‘Better Yield on 5nm Than 7nm’: TSMC Update on Defect Rates for N5
anandtech.com
anandtech.com
Its kind of sad we used have guys like shockley, moore, kilby etc. The reason why silicon valley is called silicon valley. Now we are just falling behind Taiwan, south korea, and maybe china.
TSMC margin was always at around 30%, they are killing it.
... and invest in making their own chips from end to end in a couple of years. Apple has proven that they do not hesitate to completely take over stuff with the recent Intel-ARM switch when the quality or price does not match what they perceive to be fair.
TSMC is in an extremely powerful but also very precarious position - they are worth ~400B $ in market cap, and Apple has ~200B $ in cash reserves and a market cap of 2T $. If Apple wants they can either buy up TSMC whole or they can set up a competitor.
=)
Even if TSMC decided tomorrow to sit back and extract margin, GloFo could still probably get back in the game. That ability decays with time, of course, and that decay is what TSMC buys with their low margins. They are suffocating their competition to build a moat. It's a much better strategy than "extract margin and make big bets on fancy new tech to stay one node ahead," with Intel as Exhibit A.
A) spend tens of billions of dollars on a new fab facility B) shut down production in one of the existing facilities while spending billions of dollars to upgrade it to 7nm
While not having the volume to be able to court the largest customers who are the most interested in 7nm.
They decided to stick with their existing 14nm infrastructure and specialize in the long tail of smaller customers who want custom chips and don't need a leading-edge node.
They have the money and can attract the talent
They only build their own stuff when they see a competitive advantage to doing so — when IBM stopped delivering on chip efficiency or when Intel’s chips started stagnating.
Large Nvidia chips might not be able to move quickly, but all of the MCU, WiFi, etc chips would migrate to cheaper fabs like SMIC. Which reduces TSMC's revenue, which reduces their ability to invest in new processes. Cents (and even fractions of cents) matter to those chips.
It does, and that's not only TSMC, or even fabs as such.
Taiwan is a source for a many single vendor semiconductor manufacturing equipment, and consumables. Taiwan going down means the semiconductor industry as such, globally, going down for a few years at least.
First thing that comes to mind are lithography machines, but for the very high-end ASML is the only game in town currently afaik. And they are sitting in the Netherlands (working with Zeiss in Germany). For larger feature sizes there are vendors too.
Clean room material markers, few makers of vacuum grade plastics, FOUP, film, and other carriers.
They will have to go back few generations in the resist tech to resume production with non TW suppliers.
And like that with many, many other parts of the ecosystem. It's only a tip of the iceberg.
Is this an abstract example, or is this actually the case?
Except completely reclassifying it to remove the thing that made China want to seize control in the first place: special trade status. Other than that, what could the current administration do that wasn't screwed up by its predecessors?
Instead, Trump has spent so much of his time kissing up to Xi to get his help to win the re-election (per Bolton's book). And check out some tweets from 2020: "Terrific working with President Xi, a man who truly loves his country." "In particular, on behalf of the American People, I want to thank President Xi!” "“Just had a long and very good conversation by phone with President Xi of China. He is strong, sharp and powerfully focused on leading the counterattack on the Coronavirus."
"Of his many tweets over the three-day Memorial Day weekend, when the Hong Kong issue was at the top of the news, none was about Hong Kong. During the Hong Kong protests of 2014, Mr. Trump tweeted one of his few clear statements on the plight of the territory: “President Obama should stay out of the Hong Kong protests, we have enough problems in our own country!”" https://www.nytimes.com/2020/05/27/opinion/china-hong-kong-l...
This is literally, exactly what the White House did.
It isn't enough, we need to do a lot more.
[0]: https://www.ft.com/content/fcaf0a57-ac0f-4695-af0f-82a44e6ae...
The UK and Australia are willing to flap their mouths a bit [1] over it at least, and the UK is much closer to actual sanctions (Australia has, for the time being, not terminated their special trade arrangements with Hong Kong).
[1]: https://americanmilitarynews.com/2020/05/us-uk-canada-and-au...
Ultimately only the U.S. has taken any leadership on this matter, even if I think it will be inadequate. There is more to be done; but with the U.S. Presidential election looming, one thing that would steel the resolve and buff the resources of PRC assets in American politics would be a major announcement.
