are roughly equivalent to the A100."
It's a ceiling on the performance that can be exported. In an exponentially scaling industry, this is the equivalent of an announcement in advance of a complete ban on competitive products 2-3yrs out.
Given that both political parties have placed export limits on tech to China, it seems this is the new normal.
Eric Schmidt, the former Google chairman, told Reuters in a recent interview that high-end processors should have kill-switches.
“Knowing where the chips go is probably a very good thing. You could for example, on every chip put in essentially a public private key pair, which authenticates it and allows it to work”.
What he won’t tell is that this is already a reality, as I learned after having my air-gapped system and Pixel phone wiped remotely for researching “silent speech interfaces”. There is no security when silicon trojans are in all devices.
I'm going to put this nicely, so please take it in the constructive spirit intended -- serious claims require serious evidence to be believed. And a scattering of news articles or videos are not serious evidence.
I truly believe you're very concerned about all the things you talk about, but if you want to convince people to also believe then there needs to be more from your side.
And no, suggesting that people perform ptychographic x-ray laminography on their chips isn't a practical suggestion.
Fine with me. Evidence is better when people collect it themselves. If someone isn't willing to have their own air-gapped system, ask how they could bring about the same results that I found and carry experiments accepting the risks, they aren't properly looking for scientific knowledge and I am fine with them ignoring my claims.
People with willingness and means can find me anyways.
I reply as possible. Don't need or expect people to take my word for it, need people to actually investigate what actually is on our chips, whatever their motivation. Consider:
"It's not technically hard to make a device that complies with the FCC that listens to nonpublic bands but then is quietly waiting for some activation trigger to listen to other bands," said Eduardo Rojas, who leads the radio spectrum lab at Embry-Riddle Aeronautical University in Florida. "Technically, it's feasible."
To prove a device had clandestine capabilities, Rojas said, would require technical experts to strip down a device "to the semi-conductor level" and "reverse engineer the design." But, he said, it can be done.
last I checked that's not what's at issue. they can't produce equivalent chips themselves and rely on them for various physics simulations that back weapons research
Nvidia already has plenty of spyware in their drivers
Edit: I knew this would get downvoted quickly. Hey, mr or ms judge of truth, how about you check what nvidia services run on your machine before wielding all that power.
Somehow I feel that in our world the majority of corporations don't make direct comparisons to their competitors that often, preferring to use "competing brands" or something of that variety instead.
Avoiding a lawsuit is easy in that case: The Intel lawyer would have to make the argument that Pentium PCs are, in fact, harmful. Which would have opened a whole new can of worms.
Hard to see how Intel could sue over a statement like that.
The reason many companies say "competing brands" is because they don't want to mention an alternative, because it increases the mindshare of the other brand.
In many (most?) manufacturing industries there is an upper limit on export product performance. If you make steel, there are a bunch of types/grades subject to limits related to their use in nuclear technology. If you make microphones there is a limit on those that might be used in sonar arrays. Certainly all manner of limits exist in aerospace. Even game consoles have been limited (Iran). This is not an unusual regulation.
Even travel for US citizens was restricted. I have relatives who were in the semiconductor and aerospace industries during the 1980s. They were advised by the government to not attempt to travel to the Eastern Bloc. Some kinda related reading:
I knew about the previous cryptography restrictions, but I also knew that the Supreme Court overruled them by ruling (as I understand it) that “code is speech”. I’ll have to check out the other links though
Apple made a fairly big deal of the fact that the G5 processor, when they started putting it into their machines (or maybe it was when they went to dual G5s?), had just recently stopped being export-controlled by virtue of being classified as a "supercomputer".
(What really happened was that the Feds revised the definition of "supercomputer", and suddenly the G5 configuration they were using didn't qualify anymore... it had nothing to do with anything Apple did, except perhaps lobbying for it so they could build computers in China.)
There was also a ban export of strong cryptography. So Java couldn’t use string ciphers and algorithms without adding some additional jar which basically just turned the `enableStrongEncryption` flag on. This isn’t necessary anymore but I don’t know what changed legally.
I believe this is what changed legally:
> One of EFF's first major legal victories was Bernstein v. Department of Justice, a landmark case that resulted in establishing code as speech and changed United States export regulations on encryption software, paving the way for international e-commerce
Let's not forget that these policies fostered the development of encryption in Canada, Australia and other countries (See SSLEay for example).
So it ended up being counter-productive.
The simplest way to stop competition is to subsidize a product. In this case, offer the tools openly.
The opposite was of course done when the US sold a ton of Xeon Phi accelerators to China when they built the Tianhe-2 supercomputer.
My favourite conspiracy theory is that this was done deliberately by the US, since Xeon Phi was a large pile of steaming turd, so the Chinese wasted money on a machine with high theoretical FLOPS but crappy real world performance, and also wasted the time of their scientists and programmers who were porting code to a programming paradigm that went nowhere fast.
It's not just the US government, it's many civilized governments doing these sort of things. And in this particular case this european is grateful for it because it is not about setting ceilings or hindering industry, it's about preventing the abuse of dual use goods. In other words; Not being fsckd by our own stuff.
They often do! Hobby drones, GRBL hobby CNCs, hobby micrcontrollers and SBCs, open-source SDR projects, etc. will often exceed the performance of devices that were ITAR restricted a decade or two ago. And "a decade or two ago" is a generous expectation for lethargic legislative, regulatory, or judicial systems to evolve.
Yes, in the 70s and 80s, it made sense to say that 5-axis NCs systems that can cut Inconel turboprops for ICBMs should be export controlled; those were multi-million-dollar flagships of the biggest and best manufacturers and were key to the country's military edge. Now you can get one shipped from China for a couple thousand dollars, with servo loop times, encoder and linear scale resolution, memory capacity, processor speed, axis counts, etc. that are better than the originals that cost a thousand times more half a century ago.
Today, it's true that the A100 is a cutting edge, shockingly powerful, uniquely capable processor. You can do machine learning/computer vision projects with it that you can't do with any other technology. Eighty gigabytes! 256 GB/s per package! Almost 20 teraflops of FP32! Compared to the GPU in the desktop you bought 8 years ago, yes, these are huge. But if you plot interconnect bandwidth, memory size, core counts, price, and so on, and extrapolate forward... yeah, your little hobby Kubernetes cluster of 2030-era Raspberry Pi equivalents will probably be at a performance level that's export controlled.
I expect there is a real answer and a PR answer to this question; The PR answer has something to do with missile guidance and crypto currency. But the real answer i fear is that limiting import/export allows wars to start more easily.
It's a gradient. How you throw things down on that is up to you. Guantanamo vs. reeducation and labor camps. Paying judges to help keep the local juvenile prison full vs. harvesting organs from prisoners. Suppression of free speech and the media vs. whatever we see on Twitter daily.
It's impossible to quantify oppressive. We don't know much about Chinese labor camps or re-education centers.
Even if we did, how do you measure up acts done in these places against one another? Is making license plates for pennies in the US better or worse than making t-shirts in Xinjiang? Is being wrongfully thrown in prison due to a prejudice judge seeing your only skin color worse than being thrown into prison for being Muslim who refuses to shave their beard?
I agree it’s hard to quantify, and I agree US prisons aren’t really worse than Chinese ones. Rather, I believe they are roughly the same. Thus, condemning out one without even mentioning the other is simply dishonest. Especially when the condemning side is the competing perpetrator, as is usually the case with anti-Chinese propaganda.
Again, wholly subjective. I'd say it's much worse. It's also hard to take anyone seriously who has 50 Cent Army in their profile when it comes to issues on China.
Given that US is imprisoning many more people, for largely commercial (slavery) and racial reasons, how is China worse? I'm genuinely interested in your arguments here.
As a US prisoner I was worked until my hands bled, regularly forced to do things declared torture by the USA courts. Feel free to complain to the prison guard in charge of the complaints process. You have no access to the courts until you have exhausted the prison complaints process. Complain, and you will be placed in a cell with some unsavory characters, your room will be 'shaken down' every day, including your cellies stuff, and it will be made known the shakedown is due to your bucking the system. Your cellies will make sure you stop complaining and their cell stops getting tore up. The USA system is all for looks, but is impossible to actually get any access to remedy once you are on the inside and no longer in the world.
FYI: Things like fast food franchise remodels regularly have American forced labour doing a lot of the CAD/CAM work. If you eat at American franchise restaurants, you support American prisoner slave labour. [edit] here is a pretty government website promoting you to use slave labour: https://www.unicor.gov/Category.aspx?iStore=UNI&idCategory=1...
here is the US Constitution 13th Amendment banning slavery:
“Neither slavery nor involuntary servitude, except as punishment for crime whereof the party shall have been duly convicted, shall exist within the United States, or any place subject to their jurisdiction.”
> Please respond to the strongest plausible interpretation of what someone says, not a weaker one that's easier to criticize. Assume good faith.
Along with this one:
> Be kind. Don't be snarky. Have curious conversation; don't cross-examine. Please don't fulminate. Please don't sneer, including at the rest of the community.
Nah. Your interpretation is undoubtedly correct. I've been on HN 12 years now and I've become used to the increase in Western chauvinism over the recent years – mostly all of it led by US sentiment. The implication above is that the US is part of the civilized world (colloquially called "The West") and that China (and presumably Russia, and Iran, and Venezuela, and Cuba, et al) are part of the uncivilized world.
See, for instance, on this page where Hume speaks of civilized versus uncivilized forms of government: “This chapter examines Hume's conception of government. It considers three forms of government that Hume distinguishes: barbarous monarchy, civilized monarchy, and free government (with its two subdivisions, limited monarchy and republic).” So civilized is contrasted with barbarous. https://academic.oup.com/book/12307/chapter-abstract/1618482...
You'd think with SMIC demonstrating 7nm production ( as detailed in depth in this video by Asianometry: https://www.youtube.com/watch?v=dQGnwKBxAKk ) you'd think that the US had gotten the hint that the tech cat is out of the bag – or the tech genie has escaped the lamp or whatever metaphor you prefer … A comparison of SMIC's 7nm process versus the competition: https://www.techinsights.com/blog/smic-7nm-truly-7nm-technol...
My prediction is that SMIC joins TSMC, Intel, Global Foundries, and Samsung at the top tier of semiconductor manufacturing by 2030, along with an ecosystem of Chinese design and manufacturing companies. In doing so the West's stranglehold on other companies like Huawei will be over. If Russia can survive the punitive Western sanctions until then it'll no longer be hampered by lack of access to high-end Western technology.
I worked with a company that upgraded a mainframe cluster not located in the USA. The upgrade topped out a collective rating of that user's installed computing power allowance given for civilian purposes. So, to do the upgrade, US Export controls required the client to decommission some older nodes and that had to be witnessed by US Consular representatives walked through the whole process taking video and photographic evidence. And the agreed method of decommissioning was to haul these old nodes out in to a deserted area and run bulldozers over them in a pit. Not joking. Wish I kept some of the photos though.
ITAR has limits on processing power that can be exported: supercomputers have "dual use" and can be used for military applications, such as simulating nuclear explosions or designing hypersonic weaponry, e.t.c.
Parent poster likely worked for a US company that was doing a customer upgrade abroad (i.e. exporting US technology) that would have tipped the cluster compute capacity above the permissible limits, unless some older nodes were decommissioned and destroyed.
Truly a surprise change. The new requirements do not appear to be published in the Federal Register, which is a massive screwup that potentially renders the requirements illegal and invalid. My contacts in the trade compliance world are aghast.
Will be interesting to see if Nvidia challenges it.
One way that companies get around ITAR and similar regulations for things like space grade components that have radiation tolerant features is to only rad test up to what the limits are. The components are probably much more resilient than what they are specced for but by only declaring the legal limit, companies can still export their products. The foreign buyers can do their own testing to prove the components can meet their required levels. I've done this before for a commercial chip that was rumored to be radiation tolerant despite nothing from the vendor saying so and found that it well exceeded my expectations.
I wonder if Nvidia can use a similar loophole where the chips are clocked at lower rates for export to meet the law but can be "overclocked" with hardware modifications.
Chinese brand 8fun (bafang) ebike motors have been sold this way for a long time. AFAIK, the spec for their mid drive motors has been well below output potential.
> Under the Export Control Reform Act of 2018 (50 U.S.C. §§ 4801-4852) (ECRA), criminal penalties can include up to 20 years of imprisonment and up to $1 million in fines per violation, or both.
I'll eat a hat if any major league manager in the US who could decide this ever goes to prison. To be clear, I'd celebrate it, but the mortgage crisis showed that laws and punishment hardly work on top level execs.
Notice that it was just the broker who got sentenced here. He facilitated the trade, but he's certainly not the guy who made the decisions. Same as with the Diesel scandal: VW blamed it all on some middle-management engineer who was supposedly "acting alone." Poor schmuck got seven years in the US and when he was released he got deported to Germany and locked up for another year. I'll even double down right now and say that despite the investigation going against the former VW CEO, he'll never spend a day in prison. The same would hold with NVIDIA or any other major corp.
You do. not. FUCK. with. ITAR. The penalties are very steep. When you get caught, intentional violations can have jail time attached for the responsible parties (ie. intentional data breaches would fall on the CISO or closest equivalent). If there is one thing the US government takes very seriously, it's military and defense. They have, and will, fine a company that doesn't get it's shit together into the ground.
Reminds me of how in the eighties and nineties many MMIC's topped out at 8GHz, on the datasheet.
As 9GHz X-band capable stuff was more restricted dual use stuff.
Yet ERA-2SM for example had more gain on 10GHz than the 8GHz where the official spec ended at. :)
> where the chips are clocked at lower rates for export
You'd probably have a hard time claiming that was "peak performance" for the chip if it's underclocked with "nudge-nudge-wink-wink" overclocking available.
Sounds like the 80s/90s 208hp “gentleman’s agreement” in Japan. Most of those cars put out 320+hp with trivial bolt-on mods but they were all “208hp nudge nudge wink wink”
This could work for GPUs: Instead of selling a 20 TFLOP card for $500 you could sell 2 for $250 each consuming half the power. Wonder how long you could keep this up with exponential scaling though...
China would just import the GPUs from black market of countries not facing embargo, Even specific military spec components facing tight scrutiny end up there and other countries facing embargo[1], When Nvidia comes to learn about it they'll just keep quiet until it leaks out.
Not to mention, H100 is itself made in China and it seems Nvidia is allowed to keep building it there[2].
Not only that, by the time they figure how to do it the rest of the world will have moved on by at least 2 generations. SMIC has announced 14nm (but shipments in volume aren't clear), TSMC will have 5nm chips shipping in volume by the end of the month.
how does that matter? SMIC had the 7nm news recently, not sure about that but even if it is 14nm, can't they throw money on the problem and get something to start with?
they don't have to achieve A100 levels in the first go, even Rpi level chips should be enough to give them experience
I guess they are not able to buy the tooling (think ASML EUV) to make advanced node either. So, they’d also need to throw a shitload of time and money at this.
This is one of those problems that “throwing money at it” doesn’t speed it up. China has been throwing money at the problem for over a decade and they’re still behind.
Nvidia does their main chip fab in Taiwan AFAIK, along with a large number of other companies. It's one of the reasons tensions over Taiwan are worrying. China could blockade or invade Taiwan and start a "shortage" far worse than anything we've seen recently if they decided it was necessary/worthwhile.
Indeed, China has lots of other levers it can exert to cause a catastrophic shortage entirely on its own. Chip fabs in Taiwan are a big problem because there aren't many alternatives elsewhere, and due to the cost/time required to stand up a SOTA fab facility.
We build servers, test them, and ship them (in the US). No fucks given on the technical superiority of the `A100/H100` than parts that may be a generation older.
Yeah I mean, great for running inference (and obviously training smaller models). The problem is that Scaling Laws for Transformers show us that scaling model parameter count _up_, is a sure way to improve the model's performance over your task.
So when you're playing nation-state hard-ball, since the architectures are _somewhat_ trivially copied upon being published, that means the best checkpoint goes to the one with the fastest GPU's. If China wasn't behind in deep learning, they will be.
I can’t see how this would be effectively enforced - what’s to prevent them from spinning up an arm’s-length subsidiary (i.e. controlled through trustworthy middlemen) outside of US jurisdiction and using that to sell to prohibited markets? Like, exactly what, for instance, crocs are doing with Russia - or Tetley’s did with Teapigs (for different reasons).
Here is how Biren says it stacks up on various machine learning workloads, pitting the BR100 against the Nvidia A100:
…
We presume this is for AI training workloads, not inference. The average speedup over the A100 is around 2.6X. It is not clear if the Nvidia machines were using sparse matrix features, which doubles the throughput, or not. Our guess is they were not.
I'm optimistic that this diversity will unseat the CUDA framework's dominance in high-performance compute. OpenCL, AMD ROCm, and now "Birensupa" will compete with CUDA; hopefully leading to reduced dependence on any individual manufacturer's proprietary APIs.
Does anyone actually use ROCm? It seems barely supported by AMD, with minimal development, and difficult to run in nearly any scenario. It's been a massive disappointment for someone who generally prefers AMD gpus.
Remember when that hacker group showed screenshots of multiple stolen NVidia repositories with drivers source code and alledgedly detailed chip IP stuff?
>Biren was co-founded with Lingjie Xu, who was a senior GPU architect at Nvidia from 2008 through 2010, then a GPU architect at AMD for two years after that before taking a job as manager of GPU architecture at Samsung for five years after that. Significantly, Xu then became a director at Alibaba Cloud, the cloud division of the Chinese hyperscaler that is one of the Magnificent Seven top IT buyers in the world.
644 comments
[ 3.4 ms ] story [ 504 ms ] threadthe new license requirement will address the risk that the covered products may be used in, or diverted to, a ‘military end use’ or ‘military end user’ in China and Russia. The Company does not sell products to customers in Russia.
It's possible that the US got wind of some military/dual-use supercomputers based on A100/H100 and decided to block them but that's just speculation.
> For example, if SHA-256 is used with RSA and a 2048-bit key, the combination can provide no more than 112 bits of security because a 2048-bit RSA key cannot provide more than 112 bits of security strength - NIST SP 800-57 PART 1, Page 58
Warning, PDF: https://nvlpubs.nist.gov/nistpubs/SpecialPublications/NIST.S...
As far as I can tell, at least for components, it's never gotten less restrictive. Any processor considered fast in the late 1980s is still under EAR despite being inadequate for a modern toaster.
It's truly ridiculous trying to manage the US Government's bureaucratic bullshit sometimes. I mean, really, a toaster.
0. https://www.nextplatform.com/2022/08/25/china-launches-the-i...
Semiconductors involve the most advanced manufacturing in human history, and even with blueprints of chips, China is decades behind other countries/companies who have fabs and would have no way to manufacture the chips.
If it was as simple as "steal blueprints", they would've built their own homegrown fabs that rival TSMCs, but they haven't. This is the same reason they haven't managed to build a modern jetliner that can even remotely compete with Boeing and AirBus. For certain types of products, having the blueprint is only half the battle, you need the manufacturing know-how to actually produce something that advanced, at scale, and at high quality.