Nothing is as simple as the free world bullying the nominally-free world to condemn the only major state in the world conducting a holocaust; because apparently those "allies" don't give a toss.
What can any Western country or US president can do? Do you declare a war with a country over a piece of land that is rightfully their?
I don't understand the logic behind "without a peep from the US"? like what? a WAR with China?
PRC signed an agreement with the UK that set out the legal parameters for this transfer of sovereignty, and PRC is contradicting that agreement with their actions. While it may not legally be the UK's to keep, it was not China's to take in this respect either.
> What can any Western country or US president can do? Do you declare a war with a country over a piece of land that is rightfully their?
Well, in the mean time, the U.S. has terminated the special status of Hong Kong which allowed it to serve as the conduit between the civilized world and a country with essentially no rules that regularly outweigh pure class/political privilege.
The English legal system and liberal order that the UK imposed on Hong Kong was in large part what made them so dazzlingly wealthy and dynamic; and if PRC wants to corrupt that, the result is that the benefits of the system they are corrupting disappear with that system.
Also, it is not always right to honour the law, when dealing with bandits. The CCP does not care for the law, so their assertion of a legal right to Hong Kong (which has not matured anyway) is hypocritical, and can essentially be ignored at no moral hazard.
Furthermore, even if we asserted that the CCP has some agreed legal right to Hong Kong that will mature in some years, I would not agree that it has a moral right to take its people as property.
Given that China has not always existed, how am I to interpret this statement?
> UK gave back a land that wasn't their to keep.
Portions (though not all) of Hong Kong were ceded in perpetuity. They simply were _not_ part of China before being given back.
HK isn't a sovereign nation, Taiwan is. Also, having the US go in would be questionable. Wasn't the "one country, two systems" treaty an agreement between uk and china?
Depends on who you ask.
I don't think you mean Native Taiwanese (as in Native Americans) here, so I have to say people are not unanimous about whether "Taiwan" is a sovereign nation.
I mean, you can make the same argument with "the earth is flat... depends on who you ask". At least when it comes to the Taiwan (the island) I don't think there's any doubt that the ROC government has a monopoly on violence there. They also have their own military, police force, and collect their own taxes, independent of the PRC.
Obvious a lot of time has passed, but it's not quite flat earth. It just requires a sence of 'historical justice', however unhelpfull
The U.S. started removing the extra trade agreements from Hong Kong (which arguably is a large part of what made Hong Kong what it is today) and sanctioned individuals in the Chinese government [1]. While this seems to be a "weak" response, I'm not sure what else I would like to see short of getting the military involved.
I have no idea. There's a reason I make my living as an engineer rather than as a politician or a diplomat.
The UK is trying to welcome people from Hong Kong to the UK (passports left over from before the handover) which could hurt them where it actually matters if lucky.
The original seisure of Hong Kong was an act of gunboat displomacy, and not exactly an exemplar of justice and law. Consider how its seen in donestic politics in China.
This does not mean that I approve of China's activity, just putting things in perspective.
Except the UK has much bigger issues ATM like dealing with the social, political and economical fallout of Brexit and Covid-19.
To put it mildly, even if they wanted to, it's tough for them to help put out a fire in a far away village when they have a huge dumpster fire in their own back yard to deal with first.
Obama promised to end wars close down Guantanamo bay and in the ended up keeping up all war campaigns and on top of that he started droning people around the world.
The 'big boys' of the industry have enough fingers in the pie(s) to have their say in things that would affect them.
And independence of Taiwan is definitely one.
> Hong Kong fell without a peep from the U.S.
Because it was mostly Chinese businessmen window to the world, is fall of independent HK really that big of a deal for US or UK?
Nobody is going to lift a finger for Taiwan over (checks notes) slightly smaller transistors.
We had an audition for disrupted supply chains just this year. It turned out hardly any of it really mattered.
Pretty much any company that imports stuff from China would have already had additional supplies built up to cover the normal CNY shut down and ramp up so the supply chain issues were lower than if COVID had shut down China at a different time of the year.