And similarly, I can say that rather than put up an argument of value that stands on its own, you had to add barely related political commentary at the end of your ramble. This shows someone not arguing in good faith or trying to better understand a topic, but who just wants people to know their views on something.
So the only thing "classic" is your classic conservative rambling talking points that I've heard a thousand times. Try to have a conversation about a topic without injecting commentary on the political figures your type seems to be hopelessly obsessed with.
It's no different than China's COMAC who's domestically build airliner largely importing various Western components and integrating them. Just because they integrate the GE / Safran LEAP engine in their airliner, doesn't give them the IP or the capability to manufacturer that said engine.
I wonder where does Nvidia manufacture them...
My EVGA RTX 3090’s box says “Made in Taiwan”.
Smuggling electronics beats smuggling drugs, danger-wise.
So you have noticed a lot of retail demand for those consumer goods, what would you do?
I know nothing about this.
"any company or individual can be investigated, charged and sentenced because of a remote link with US jurisdiction, such as a transaction in US dollars, the existence of a subsidiary onto US territory, or an exchange with a US citizen."
I may be wrong (I'm just guessing) but I suppose that this may be one of the non-technical and legal reason crypto platforms prefer USDC or USDT rather than pure USD.
What you mention is about an individual buying from an e-commerce site or a retail seller. That can happen in the local currency.
An example:
https://www.reuters.com/business/media-telecom/us-judge-rule...
ZTE, a Chinese company, was blacklisted for years and paid something like a billion dollars for evading US sanctions.
It is really quite hard to not have connections to the US economy. If you really don't want to do business with us, you probably want to bank with banks that have connections with us. If you don't play ball with our legal system, we'll go after your assets at your bank, and if your bank doesn't want to play ball we'll sanction them and completely cut them off from any bank connected to us. There is no foreign bank that is "too big to fail" if they don't comply with our courts, and really there aren't any that are going to stand behind somebody else's fraud.
I'm not sure I understand. AI startups are everywhere. It seems it would be trivial to set up a Kazakhstan AI "startup", buy a bunch, then ship them to Russia/China. Or are you saying that any non-governmental entity that would be interested in these would also be interested in the US market? And, someone friendly to the US would eventually peek in their server room and rat them out?
Like everybody involved has to be fine with being airgapped entirely from the western banking system and any allied country or risk quite a long time in prison.
Who's got tens or hundreds of million dollars and wants to risk it evading export controls for a percentage profit?
Perhaps you don't understand, your Kazakstan bank would freeze your funds for the Americans because they didn't want to lose the ability to transact with any bank attached to the US or any bank attached to any bank attached to the US. The financial reach of US courts is long and not many people with lots of resources want to be blacklisted from ever traveling, spending money, or doing any transactions with the west.
And do you think the CIA isn't watching some random ex-soviet satellite state startup with sketchy funding sources buying enough A100s to build a supercomputer and then producing nothing with them? It's their whole job to look for interesting things, and they find them more than is imagined. If you hear about it, they're doing a bad job.
Everybody's entire phone records and so much more are part of an NSA dragnet. Do you really think large transactions for high potency computer equipment is being ignored?
If Kazakh companies suddenly started buying GPUs to go in their datacenter that probably doesn't exist suspicions would be raised.
If the image generation market is getting too saturated, maybe OpenAI can sell something to the government. No one has deeper pockets.
Hard to see any particular country having the desire or where for all to undertake such an endeavour.
Also China has desire to resolve the civil war by other means.
Wherewithal?
The cheapest solution out of all the possibilities is to just use a back channel to get them, of course.
I doubt, though, that they will restrict the export of the raw material for creating AI: data.
Except these companies do way more than just serve innocent ads, as the Snowden reveals have showed; They are complicit in way more than that.
[SDI]: https://en.wikipedia.org/wiki/Strategic_Defense_Initiative
[W]: https://en.wikipedia.org/wiki/Export_of_cryptography_from_th...
The only way to really do this is to geo-fence disable the chips.
Just buy less GPUs and run them for more time. It's not like GPU clusters are running at 100% anyway.
(Also, implying China does not have enough money to buy them like that)
"Just get way better at managing compute resources. You know, that area of research every giant, multi-national cloud provider is pouring tons of money into. Just do that."
High tech is still high tech for a reason. Sure, in the long term China is going to get there eventually. But in the intermediate decade or two, there's plenty of time for trade sanctions to work to shape markets.
AMD and Google have already beat Nvidia at the ML accelerator hardware game in the past. Google TPUs have had the edge over Nvidia cards for a few years, and AMD spent almost a decade with an edge in compute over Nvidia at every level. It's just that Google TPUs aren't commercially available and AMD GPUs lacked the software stack and critical mass.
Actually building a fast ML accelerator is far from a monumental task. It's difficult, but well within the reach of a country which fully designed a Top10 processor by itself. And now has the means to manufacture super high end chips.
It's not a case of intermediate decade or two. The Chinese are pretty much already there. They already have at least one indigenous exascale supercomputer and their own 7nm process. They can make an ML accelerator comparable to the A100 if they want to. Believe it or not, there are only so many ways of doing matrix multiplication in hardware. The hard part is lithography, and they did it.
EDIT: I forgot to mention, but SMIC wanted to keep their 7nm production secret to reduce incentives for the US to force ASML to stop 28nm lithography machines, keeping the impression that banning the sale of EUV machinery was enough to stop their advances. That's why they mix everything under 28nm in their financial reports, to show their very strong profitability without tipping their hand.
This has been discovered by studying with an electron microscope some samples of ASICs made at SMIC for crypto mining.
This story has been reported in a large number of publications and it is possible to buy a detailed report with the result of the investigations about the secret 7 nm process of SMIC.
Because of the similarities with the TSMC process, and because it is known that SMIC has hired former TSMC employees, it is believed that SMIC might have copied various parts of the TSMC process, which could be a reason to not bragg about it.
However such copying is not as simple as non-specialists may believe. I have worked in semiconductor manufacturing, and even if you have all the details about how a technological process is done in another plant, when using other equipment you will not be able to reproduce the results without a lot of tweaking and changes.
Having stolen information about how the competition does it is helpful and it can shorten a lot the development time, but it cannot replace the need for a team of competent engineers, able to develop such a process by themselves, even if in a longer time.
Downvoter: Just literally read the news about Jack Ma and BABA and Tencent etc.
If you look at the SMIC earnings report, you'll notice they added a new category called "28nm or less". That's their way of telling investors they have a money factory without telling the US to sanction them even more! Also, a very large part of SMIC >32% is owned by the Chinese state, and their profits have been skyrocketing in recent years.
That is already in the past. They have already designed a datacenter GPU faster than NVIDIA A100 (made at TSMC):
https://www.servethehome.com/biren-br100-gpu-for-datacenter-...
Alibaba has already the fastest per socket server CPU for non-floating-point workloads, which has 128 Armv9 cores and DDR5 and PCIe5 before the launch of Intel Sapphire Rapids and AMD Genoa.
At this time, China is at least at parity with USA in chip design, so no export bans of US chips can hurt them.
Their only vulnerability is the dependence on the superior manufacturing capability of TSMC, so only banning exports from TSMC to China can hurt them (like Huawei was hurt when they were no longer able to make chips at TSMC).
Replacing CUDA programs will take work, but they have enough people for that and this is an easier task than designing a GPU.
Russia did that? Right...
China on the other hand, has already designed faster GPUs than those that are banned now for export to them, so they will not be affected. NVIDIA will be more affected by the ban than China, as they have already announced that they will have to delay the deployment of H100 until they will succeed to migrate the assembly of cards to locations outside China.
Other comments have already provided links to benchmarks comparing A100 with the Chinese Biren GPU.
For example they could set-up a couple of "mining startups" in Dubai, Egypt, Serbia or India. Then buy 1000s of GPUs under the companies name and then put them in a yacht and ship to Moscow.
[1] https://www.telesurenglish.net/news/CIA-Allegedly-Targets-Tu...
Especially since China is incredibly advanced in hardware manufacturing and is not at war with the West.
At best, this is just pushing for escalation so they can restrict exports on their side too.
Even, for them. In one of their speech (in France, on a YouTube channel called Thinkerview for those who know), one of their representative quoted a Chinese proverb saying: "Nothing is more favourable than stability, nothing is more harmful than chaos".
Social stability, free markets and long-term peace are in the interest of both parties.
What I am calling for: closer ties and negotiations and collaborative work.
The first step when you find yourself in a hole is to stop digging. Too late for that now, I have to admit.
You mean that the point of no return is already passed and that there is no reason to try to calm down things using diplomacy and cooperation, that it's purely on a power level that there is the solution ?
That kind of cooperation and diplomacy is called coercion and incitement.
The problem is, China and India have grown up now and they are going to behave like adults. And how do adult role models behave in the century XXI? Afghanistan, Lybia, Iraq, that's how.
Meanwhile, Kuril islands: contested by Japan but 70 years under Russian administration.
> any sign of appeasing would be positive
...yada yada doomed to repeat yada yada...
EDIT: Oh I think I get it, you want more cooperation, peace, among humans/countries etc.
Yeah, most people want it, but in today's world, it is increasingly unlikely.
On one hand, I don't like the fact that Western world emphasizes diversity and multiculturalism too much. We are too different, and forcing everyone to accept everyone is just pipe dream. At one point my values will clash with someone else's. I think fences, borders, are good. Deglobalization is good, slower economy is good. Just make sure we don't invade one another (I know, also pipe dream).
I also quite admire Xi Jinping. He is a dictator, and knows that he serves his country, an ethnic nation. He doesn't pretend to care about other ethnic minorities or other nations. He made China great. Hopefully China doesn't invade other countries, and prefer to be defensive.
In short: fences make good neighbors, let's all slow down on (forced) diversity and multiculturalism, let's all go back to our own fences, and cooperate from within the fences. We can visit our neighbors sometimes, have a good laugh, and after that we go home to our own fences.
Cooperation is overrated. I'd settle with peaceful coexistence. It is heaven on Earth.
I don't seek downvote, but I know I will be downvoted for this opinion, just make sure you squint to read the light colored comments.
I'm maybe too dreamy, but I feel we are all prisoners on Earth, until we reach another planet.
And I believe game theory shows that prisoners are better collaborating ;)
As an entrepreneur, if I'd be running NVidia (or just investor in $NVDA), I'd be extremely annoyed that I am forbidden to sell my products to a large market, just because of political pressure.
Regarding your solutions and analysis, I don't know much about China politics other than gifs or memes that I saw on reddit, so I cannot say :D
China is the only old civilization that still strive today. As a country, they are really big. It takes tremendous amount of leadership to bring China to what it is today.
Yes the memes about China are true, like the Uyghur Muslims treatment, the police state etc. But despite all of these, despite the lack of diversity, despite the lack of Western values, they still made great progress. It shows that a nation doesn't need diversity and multiculturalism to be strong.
I'm fine with diversity and multiculturalism, I just prefer the fenced version of it.
To tell the truth, I am scared of China too (and I am scared of their influence and power over my home country, I currently reside in the US). But I respect their strength.
i doubt that is actually true - the old chinese civilization is long gone. It's like saying that the american indian civilization is still here today.
> they still made great progress
they did - and to obtain that progress, they had to open up to world trade, to become part of the international community, and to cooperate.
It's proof that free market and international trade is not zero sum, and that all parties benefit.
But now that progress has been made, the chinese gov't seems to want to usurp more - both politically as well as economically. They desire the advances that their western trade partners enjoy, but don't desire the political freedom, and self-determination that would also come with such advances.
I think the US is not in a good strategic position to overcome china economically, unfortunately. The fact that the US has resorted to politically motivated bans like that is a clear sign. And china has progressed enough that i dont believe such bans will be effective at all - at most it will slow down the progress by about a couple years.
Native American Indian didn't speak English, got obliterated by European settlers, their culture and their language destroyed. The American Indians have zero continuity today. They are a lost people. The Chinese was and still is the same, the Han. Despite Mongol and Japan's invasion, they still retained their culture, their buildings, their language, and their genetic make up.
> they did - and to obtain that progress, they had to open up to world trade, to become part of the international community, and to cooperate.
> It's proof that free market and international trade is not zero sum, and that all parties benefit.
I agree. Hence why they showed that a nation can become strong without forcing diversity. They just stick to their own self, their own culture, and generally absorb world's technology, research, and information. In today's free flow of information, we don't need diversity and multiculturalism.
Although, I argue that diversity and multiculturalism has zero correlation with why US is a strong nation. US is a strong nation simply because this land is abundant, rich, has a lot of river, geographically advantageous, and largely left alone when the whole world was fighting.
> But now that progress has been made, the chinese gov't seems to want to usurp more - both politically as well as economically. They desire the advances that their western trade partners enjoy, but don't desire the political freedom, and self-determination that would also come with such advances.
Yeah, I can understand why the Chinese govt want that. If I were the Chinese govt, I would think the same too. If I were an average Chinese citizen, I'd want the opposite.
> I think the US is not in a good strategic position to overcome china economically, unfortunately. The fact that the US has resorted to politically motivated bans like that is a clear sign. And china has progressed enough that i dont believe such bans will be effective at all - at most it will slow down the progress by about a couple years.
Yeah, totally agree.
Han Chinese is the biggest Chinese ethnic make up.
That’s not what you said.
> Han Chinese is the biggest Chinese ethnic make up.
Didn’t say they weren’t.
Thank you generous Western nation. I hope my contribution, my taxes, my hard work benefits you.
One day I will go back home. I won’t be part of you anymore.
When that time comes, I hope you’ll leave us alone in our own affairs and don’t invade us or bomb us anymore.
Sincerely, A very hard working immigrant who could outwork most of your citizens, who just want to be left alone and prefer peace.
Enjoy encountering millions of people like me in this world. The silent majority.
Often times I wonder why Westerners never really learn that they shouldn't impose their way of life to the rest of the world.
Then I look at social media, and forum like this. I understand.
They still have their superiority complex, deep down to the bone.
> Sincerely, A very hard working immigrant who could outwork most of your citizens
The irony thickens
https://www.reddit.com/r/Damnthatsinteresting/comments/x332z...
See the evil in your people first before judging others.
https://www.reddit.com/r/Damnthatsinteresting/comments/x332z...
See the evil in your people first before judging others.
US already has inside chaos (and will get way way, way way, way way worse), don't need to worry about other countries chaos.
You should look at the GDP per capita of the US compared to literally every other country in the last 10 years.
i dont think it's wrong to compare the US to IBM. The GDP per capita is high still because the returns from investments made many decades ago (in fact, almost a century ago!) are bearing fruit. I'm talking about highways, infrastructure and research/development.
But those investments will eventually be completely consumed. So what investments are being made today that would replace it?
IBM was still very profitable when their downfall began.
SMIC 7nm: https://www.tomshardware.com/news/china-chipmaker-smics-7nm-... (ignore the journalist-speculative headline, but pay attention to the following)
> "This low-volume production product may be the steppingstone for a true 7nm process that incorporates scaled logic and memory bitcells. Since bitcoin miners have limited RAM requirements, they likely do not feature the typical bitcell memory that the true 7nm technology definition requires (both scaled logic and bitcell adoption). This chipset likely demonstrates the logic part but not the bitcell aspect," said TechInsights.
Biren: this looks to be diffused by TSMC.
Biren is designed in China, like A100 & H100 are designed in USA.
As long as TSMC is not forbidden to export to China, forbidding only the export of chips designed in USA can no longer affect China, when they have pretty much identical chip design capabilities with USA.
A major reason the US has had sustained influence in the world without using much force is because it has not weaponized trade. It's the reason for the strength of the dollar, the dominance of US tech and finance, and so on. Throwing that away to merely stagger a country that's destined to catch up eventually doesn't seem wise.
Are you sure? All the economic sanctions which the US heaps on "enemy state"s are nothing but weaponizing trade.
US semiconductor manufacturing has a bright future. The next decade will see an amazing push.
To strengthen its claims in the West Pacific, the PRC has embarked on an unprecedented island building program in the region, and turned those islands into military outposts.
Here's a good breakdown of the history of PRC claims in the West Philippine Sea:
https://youtu.be/H3NzRvvnjRQ
Nvidia has been with 5 fabs so far, with at least three generations of GPUs split between fabs.
ST, TSMC, IBM, UMC, and Samsung.
Mining chips don't just simply pay for themselves. If it was cheaper to get the chips fabbed at Samsung or TSMC at any other lithography they'd be fabbed there. The company that made the miners isn't even Chinese. That it was fabbed there means that not only was there yield leftover from whatever state project was required, it was also commercially competitive.
If their yield was abysmal, no one would be willing to buy the chips since they aren't selling them at a loss and have plenty of public sector demand. The fact that they were able to sell them commercially and thus competitively means that their yields are okay at least.
I would consider that a profound hurdle to advancement.
Nvidia has already moved to 5nm (released later this year).
Even with perfect knowledge about how a process is done at TSMC, implementing it in another fab with different equipment requires a very large amount of work done by very competent engineers, to substitute some process steps and tweak the technological parameters for the others.
Someone who succeeds to reproduce a stolen manufacturing process is very likely able to further develop such a process on their own.
Even within a single company, i.e. when having all the technical documentation and when having continuous access to all the people working there, it is extremely challenging to build a second semiconductor plant that will be able to make the same chips as a previous plant, with the same manufacturing process.
It took Intel many years of huge efforts until they have perfected a methodology to "copy exactly", to be able to build additional fabs for expanding the production of an already existing processor.
Of course, and this is partly why they were successful: https://www.bloomberg.com/news/articles/2022-07-27/synopsys-...
> Someone who succeeds to reproduce a stolen manufacturing process is very likely able to further develop such a process on their own.
And, this is why we'll have to wait and see if their current, successful, strategy will result in future, independent, advances.
Eh, this, and probably others in your list (Nortel breach with 5G is a candidate), weren't really independently developed [1]:
> When the Soviets refused to part with their Su-33 design secrets, China purchased an Su-33 prototype aircraft from Ukraine, dubbed the T-10K-3, and quickly set about reverse engineering it.
> The J-20, ... America’s F-22 Raptor. Plans ... were stolen by a Chinese national named Su Bin, who was sentenced to 46 months in federal prison for his crime.
> Lockheed Martin’s F-35 Joint Strike Fighter was also compromised by Su Bin, leading to China’s J-31 program
I think it will take some time to see if they can become independent. Or, maybe independence doesn't matter! Copy and improve may end up being a successful strategy.
1. https://www.popularmechanics.com/military/aviation/g23303922...
And that new world could be dominated by a new technology. Truth is that yes, with enough resources they can catch up technologically as many nations have in the past. That's no big surprise really, being able to produce ones own chips is so useful for defense you'd expect China and India to have that on a priority list somewhere. It wouldn't give them any advantage over other world powers
They do have indigenous CPUs and GPUs.
That's not how tech works. They will master and improve on 7nm, and probably future nodes. But the "reach" isn't 7nm, it's the continuous advancement. Rumors put Nvidia and AMD on 5nm, and Apple on 3nm.