I guess the real danger here in our unforeseeable world will arise if the American government becomes either unwilling or unable to defend Taiwan from a sea invasion from the CPC, something that would probably be more serious to American (and world) security than losing a few fabs in Asia, I think.
That's not the case since 1991.
With China about 100 miles from Taiwan, I don't see how anyone could get close without mass casualties. China has the shorter and more secure supply lines. They could send an invasion force using numerous small water vessels in waters they could easily defend. In the end, the losing side would just destroy all the important places in Taiwan.
So yes, I think the conflict will definitely remain frozen for the next decades, and it will be at least until either the CPC loses control of the mainland, the USA lose their grip on the region, or some US politician goes bananas and pushes for Taiwanese independence.
Russia for example still keep tries to have a 100% internal military semiconductor supply chain. Their best chips made on a domestic process are 10μm. TI military foundry is not much further ahead.
I doubt that any actual "strategic planner" actually put his hand on any American policy making process in the last 30 years. The way how terribly unprepared the US was to even such an insignificant nationwide emergency like the virus, is the best proof.
And we are only talking about digital chips. Analog chips case will likely be hopeless without a specialty analog foundry.
In the peacetime, it is very, very unlikely they will be able to "reorient" to make defence chips, as it will seriously disrupt the company.
And in the end, it may not matter at all if the supply chain for materials will go down in case of a military conflict.
TSMC's ties with the U.S. are, on the other hand, a big strategic threat to China, and maybe to other countries that aren't the U.S.
TSMC isn't real bottleneck here. It's US semiconductor manufacturing equipment firms fully under US jurisdiction that underpin the supply chain for any modern fab.
https://www.nextplatform.com/category/hyperscale/ has interesting stuff on uses and development you don't see from a consumer POV
Intel, AMD, Nvidia, Lam Research, Applied Materials, KLA, GlobalFoundries, Broadcom, Apple, Xilinx, Maxim Integrated (now part of Analog Devices), Linear Technology (now part of Analog Devices), Marvell, Cisco, Cypress, Google, Cadence, Synopsys, Western Digital, Seagate.
also, why would this make sense:
> of course, excepting Apple
Apple is probably the single largest customer of TSMC, and their products depend on their semiconductor design team (especially true now that they are moving to in-house CPUs for laptop/desktop). Why throw them out from a discussion of companies that care about fabs?
But if you ask the question of which sectors the Bay Area has the largest investment and footprint in, then semiconductors would (unfortunately) be only a small part of that. That uncomfortable truth (which is exactly what I'm trying to point out in the above comment) is that market valuations today seem to be driven by the things built on top (in particular, the "data economy" and the effectiveness of personalization-driven consumption increases), rather than (for eg.) the semiconductor engg which is crucial to build those platforms.
An analogue of this is when the Ford CEO speculated in 2018 to the effect that it might be more profitable to sell customer data derived by selling cars, than from selling cars in the first place.
1) government support - i.e. the national government saying "we want this sector to succeed" and providing funding via grants/loans/subsidies, (or by fiat w/r to PRC)
2) actual labor costs?
3) talent, i.e. American-trained Chinese/Taiwan/Korean talent coming home, then generation 2.0 starting up research at Chineses/Taiwan/korean universities?
Assuming a long-term visionary US government wanted to compete; would we have to fund fab research, development, and capital i.e. similar to the pentagon funding Newport News shipyard to keep strategic know-how alive in the US market?
2) labor cost: probably very insignificant. While they employ high-paying high-tech workers, it's a capital-intensive, not labor-intensive, business. Both TSMC and Samsung spend over $10+B annually on CAPEX.
3) brain drain?: Morris Chang worked for TI for decades before founding TSMC. He pretty much started the outsourcing business in the late 80 single-handedly.
Sure. I don't agree with Trump on a lot of things, but he's at least doing something to bring them back.
However it is utterly inconceivable that anyone would “go back” to using some old social media platform.
Anyway, my point is that having faster and more energy efficient chips would make the training of stuff like GPT-3 cheaper.
I haven’t reacted my compuserve account, but I’m fine using email, IRC, phpBB, mailing lists...
I was really talking about, for example, Vine. You'd rather use TikTok than Vine. There might not actually be any social networks from late 2016 no one uses today that I've heard of, so it's hard to give an example that looks back to when 14nm was the typical node size for phones. Which is sort of my point, it is in some strict sense of the word literally inconceivable to go back. Whereas I remember what an iPhone 7 was like, and it's not that big of a change.
These are insane numbers.
New GIGAFAB takes time to build.
Specializations change over time. The US, specifically Silicon Valley, has different specializations today than it did in the 60's. A lot of the semiconductor specialization moved overseas, costs came down, and in general the world is richer. Other specializations include things like China and injection molding, Japan and mechatronic technologies in the 80's and 90's, Thailand and lenses, etc
Obviously the protectionist crowd would take issue with all this, but free trade theory is studied and applied all over in trade agreements, specialization zones (like Shenzen), etc. Generally speaking though, the cost of attempting to be self sufficient in everything (like semiconductor manufacturing) means closing borders to competitors outside, which eventually leads to inferior product. With robust competition, the consumer usually wins on both quality and price. Obviously there are other consequences, such as job migration etc. In general though, everyone is lifted up and made richer over time (evident by the fact that there is quantitatively less poverty in the world today and the 1960's and even the most poor can enjoy things like cellphones, leather jackets, and shoes that fit well that even the kings of a century ago could not). The system isn't perfect, but atleast in my eyes it is the one that does the most good for the most people.
This is an interesting issue. The existence of national borders and the restriction on the free movement of people reduces the efficiencies achieved with specialization. Imagine how more efficient it'd be if Taiwan had easy access to US semiconductor talent. Similarly, it'd be more efficient if the seasonal migrations that help agriculture didn't happen and those people actually lived near the crops they help maintain.
I mean it's more complicated than that, most seasonal migrations of agriculture workers happen in large part because the seasonal workers come from areas with lower standards of living and lower cost of living. Partially, they are attractive as workers because they are able to take lower wages than the people who live where the crops are grown.
When I first started my career in software in 2000, you could hire 6 Chinese engineers for the price of one American engineer, or even 2 for the price of an American high-school student. Nowadays, the good software engineers in China (the ones who work for Google etc.) get paid just as much as their American counterparts. Similarly, the good American engineers have largely maintained or improved their salary - hell, L7 compensation at Google (and the equivalent at Facebook/Snapchat/Apple/Lyft/etc.) is ~4x the highest compensation I remember seeing 15 years ago. Wages for mediocre software engineers (say those whose only skill is building PHP webapps by copying code from StackOverflow) have marginally improved in China and have fallen significantly in the U.S. We replaced a geographic wealth gradient with a skill-based one, which IMHO is significantly more fair but probably doesn't feel like it if you were on the losing side of the gradient.
Not necessarily, or indentured servitude would have been impossible.
> Nowadays, the good software engineers in China (the ones who work for Google etc.) get paid just as much as their American counterparts
This seems to be more of a result of the low friction of hiring a good remote developer. It's also not completely true - a top developer in Berlin or Dublin (which is where I live) makes less than the same developer would in the valley (I know that because I was offered to move) and a lot more than a developer makes in Brazil (where I'm from). I am not aware of any remote developer who gets paid what they'd be if they were living in London or San Francisco.
That's not true. I can provide a data point: a Googler who works in Beijing, 3 years into his career with 2 years at Google[1], get paid roughly 72k USD per year. I don't think this is even comparable to the compensation level in Valley. That said, if you compare with the Japan or Europe (except Switzerland) counterpart, it does seem to be on the same level, so the outlier here is US.
[1] To be fair this does not matter because he is stuck at L3 due to works nullified by internal protests.
If there was a year round harvest they wouldn't be migrant workers. It brings to mind a "silly stopgap" solution mix of paying people in low wage countries to operate farming drones to subsitute for the AI in spite of the cost of building the body being higher and the infastructure and tools required to control a bot not corresponding with migrant farm worker origins.
My bank account tells a different story. Only very few people got richer. General purchasing power fell off like cliff in the US after the 70s when all the outsourcing happened. There is nothing positive about globalization. The bad effects are somewhat mitigated by a massive amount of technological progress.
> closing borders to competitors outside eventually leads to inferior product.