As is, they've shown they can get something very close to the TSMC process working [1], partly by buying talent [2]. We'll have to see how that translates to long term improvements. But, they definitely can't rely on that strategy, long term.
1. https://www.theregister.com/2022/08/01/column_7nm_chips_chin...
2. https://www.bloomberg.com/news/articles/2022-07-27/synopsys-...
https://www.top500.org/lists/top500/2022/06/
Traditional hpc needs a general purpose machine, AI benefits much more from clever but immutable additional hardware.
https://www.reuters.com/technology/amd-says-us-told-it-stop-...
I think this ban will result in worse off situtation for both countries. China is a huge market for Nvidia after all. While China will become ever more isolated in the landscape of AI/Graphics community, which is a big bad news for it apparently.
China will create its own GPUs eventually, but probably with an incompatible standard that suits only Chinese market. Like Japan in 80s/90s.
If high speed trains are anything to go by (China sends regards to the California High Speed Rail Project) then the US will be left behind by the end of the decade.
Which I don't think is a given, it will become more like cold wars, where market access will become limited to allies only.
https://howmuch.net/articles/trade-timelapse-usa-china
Most of the world trades more with China than with the USA, if it is contingent on that then not only does China have access they are most often the biggest trading partner.
> it will become more like cold wars, where market access will become limited to allies only.
Yes, it seems like the USA has admitted marketplace defeat and will proceed to isolate themselves as much as possible.
edit.... but apparently the AMD MI200 datacenter chips are subject to this
https://www.reuters.com/technology/amd-says-us-told-it-stop-...
> Aug 31 (Reuters) - Advanced Micro Devices Inc (AMD.O) on Wednesday said U.S. officials have told it to stop exporting its top artificial intelligence chip to China, according to a spokesman.
> The company said new license requirements prevent shipment of its MI250 chips to China but it believes its MI100 chips are not affected by the new requirement.
A close analogy might be to East and West Germany: both agreed that there should be just one place called “Germany”; they just didn’t agree on who should be in charge of it.
Which is only ridiculous and unknown to consumers of China and Russia MSM.
It's not China, because of too much history.
To start, why Taiwan exists?
Its roots are related to how China became a republic in first place, when the Emperor was booted from power, China quickly became the land of warlordism, with generals treating their subordinates as personal army, and each warlord promoting a different ideology.
Although officially in history books people separate the period as many smaller civil wars, the fact is that China is having a single civil war that started back when the monarchy ended.
Taiwan itself originated when one of the groups in the civil war (The right Kuomingtang, beacuse by the way, the "Left Kuomingtang" is a thing and one of the few legal parties in China) was losing the war on mainland China, and had a golden opportunity in Taiwan:
Taiwan had been conquered some time before by Japan, and after WW2 ended, it was returned to China, the governor of Taiwan was very corrupt and ended pissing off the population that rioted, so Kuomingtang sent soldiers there to massacre the rioters (alongside a lot of random people). After a while they ended just sending everyone over and basically conquering the island by force.
So in a way, communist party "won" the civil war, they had almost the entirety of Mainland China, and started their communist dictatorship.
Meanwhile, Kuomingtang as they arrived on Taiwan, concluded the Taiwanese were too "japanified" and weren't "chinese enough", and decided that the right thing to do is murder people and terrorize them until they become Chinese, similar to what Mao was doing to "not communist enough" people. People call this period of time the "white terror"
So currently there are three positions about this:
1. Mainland government and a lot of people living there, support the idea Communist party must finish winning the civil war and erase Kuomingtang from the face of the earth, and take Taiwan.
2. Kuomingtang still exists, and their supporters among Taiwan's population believe they should prepare and then fight back, and win the civil war, wipe communist party out of the face of the earth, and re-take mainland China.
3. A variety of people, including a lot of native Taiwanese and pro-japan Taiwanese, want Taiwan to be a real country, that is not China, those that are more dictatorial in views want both the Communist party and the Kuomingtang gone.
Views 1 and 2 are both called "one china policy" and is how you manage to have strategic ambiguity. View 3 is what communist party is accusing Pelosi of spousing.
As for Taiwan current views: officially the government mostly claims there is one China, to avoid resuming the hot war. But pro-independence parties been making gains in elections. It might be just a matter of time until someone officially declares independence and the temporary (even if long lasting) truce ends.
I know a little some Taiwanese people, and can say they have a variety of views, one of the most obvious ones is a guy making indie games that put flag of Taiwan on his games splash screen with the phrase: "If you are not in favor of Taiwan independence you aren't human, and fuck Kuomingtang". I won't say name of games don't ask.
This is a rather outdated caricature. If anything, Kuomintang is now seen as being too soft on the PRC. Of course, the KMT is not a homogeneous blob and there are diverse viewpoints vis-a-vis Cross-Strait relations within the party. But there is zero support for taking the mainland by force.
https://www.taiwannews.com.tw/en/news/4576006
The majority wish is peaceful reunification, with a small (but growing) wish for independence.
How is he going to make 5 Trillion in ShareHolder Value by 2025?
-
We don't need to "officially" recognize Taiwan when many congress folks invest heavily there. Don't unplug the circuit that feeds you.
--
https://www.congress.gov/bill/117th-congress/senate-bill/442...
Whit the the new delegation headed to Taiwan, they probably do not want to upset the Logan Act:
---If we officially recognize Taiwan - then we are in dispute with China - and that fucks up a lot of things.
Its also why certain investments are funneled through Proxy Countries, such has been done recently through Ukraine...
Fun fact about laws: yeah you do. You can only litigate the letter of the law, not the spirit. If the law says "you cannot export to China" and the fabs are in Taiwan, then either NVidia's breaking the law by default, or they're not and the US officially recognizes Taiwan as not being part of China.
Whether that's what they intentionally went for when they wrote this: preeeetty sure it wasn't, but they just created a potentially massive legal situation.
So "by default" the US already "officially" recognizes Taiwan as "not China." I think the mistake here is confusing de facto with "officially."
(To put it another way: even assuming your point that "technically" the US could try to prosecute Nvidia for this and a court would get the question of "is this China" - but do you actually think the US will? Or will they continue to walk the same line they've walked for decades in the same situation?)
Yes, that's literally what the original question was?
China cared a lot. For sure, the US didn't believe china would do anything.
The stance of the US is clear though. China is under US protection.
Doubly so in light of the recent 1.1 billion dollar military aid deal: https://www.bloomberg.com/news/articles/2022-08-30/biden-adm...
>China's final warning
It's hilarious when people daftly throw this Soviet meme from the 60 which ended in Sino-Soviet border war in 69. PRC has gone to war with US+UN, USSR, India... threatend UK over HK handover, and supplied Vietnam against France over security concerns much smaller than TW. That makes PRC the only country who has no qualms with going hot with EVERY NPT nuclear state. China's final warning has consistently preluded to war.
China's repeat statements about how it vows to bring Taiwan into the fold are like the gif of a truck about to crash from multiple angles that never actually crashes: https://c.tenor.com/cD8WqQ-ZGXcAAAAC/truck-crash.gif
The USA would not merely be defending an ally; it would be defending a part of its supply chain that somehow, somewhere along the line, became essential to pretty much every moving part of the country.
Moving enough fab to the continental United States to make a meaningful difference could be done long-term as a matter of strategic policy, but you can't make a change like that fast enough to address the USA's _present day_ geopolitical concerns. For now, Taiwan is kinda important.
That's like saying North Korea is a democracy because everyone calls it the Democratic People's Republic of Korea.
Yes the US is very much determined to win such a war. A free and independent Taiwan is the linchpin of American foreign policy in Asia. America is willing to spend great fortunes in treasure and blood to keep it that way.
Let me guess you also oppose aiding Ukrainians?
Completely different scenarios.
When was the last time PLA had a live conflict?
In this case, the US is obviously confident it can win. This one is easy to see in this scenario.
China on the other hand questions the commitment of the US: “really, you’re going to go to war to protect Taiwan? It’s China! What do you care?”, and this could be true if geographic and geopolitical realities were different.
But what the CCP somewhat and I think many Chinese citizens misunderstand is that Taiwan isn’t just some island. Taiwan is a crucial linchpin to all of America’s presence and legitimacy in Asia.
So for China (more so Chinese citizens) it’s just some island that belongs to China anyway. Of course China would win and take over because Americans don’t really want to fight or think they can win right?
There lies the miscalculation that leads to the war. It’s not about Taiwan the island, it’s about existential foreign policy for America. The US is extremely determined to maintain the status quo. China misunderstands this and thinks that an invasion over Taiwan won’t lead to an actual war because they don’t understand what the war is about.
Note: When I say China, I mean people who question whether the U.S. is determined to fight a war over Taiwan. Certainly main Chinese including leadership in the CCP understand what’s at stake, which is why they haven’t undertaken an invasion as of yet.
A real military conflict with China near China's front door is another thing. I doubt anyone with some common sense should call that an easy win.
The US has lost a lot of their soft power, beginning with the Iraq war, but exploding during the 45th Presidency. Their failure to convince Germany to drop North Stream 2 is the best example of that loss - particularly the gang of senators which thought it would be a good idea to threaten a harbor with sanctions [1] was perceived as extremely rude here. And a lot of nominal allies are highly skeptical of how reliable an ally the US themselves are, given the very high likelihood of the Republicans getting obstructionist power again either this November or in two years and the lack of decent candidates on the Democrat side.
[1] https://www.dw.com/en/nord-stream-2-us-senators-threaten-ger...
Though I think your point about obstructionist Republicans is well-founded and accurate and extremely accurate when it comes to domestic policies. With that being said I don't see a significant threat there in terms of US vs Russia or US vs China given that even when their vote was mostly inconsequential (i.e. they could try and virtue signal) most members of Congress, either Republican or Democrat, tended to vote to support US efforts against Russian aggression. The Senate was voting something like 99-1 to support Ukraine (don't quote me on the exact bill or bills) when they didn't really need to and could have tried to stick it to the Democrats as war mongering and followed in Trump's footsteps there.
For now it is, the interesting question is if the US voter class is ready to go to war again after the disasters of Iraq and Afghanistan - at least from an European POV it looks like wide swaths of the population do not want further overseas engagements and wish to reduce military spending in order to have more money available to fix pressing issues (crumbling infrastructure, healthcare, poverty, high energy prices). Additionally, while Trump himself is a very vocal opponent of China and I think Taiwan will be secured even if the Republicans win either of the next elections, it is nowhere near certain that Trump as a President will apply the same towards Russia or other enemies of the Western sphere - particularly not given strong indications of both himself and his close environment being influenced by Russia, both financially and via "kompromat" rumors (e.g. the "pee tape" allegation).
The main threat to the US is dissolution of the democracy, but I think even in that case there would still be strong (maybe even stronger) support for defense of Taiwan because it's an existential issue for the US. For China to invade Taiwan with impunity would dissolve all American credibility in the Pacific and the ramifications of that are significant. It also would kick the US out in terms of trade in favor of China, which would do damage to the US economy.
There certainly is some concern that China may try to engage in a fast, hard engagement with the US and that could cause the US to withdraw, but given the significance of maintaining economic and military interests it's likely that it would just lead to a swift, hard response and from there we're probably in a pretty destructive war. I don't think even Trump or someone similar could prevent a US response here.
What the US can do however, it make the conquest of Taiwan very expensive. Sort of what they're doing in the Ukraine but probably with a shade more direct involvement. China will experience a abrupt cutoff from the US lead international order in addition to some limited skirmishes with western forces. But they also won't be physically prevented by western forces from occupying Taiwan. Western arms and intel will help fuel the Taiwan government until it's defeated. The goal isn't to win. It's simply to make the invasion very expensive.
Now, what I think western leaders are missing in this calculus is that China will not stop at Taiwan. The expense of exclusion from the western system is already priced in. It'll be paid for with additional expansion. So, after a short rest, China will ramp up with the annexation of additional territory within and around the nine dash-line setting up further conflicts with those countries. Taiwan's violent subjugation will give those countries pause and likely result in some combination of territorial concessions or vassalization; similar to the recent integration of the Solomon Islands into China's new extra-national order.
In parallel, Russia will solidify their holdings in western and southern Ukraine. I think at some point they'll test European and Nato resolve with a handful of attacks / incursions. But don't expect Nato to invade Ukraine or Russia anytime soon, because, you know, nukes. Ultimately, I predict Russia will move to finish their conquest of Georgia or accept their capitulation. Armenia will welcome Russian forces and Russia will begin moving further south into the middle east where they'll support Iran in conflict with some of the wealthy but poorly defended Arab states in the late 2020's or early 2030's.
> But they also won't be physically prevented by western forces from occupying Taiwan. Western arms and intel will help fuel the Taiwan government until it's defeated.
I'm a little perplexed because you seem to understand what's at stake, but you don't think the US would fight? All evidence points to the contrary. I mean not only is the US military present, the actual stakes are existential physical and economic security. If you aren't going to fight for those stakes I don't know what you do.
> In parallel, Russia will solidify their holdings in western and southern Ukraine.
I think this remains to be seen. I'm curious about how well Ukraine will do under current operations. It's bewildering to me that Ukraine can just strike and blow up Russian targets with impunity, including in Belgorod which presumably had some sort of air defense capabilities. I don't think Russia wants a fight with NATO unless the goal is to make defeat easier to swallow, because engagement with NATO necessitates a nuclear response from Russia due to what would clearly appear to be a decisive victory for NATO based on Russian performance in Ukraine and I'm not sure Russia wants to use nuclear weapons either. Engagement with NATO seems to be a lose/lose.
The US is talking the talk but won't walk the walk. There remains virtually no political will to fight other peoples battles. This is why politicians from both major parties have argued for leaving NATO. Americans are tired of endless wars and the prospect of another is unbearable. Moreover, a protracted war with China will bring the sort of economic discomfort to the american populace not experienced since world war 2. American's don't have the will for it.
> I think this remains to be seen. I'm curious about how well Ukraine will do under current operations. It's bewildering to me that Ukraine can just strike and blow up Russian targets with impunity, including in Belgorod which presumably had some sort of air defense capabilities. I don't think Russia wants a fight with NATO unless the goal is to make defeat easier to swallow, because engagement with NATO necessitates a nuclear response from Russia due to what would clearly appear to be a decisive victory for NATO based on Russian performance in Ukraine and I'm not sure Russia wants to use nuclear weapons either. Engagement with NATO seems to be a lose/lose.
Russia's war strategy is very different from the west. Where the west see's troop and equipment loses as a negative, Russia views them as a mixed bag. Sure, it sucks to loose hardware but armed peasants will fight harder with the clarity of what loosing means. It's the pattern that won them world war 2 and brought them success in Chechnya.
Now more strategically think about it this way: Russia has no worries about loosing territory in the current conflict with Ukraine. Even if tomorrow they, say, provoke a Nato member by invading it... Nato will respond only to push Russian forces from their territory. Nato will not use nukes fist. And Nato lacks the political will to invade Russia or impose regime change. Since that's the case, Russia can really do whatever they want. Russia's warning to the west remains. Touch our land and we'll nuke you. And given no credible counter to this threat, Western powers will remain very wary of provoking a 'miss-calculation' that could lead to the use of nuclear weapons.
The same rational exists in Taiwan. So, if China invades Taiwan, Western forces might have a few exchanges with them, but do you really see American boots-on-the-ground fighting and killing hundreds-of-thousands or millions of Chinese troops? When would American troops arrive even? After China starts the war? Because that's what it will take to 'win' for Taiwan. Okay, maybe you think American would do that. That's just defense. Can you imagine American troops landing in China for the purpose of threat neutralization? It'll never happen. Since the west won't violate China's borders China is also secure.
Territorial security is a partial basis of external belligerence. This is the lesson America has been teaching the world for the last 70 years.
I'm not sure this issue over Taiwan, certainly at the State level, is viewed as other people's battles. I think this issue is viewed squarely as what it is, which is global supremacy. And to your point about both parties arguing for leaving NATO, that tends to occur on both the far right and far left, but you see in practice massive support (99-1 kind of votes in the Senate which are unheard of) to support Ukraine even at the risk of escalation and in spite of the energy issues. Gas in the Great Lakes region was pushing over $5.00/gallon and not a single thing changed w.r.t US support for Ukraine. Europe is far more likely to bow out than the US here. But insomuch as both parties may have people who are arguing to leave NATO (which, no surprise is primarily due to Russian disinformation campaigns), virtually nobody is arguing to leave Taiwan to China. You can see continued US + ally action that runs contrary here. I don't believe that there is any evidence to suggest that at least the political leadership and State apparatus of the United States has anything in mind but maintaining dominance.
> Moreover, a protracted war with China will bring the sort of economic discomfort to the american populace not experienced since world war 2. American's don't have the will for it.
I agree that this is a risk, but the US is in a far better position here than China. While Americans may have a hard time obtaining iPhones, Chinese would have a hard time importing food or energy, especially as the US completely cuts off all shipping into the mainland. You're also guessing that this would be a protracted war, when that may not be the case. I'll dig it up if you are interested but the RAND Corporation did a pretty good study about the likely trajectory of the war and surmised that both sides would be incentivized to make the war as brutal and destructive for the other side as possible to get them to capitulate early, but that runs the risk of redoubling efforts to win the war versus capitulation. That is to say, it's possible (perhaps likely) that in the outbreak of a war we eventually wind up in a protracted conflict, but we may also wind up in a position where one side completely dominates the other and the war is short and intense. The scary thing here would be a long and intense war which I believe is most likely unless the US wins decisively and early.
> Russia's war strategy is very different from the west. Where the west see's troop and equipment loses as a negative, Russia views them as a mixed bag. Sure, it sucks to loose hardware but armed peasants will fight harder with the clarity of what loosing means. It's the pattern that won them world war 2 and brought them success in Chechnya.
I agree that Russia views this very differently than the west and it's why so many were surprised by Russia launching the invasion. They'd say things like "but why don't they just stop and sell things and coexist peacefully?" which is to misunderstand what the aim of Russia is here. However, that being said Russia's strategy, as you argue, of just losing troops until they win only works to the extent that they don't lose too much military equipment. This isn't the Red Army. You can also see this in the lack of Russia actually fully engaging in the war in terms of conscription, calling it a "special military operation", etc. Putin rightly fears that public opinion would turn against him and the war should Moscow and St. Petersburg and other more ethnically White Russian areas start to feel the effects in terms of a draft or something like that. So the point there is that Russia does have a limit on what it is willing to lose. P...
Ultimately I hope for peace and kindness for us all; as I suspect you do too.
How long will it take for the emerging Chinese middle class to force changes internally, move up Maslow's hierarchy, and demand freedom of conscience, speech, religion, assembly, press, petition, and privacy? IMO it's a bit of a race between the Chinese middle class' moderating influence and whatever spark (Taiwan?) sends us tumbling to war. Maybe if we can make it 60 more years (2 generations) without a war, the Chinese middle class will pick up some momentum in demanding their rights. That's the larger, longer-term tragedy of Tiananmen Square, that it set this process back decades (again, IMO).