There is no proof for that. With close borders you can simply steal the intellectual property of your competitor and match the quality within a year. The Chinese are doing this for decades.
> In general though, everyone is lifted up and made richer over time
This is quantifiable wrong. Maybe you have to get out of your filter bubble. If anything globalization leads to more inequality. The statistics about how the top 1% make all the money whereas everybody else loses are neither controversial nor new.
Globalization does increase inequality, particularly in countries with a lot of capital, but it also makes the very poor better off. What the US calls the middle classes do stagnate, though, and capitalists are the biggest winners - which, don't forget, also includes pension funds as a significant fraction.
Most of the top 1% are wage slaves like everyone else. The top 1% by income isn't what's interesting; it's the top 1% by assets. Wealth inequality is much bigger than income inequality.
So the reason why Chinese masses are so much better off today is this very "one-sided" version of globalism. You might as well call it nationalism.
Which is more or less what the US did with China.
Nationalism made China rich. Not globalism.
On the other hand wages in China have shot up and the huge pool of cheap rural labour is mostly tapped out. Low end manufacturing is moving to other countries, Samsung now does most of their smartphone final assembly in Vietnam but those countries are smaller. That trend is playing out it's end game. Employment in the US and Europe was rising, and came close to 100% before the Pandemic.
So I actually agree with your criticism, but not that there's nothing positive about it. Things change, overall I think things are changing for the better but that's never a painless process. That's why although I'm a believing market capitalist, I also support social programs and universal health care. If corporations deserve bailouts, so do ordinary citizens. We're all in it together.
I would have no problem with that if it didn't happen at the cost of American worker which it clearly did.
> Employment in the US and Europe was rising, and came close to 100% before the Pandemic.
You can thank the "nationalistic" trade policies of Trumps USA towards China for that. In Europe on the other hand (which had no such policies) unemployment especially in the periphery was and is still sky high and the situation was looking pretty grim also before the pandemic.
We have a long way to go before the US system feels like a a kindly helping grandfather rather than an abusive angry father that blames you for not working harder.
As for globalisation, the US became the dominant global economy how exactly? Protectionism, aside from being an autocratic intrusion into the freedoms of citizens, just doesn't work.
Take the US steel industry, a main target of Trump protectionism. Yes blast furnace employment is marginally up, but steel mill employment has been hammered because it largely depends on working imported steel. The domestic furnaces can't expand fast enough to meet supply, and don't produce some of the steels needed by manufacturing anyway. As a result manufacturers consuming steel are facing inflated prices and reduced competitiveness. It's unintended consequences all over the place.
In Asian countries, i.e. Taiwan, China, SK, Japan, Vietnam, as well as, say Mexico after NAFTA/NAFTA 2.0, what are the numbers regarding general purchasing power?
As we've seen with Brexit and Trump, nationalism trumps economics. America is having a hard time tolerating a growing China, and frankly, China is worth opposing. It had been hoped their growing middle classes would lead to demand for a more liberal government, but that hasn't happened, and instead the CCP sees Western liberal ideas as a kind of alien mind virus which must be extirpated. Another cold war with local hotspots looks almost inevitable.
The most aggression we'll see is covert secret service actions, disinformation campaigns, cyber war, etc. They'll never dare drop real bombs.
See e.g. https://www.storm.mg/article/2971332
Nationalism in its modern (post-1865) form is a product of technology + economics. The U.S. before 1865 was a relatively lose confederation of states: this is enshrined in the 10th amendment (now largely forgotten, but which reserved all powers not explicitly delegated to the federal government or people "to the states"), and indeed in the word "state". Same thing in other nations like Germany and Italy: before 1871 these were independent states like Prussia, Bavaria, Saxony, Hanover, Brandenburg, Westfalen, Bohemia, Sardinia, Tuscany, Venetia, Florence, etc.
The driver for national unification was largely industrialization. New technologies like the railroad, telegraph, steel-making, mass production, etc. gave large advantages (particularly on the battlefield) to larger political unions that could draw on the labor of many regions and specialize. Political bodies that didn't adopt these technologies got rolled. The North likes to view the American Civil War in terms of slavery, and the South likes to view it in terms of states rights, but fundamentally it was a clash between the industrial system in the North vs. the plantation system in the South. And just like Brexit and Trump are doing now, the South fought viciously to preserve their self-determination and way of life, but the industrial system was just more efficient.