We're not going to be able to avoid a war with China, because Pooh Bear has to have a scapegoat and the U.S. will be it, but make no mistake, China is on the way down, not up...
Standards of living in China were low very recently, for now middle class remembers this, and not having access to a parallel timeline (plus being under influence of ever-present propaganda) they think they should thank CCP for that. I suspect many of them are dissatisfied Taiwan is not attacked yet, hardly a moderating influence.
As a Chinese have a mostly negative feeling about CCP government, I'm not sure if this is true.
Imagine hypothetically, Chinese government got magically replaced by a democratic regime. Is the intensifying competition between US and China going to stop? I think the same thing like US and Japan competition in the 90's, fuelled by purely economical reasons, will still be there.
While I deeply hope China can get a more liberal government in the future, I don't believe the current concern of US is driven by "belief in democratics".
Also look at non-democratic non-liberal at all countries like Saudi Arab, why US sell F15s to them but have concerns selling chips to China?
Anyway, Chinese determination to recover Taiwan will not change even if her GDPPc is 50k or more. The quest of national restoration will not be abated by a mere richer country.
The importance of TSMC has been vastly exaggerated. In case of war their factories would be destroyed anyway. Remove TSMC and the importance of Taiwan remains exactly the same.
If the USA doesn't protect Taiwan it'll be their Suez canal moment. The moment the British and the French realized they were not great powers anymore.
A majority of Americans support defending Taiwan from Chinese invasion. How much are the Chinese people willing to pay to attempt the largest amphibious invasion since D-Day while being starved of oil and international trade?
In the last 80 years the Chinese have become masters of holding diametrically opposed ideas. I believe they can back down from this without breaking a sweat.
https://m.youtube.com/watch?v=XjkoYlpf3EA
http://www.cnn.com/TECH/computing/9909/17/g4.ban.idg/
Not going to happen.
I didn't mean this to snarky, sorry!
At what power-compute capability is worthy of banning? Especially when we saw billions of cores deployed in mining ops...
Maybe there is a GLOBAL-HEAP (or whatever the original CRAY 'heap' sense was called) capability for billions of M1s...
-
SERIOUS QUESTION:
How much malware is mandated as being installed on any chinese product (irrespective that it may be ostensibly an 'Apple' product?)
How do we know that every iphone doesnt have super secret ccp shit?
I mean we fn arrested the HUWAI CFO for same...??
I dont trust shit.
I'm wondering what contingency plans US companies have for this? It seems to me like Apple would be screwed. Microsoft's hardware businesses would have problems. Google and Meta and Amazon are probably OK, though it may be tough to build and maintain datacenters, but everyone will be in the same boat.
This would mean catastrophic economic damage. Likely, would be far worse than the great depression. We important so many goods from China.
Err, you other points aside - that's how you'd contest a nail blockade, with an opposing naval force.
Saudis couldn't defend their oil assets from Iran drone attack for example.
[1]: NYT: How China Could Choke Taiwan (https://archive.ph/gS6qb)
Of course it will. US isn't capable of defending TW anymore. They may spoil an invasion, but there's nothing US can do to prevent PRC from trivially destroying infra the keeps fabs operating. You should expect the entire semi supply chain in East Asian (which supplies the world) to stop functioning if US tries to intervene on TW, especially using JP or SKR infra, nearly everythign critical is within 60m PRC missile range. The reality is PRC has much more leverage over global semi supply chain in even of war vs US who has leverage during peace via sanctions.
> major reliance
Are all in US in terms of high tech industry supported and actual patent fees extracted from production of semi. PRC will lose bunch of manufacturing jobs, but indigenous semi efforts will keep chugging along. Both will be reset to as far back as 1980s tech depending on how PRC escalates, which ultimately cripples US as tech leader more.
Lot harder to justify a new PC when no longer receiving a paycheck, though it will likely be even more important to have one then.
Just got a new System76 machine and I have to say, I'm really loving it.
Hard times ahead, unfortunately after that, it's even harder times.
For more than a decade money was more or less free and the only sane choice for investors was to pour cash into anything that might one day get big enough to magically flip the profit switch and make a huge amount of net positive revenue.
I thought the pandemic would be the end of the tech boom, but I was wrong. Not only did more money get pumped into the market but into consumers wallets as well. This led to extremely rapid expansion in the tech industry as investors were willing to pump even more cash into any company (hoping for a rapid IPO and a the ability to cash our there chips before the casino closed) and consumers had a surprising amount of extra cash lying around from being locked up for months.
Unfortunately reality has started to show it's ugly face. Inflation is starting to hit hard, not because of the money supply, but because we are hitting some real resource limits due to a combination of supply chain problems and the Russian war with Ukraine (which, to be fair, is also happening because of resource strain). These are not temporary causes of inflation.
This has caused the central banks to rapidly attempt increase interest to combat this inflation. Ending the era of cheap cash. This is coupled with the fact that things cost more so consumers also can't spend as much (even though they are also taking on more debt then ever, an unsustainable trend).
Sure inflation is cooling a bit but that's largely because of a drop in oil prices (at least in the US) which is a political play from the democrats resulting the record low strategic oil reserves in the US (since 1984) [0]. This cheap oil cannot be sustained and will end after the next election cycle.
The immediate impact of all of this is that direct to consumer companies are seeing an immediate hit (e.g. streaming service subscribers are churning faster than ever). Then non profitable ad revenue companies start to take the hit (see Snap's layoffs today). Next it will be non profitable B2B. But then we'll start to see how much of those profitable companies bottom lines comes form non-profitable companies.
This will have a cascade effect, since as I mentioned, it won't be as easy for us tech workers to find an new job with anywhere near the same TC. As a 200k+ engineer, consider all the crap you pay for without thinking of it because you don't have to. All the ubers and door dash orders that are surprisingly pricy to the average person, but really a brainless decision for any tech workers.
It's just getting started. If you work at a B2B tech company I encourage you to go poke around your internal data and look at how your customers are doing. I promise it's not great.
Then look at the state of venture funding [1]. Sure it's still above 2018 levels but it's rapidly falling. The start up scene is going to dry up quickly.
There's is more I could dive into but this is a good start.
0. https://www.reuters.com/business/energy/crude-us-emergency-r...
1. https://www.cbinsights.com/research/report/venture-trends-re...
+ with such goods if you are in the EU, you have up to 14 days to send them back and get a full refund, so basically you have 14 days to resell them double the price with no downside risk.
ideally you know an insider working there to reserve the products for you.
The headlines about Taiwan, I'm guessing, basically add up to a lot of desire by the media for eyeballs for ad revenue and a lot of face saving by Xi, that is, don't let the media and politician talk erode the long existing images and status of Taiwan.
Net, China has plenty to do now without making a big, real issue about Taiwan.
But there is no force guaranteed to keep Xi from making a mistake. E.g., he had China muscle into Hong Kong and violate the "one country, two systems" promise.
I'm guessing that if Xi does to Taiwan what he did to Hong Kong, then TSMC will be producing in the US and on their way out of Taiwan will junk their factories so that China gets nothing.
https://twitter.com/unusual_whales/status/156511407338962534...
https://twitter.com/quantian1/status/1476589270454390787
Similarly Rand Paul, if he tried to insider trade, lost money too.
https://twitter.com/quantian1/status/1425981488408010756
Remember, most of the time, the boring normal story is the accurate one.
That body language is pretty damning
This is just another form of reporting an executive "suddenly sold a bunch of stock" (when they filed a 10b5-1 plan about it months ago).
https://news.ycombinator.com/newsguidelines.html
The chances of nothing getting through is zero, but it’s harder to do at scale and much more expensive. Plus the US does go after foreigners that help targeted nations dodge sanctions so can make it very miserable for China’s smugglers further exacerbating the shortage in China.
https://www.sec.gov/Archives/edgar/data/1045810/000104581022...
Note a User-Agent header with an email address is required for accessing www.sec.gov.
https://www.sec.gov/os/accessing-edgar-data
Normally I do not send a UA header so I configure the localhost forward proxy to add one automatically for www.sec.gov
It does not appear that the SEC verifies email addresses sent in UA headers.
A suprising policy for accessing public information hosted by a taxpayer-funded organisation.
I am not a "bot", I just prefer not to use Chrome/Chormium-based/Safari or other graphical browsers. I use a text-only browser. Generally, SEC filings are just text.
It's a ceiling on the performance that can be exported. In an exponentially scaling industry, this is the equivalent of an announcement in advance of a complete ban on competitive products 2-3yrs out.
Given that both political parties have placed export limits on tech to China, it seems this is the new normal.
Edit: why disagree?
“Knowing where the chips go is probably a very good thing. You could for example, on every chip put in essentially a public private key pair, which authenticates it and allows it to work”.
hxxps://www.reuters.com/technology/chip-challenge-keeping-western-semiconductors-out-russian-weapons-2022-04-01/
What he won’t tell is that this is already a reality, as I learned after having my air-gapped system and Pixel phone wiped remotely for researching “silent speech interfaces”. There is no security when silicon trojans are in all devices.
intel SGX already involves a per-chip keypair with intel as the root of trust. it just isn't required to use the rest of the chip
https://news.ycombinator.com/item?id=32394353
> Eric Schmidt, the former Google chairman, told Reuters in a recent interview that high-end processors should have kill-switches.
“Knowing where the chips go is probably a very good thing. [...]
I truly believe you're very concerned about all the things you talk about, but if you want to convince people to also believe then there needs to be more from your side.
And no, suggesting that people perform ptychographic x-ray laminography on their chips isn't a practical suggestion.
People with willingness and means can find me anyways.
"It's not technically hard to make a device that complies with the FCC that listens to nonpublic bands but then is quietly waiting for some activation trigger to listen to other bands," said Eduardo Rojas, who leads the radio spectrum lab at Embry-Riddle Aeronautical University in Florida. "Technically, it's feasible."
To prove a device had clandestine capabilities, Rojas said, would require technical experts to strip down a device "to the semi-conductor level" and "reverse engineer the design." But, he said, it can be done.
https://www.cnn.com/2022/07/23/politics/fbi-investigation-hu...
One way it can be done:
Ptychographic X-ray Laminography; No trade secret or hardware trojan can hide! https://spectrum.ieee.org/chip-x-ray
Edit: I knew this would get downvoted quickly. Hey, mr or ms judge of truth, how about you check what nvidia services run on your machine before wielding all that power.
https://itigic.com/minix-the-hidden-operating-system-in-inte...
"As for Pentium PCs... well, they're harmless."
without that resulting in lawsuits.
Somehow I feel that in our world the majority of corporations don't make direct comparisons to their competitors that often, preferring to use "competing brands" or something of that variety instead.
Hard to see how Intel could sue over a statement like that.
The reason many companies say "competing brands" is because they don't want to mention an alternative, because it increases the mindshare of the other brand.
This is not really new but NVidia was caught by a surprise change in the limits.
Assuming said steel is actually of the quality it's certified as [0]
[0] https://asiatimes.com/2017/10/nuclear-tentacles-kobe-steel/
https://www.ecfr.gov/current/title-15/subtitle-B/chapter-VII... https://www.federalregister.gov/documents/2006/04/24/06-3647...
You'll probably have to dig quite a bit to find the latest rules and regulations.
These regulations go back to at least the Export Control Act of 1968. Every CPU/GPU maker would have a legal team that understands these rules.
https://www.bis.doc.gov/index.php/regulations/export-adminis...
For example, high resolution analog to digital converters have restrictions. In the past strong cryptography even had restrictions:
https://en.wikipedia.org/wiki/Export_of_cryptography_from_th...
Even travel for US citizens was restricted. I have relatives who were in the semiconductor and aerospace industries during the 1980s. They were advised by the government to not attempt to travel to the Eastern Bloc. Some kinda related reading:
https://academic.oup.com/dh/article-abstract/43/1/57/5068654...
https://en.wikipedia.org/wiki/Bernstein_v._United_States
(Because the opinion was withdrawn, it's "persuasive" but not "binding".)
(What really happened was that the Feds revised the definition of "supercomputer", and suddenly the G5 configuration they were using didn't qualify anymore... it had nothing to do with anything Apple did, except perhaps lobbying for it so they could build computers in China.)
They made a big deal about it in their commercials at the time: https://www.youtube.com/watch?v=OoxvLq0dFvw
https://www.eff.org/deeplinks/2015/04/remembering-case-estab...
My favourite conspiracy theory is that this was done deliberately by the US, since Xeon Phi was a large pile of steaming turd, so the Chinese wasted money on a machine with high theoretical FLOPS but crappy real world performance, and also wasted the time of their scientists and programmers who were porting code to a programming paradigm that went nowhere fast.
Other examples can be found here: https://www.wassenaar.org/app/uploads/2021/12/Public-Docs-Vo...
Yes, in the 70s and 80s, it made sense to say that 5-axis NCs systems that can cut Inconel turboprops for ICBMs should be export controlled; those were multi-million-dollar flagships of the biggest and best manufacturers and were key to the country's military edge. Now you can get one shipped from China for a couple thousand dollars, with servo loop times, encoder and linear scale resolution, memory capacity, processor speed, axis counts, etc. that are better than the originals that cost a thousand times more half a century ago.
Today, it's true that the A100 is a cutting edge, shockingly powerful, uniquely capable processor. You can do machine learning/computer vision projects with it that you can't do with any other technology. Eighty gigabytes! 256 GB/s per package! Almost 20 teraflops of FP32! Compared to the GPU in the desktop you bought 8 years ago, yes, these are huge. But if you plot interconnect bandwidth, memory size, core counts, price, and so on, and extrapolate forward... yeah, your little hobby Kubernetes cluster of 2030-era Raspberry Pi equivalents will probably be at a performance level that's export controlled.
China not civilized?
And before human rights becomes the main talking point: Guantanamo.
Civilized or not is up to your perspective.
Even if we did, how do you measure up acts done in these places against one another? Is making license plates for pennies in the US better or worse than making t-shirts in Xinjiang? Is being wrongfully thrown in prison due to a prejudice judge seeing your only skin color worse than being thrown into prison for being Muslim who refuses to shave their beard?
Let me know when anyone is actually allowed to verify what is happening there, until then comparisons are useless.
FYI: Things like fast food franchise remodels regularly have American forced labour doing a lot of the CAD/CAM work. If you eat at American franchise restaurants, you support American prisoner slave labour. [edit] here is a pretty government website promoting you to use slave labour: https://www.unicor.gov/Category.aspx?iStore=UNI&idCategory=1...
here is the US Constitution 13th Amendment banning slavery:
“Neither slavery nor involuntary servitude, except as punishment for crime whereof the party shall have been duly convicted, shall exist within the United States, or any place subject to their jurisdiction.”
You are very generous with your interpretations, I admire your optimism in how civilized comments in Internet are.
> Please respond to the strongest plausible interpretation of what someone says, not a weaker one that's easier to criticize. Assume good faith.
Along with this one:
> Be kind. Don't be snarky. Have curious conversation; don't cross-examine. Please don't fulminate. Please don't sneer, including at the rest of the community.
They are all honestly pretty great: https://news.ycombinator.com/newsguidelines.html
See, for instance, on this page where Hume speaks of civilized versus uncivilized forms of government: “This chapter examines Hume's conception of government. It considers three forms of government that Hume distinguishes: barbarous monarchy, civilized monarchy, and free government (with its two subdivisions, limited monarchy and republic).” So civilized is contrasted with barbarous. https://academic.oup.com/book/12307/chapter-abstract/1618482...
You'd think with SMIC demonstrating 7nm production ( as detailed in depth in this video by Asianometry: https://www.youtube.com/watch?v=dQGnwKBxAKk ) you'd think that the US had gotten the hint that the tech cat is out of the bag – or the tech genie has escaped the lamp or whatever metaphor you prefer … A comparison of SMIC's 7nm process versus the competition: https://www.techinsights.com/blog/smic-7nm-truly-7nm-technol...
My prediction is that SMIC joins TSMC, Intel, Global Foundries, and Samsung at the top tier of semiconductor manufacturing by 2030, along with an ecosystem of Chinese design and manufacturing companies. In doing so the West's stranglehold on other companies like Huawei will be over. If Russia can survive the punitive Western sanctions until then it'll no longer be hampered by lack of access to high-end Western technology.
I don't know what this means. Would you explain?
Parent poster likely worked for a US company that was doing a customer upgrade abroad (i.e. exporting US technology) that would have tipped the cluster compute capacity above the permissible limits, unless some older nodes were decommissioned and destroyed.
Will be interesting to see if Nvidia challenges it.
I wonder if Nvidia can use a similar loophole where the chips are clocked at lower rates for export to meet the law but can be "overclocked" with hardware modifications.
I work in CPU design and I wouldn't risk it.
That’s why an individual contributor should not take part in it. It’s better to get a different job.
This is just the first of many ITAR violation stories that I found when doing a quick search. ITAR violations are regularly enforced.
[0] https://www.jdsupra.com/legalnews/businessman-sentenced-to-o...
Yet ERA-2SM for example had more gain on 10GHz than the 8GHz where the official spec ended at. :)
You'd probably have a hard time claiming that was "peak performance" for the chip if it's underclocked with "nudge-nudge-wink-wink" overclocking available.
Not to mention, H100 is itself made in China and it seems Nvidia is allowed to keep building it there[2].
[1] https://www.reuters.com/technology/chip-challenge-keeping-we...
[2] https://www.cnbc.com/2022/09/01/nvidia-says-us-government-al...
they don't have to achieve A100 levels in the first go, even Rpi level chips should be enough to give them experience
If china had the layout, they would just get it made in taiwan.
Though Nvidia has probably already sourced some niche provider at a much higher cost elsewhere in anticipation.
Worked on a three 50U rack recently with quad T4's per 2U and we're not even through 10% of that order yet.
We build servers, test them, and ship them (in the US). No fucks given on the technical superiority of the `A100/H100` than parts that may be a generation older.
So when you're playing nation-state hard-ball, since the architectures are _somewhat_ trivially copied upon being published, that means the best checkpoint goes to the one with the fastest GPU's. If China wasn't behind in deep learning, they will be.
Fabbed at TSMC. (I guess we are about to see if SMIC can do their 7nm at scale.)
https://www.nextplatform.com/2022/08/25/china-launches-the-i...
CHINA LAUNCHES THE INEVITABLE INDIGENOUS GPU
…
Here is how Biren says it stacks up on various machine learning workloads, pitting the BR100 against the Nvidia A100:
…
We presume this is for AI training workloads, not inference. The average speedup over the A100 is around 2.6X. It is not clear if the Nvidia machines were using sparse matrix features, which doubles the throughput, or not. Our guess is they were not.
https://www.servethehome.com/biren-br100-gpu-for-datacenter-...
https://www.tomshardware.com/news/baidu-unveils-kunlun-ii-pr...
Now China announces a powerful GPU.
https://videocardz.com/newz/hackers-now-demand-nvidia-should...
https://www.nextplatform.com/2022/08/25/china-launches-the-i...
The grand tour.
I've never heard of this term before. Anyone know what companies it's supposed to include?