Right now, we're seeing protectionist lip-service. My iPhone is still made all over the world, my employer still makes a majority of its money in foreign countries, I still have to compete with foreign buyers when buying a home (though it's better now than it was in 2015-2018). It'll be interesting to see what happens if there's a serious threat to that. We had a previous era of globalization in 1900-1914 that ended with WW1 - but arguably that was because the world wasn't ready for it, and globalization was entwined in the lives of the rich but not the common people. Now there's a significant number of people who are largely unaware of how dependent their lifestyles are on globalization, and I'm curious what'll happen if it's actually threatened.
Take a good look. Back in 2012, people were genuinely thinking that Zuckerberg was going to run for the election. And they rightly though that if facebook was for example post "Go Vote Zuck" on the frontpage, he might've well won.
Minor nitpick, but the Confederacy viewed slavery as the cornerstone of their government[1]. Those who now claim the Civil War was about 'states rights' are engaging in simple, brazen revisionism.
> In what’s now known as the “Cornerstone Speech,” Stephens told a Savannah, Ga., crowd in 1861 that “our new government is founded upon exactly the opposite ideas [as those of slavery foes]; its foundations are laid, its cornerstone rests, upon the great truth that the negro is not equal to the white man; that slavery, subordination to the superior race, is his natural and normal condition. This, our new government, is the first, in the history of the world, based upon this great physical, philosophical, and moral truth.”
[1] https://www.csmonitor.com/Books/chapter-and-verse/2015/0708/...
This has implications for how the future could look that go far beyond what seems most "reasonable", what looks most like the past and what will make each individual person feel best. It doesn't necessarily have to be war - it could also happen through peaceful but intense competition and trade.
And even if you take the rewards of comparative advantage and redistribute them to account for this, you are still destroying the social fabric of families and communities. Vitality, self-respect, and good values come from being engaged in productive work.
This is very comforting to regions that are specialized in high-cost, high-margin, high-technology products, and not very comforting to regions that specialize in low-cost, low-margin, low-technology products. The countries growing the bananas that go into your smoothies have specialized in growing bananas, but they aren't making any money from it.
If they made great smoothies (in addition to growing bananas), then they wouldn’t be leaving that market to blond Californian influencers living in Bali.
This happened for bananas going into the smoothies, rubber going into tires (at the time), coffee going into lattés, oil going into tires (at a later time), etc
Low capital banana growers would do far worse if they tried to compete by growing grain crops in the region as they have to contend with high capital mechanized and chemical using operations with massively better yield and closer markets.
Specialization is at least a local maximia to them even if there are other options better in the long term. Take say West Virgina's coal mining vs a more diversified industry.
"We" (the US) still lives in a violent world, or at the very least a slow, soft violent world where soft power might cut us off from Taiwan.
I find the fact that one half of the "techno-elite" crowd thinks that giving up on hard problems is a long term feasible position. We must be able to manufacture and produce the absolute best fundamental technologies from a strategic standpoint, even if it isn't a core, market efficient concern.
The rest of the world is modernizing, dealing with modern problems (disease spread, high end technical manufacturing, green technology), and we can't even get out of our own way. Frustrating.
Don't get me wrong, I understand the cynicism here. There is no perfect system, but I just don't know of one that has empirically produced better results. Certainly not socialism or communism.
Capitalism is not a singular thing, economic policy of the 30s, 60s and 2020s is very different. There are various schools of though: Shumpeterian, kensyan, neoliberal, austeian, etc. Lately neoliberal system has been dominant, and delivered remarcably poor results for most of us.
We should be discussing the economic system we currently have and various ways it is / was / can be structured. Even the justice system, the cornerstone of civilised society, has been corrupted.
For example why is it that we have publicly funded police but not publicly funded lawyers that can actually hold firms to account for contract violations? Why do companies that pay no taxes enjoy priority when it comes to public prosecution over regular Joe. Why is it, that if I owe the state a few grand, I will be hauled before a court within a day, but if BigCorpX owes the state a few billion they get to negotiate a special tax deal with massive 'discounts'? The list is truly endless.
War. Wartime has always been a fantastic generator of innovation. It unfortunately destroys mass amounts of capital and kills and wounds many people in the process.
> wealth, and prosperity
Charity and taxpayer-funded government investment. The Marhshall Plan dramatically improved the prosperity and efficiency of much of Europe and was paid for by US taxes. The Green Revolution was funded by the US, UN, and Rockefeller Foundation.
Capitalism is a valuable tool, but too many accolades are laid at its feet these days.
I fear as a country we are not capable of that type of thing any more and I am not only frustrated, I'm saddened.
BTW if anyone has any examples of more recent (post-2016) successes of a similar magnitude and impressiveness, I could definitely use a mood boost.
It's hard to assign an achievement to a specific country unless it's a government project.
I'm not a Tesla fanboy, but Tesla is an American company.
Also historically speaking, America (compared to some other governments and policy bodies in the world) doesn't generally over regulate or adopt policies on young industries. This is generally in the spirit of not quashing broad innovation too early.
To your point though, we are indeed struggling with modern problems. Many of our modern problems though are partially due to modern marvels. I for one am optimistic we will make progress in some of these areas.
Are you sure about that? "Total global trade" seems unstable to me. Look at the coronavirus, and countries mostly unable to produce their own masks, having to import instead from far away countries — but that country sold to the highest bidder primarily. Here were I live, I couldn't buy any masks (back in April and May).
> the cost of attempting to be self sufficient in everything (like semiconductor manufacturing) means closing borders to competitors outside
Can that be done differently? A country can have a small amount of domestic production, to keep knowledge and tech available in the country. Whilst importing most things from abroad, because of the benefits you mention with global trade. — Then you get both stability in case of crisis, and global trade efficiency? And more local jobs
In regard to the masks, examples exist and the world isn't perfect. No doubt it will be a business school case study for decades to come.
As far as personal anecdotes (because seeing numbers and seeing it with your own eyes are not the same), I have been traveling for more than three decades to developing countries and have seen the changes, the access to medicine, clothing, school supplies, food availability. From the developed world, we would still call these people poor, but the quality of life for the same class of people over a few decades has changed dramatically.
> Can that be done differently? A country can have a small amount of domestic production, to keep knowledge and tech available in the country. Whilst importing most things from abroad, because of the benefits you mention with global trade. — Then you get both stability in case of crisis, and global trade efficiency? And more local jobs
Just run through the thought experiment. It can be done, but that small amount of domestic production needs to find either subsidy or a customer that must have things made domestically (like maybe a government). Subsidy means the taxes you and I are paid are allocated to create sub-efficient industry. Domestic production needs, while real for purposes of national security etc, are tricky. We have seen in recent years Trump (his predecessor too), using executive power to force things to happen on the basis of national security. Some are fine with it, other call it abuse of power. My point is just, it is messy.
“At the insistence of the United States, in 1986 Japan agreed to limit its exports of semiconductors, mainly the "dynamic random access memory" (DRAM) chips, to America. These chips are used in high-tech consumer electronics equipment like computers and video cassette recorders. The agreement expires this July 1, and the Bush Administration thus soon must decide whether to renew it. Doing so would make Washington a hypocrite in its free trade efforts to open markets abroad for American products. The 1986 chip agreement, after all, restricts trade, ostensibly to help some American segments of the semiconductor industry. The agreement in fact has harmed American computer manufacturers, who have found themselves paying higher prices for computer chips. This makes American computer manufacturers less competitive and drives up computer prices for all Americans.” [by Bryan Johnson, published January 24, 1991 – 20 min read]
!! @ “high-tech consumer electronics equipment like […] video cassette recorders”
https://www.heritage.org/asia/report/the-us-japan-semiconduc...
It would be an interesting couple of evening's research to compare and contrast the competitive rise of Japan in the 80's and the rise of China in the 10's and how the US reacted to both and how both reacted back.
Really? In the 1970's in the US it was possible to own a house and support a family of 4 on a single income in the manufacturing sector. It was also possible to work your way through college and exit with a degree and no debt, in a job market which made that worth it.
Yes global poverty has decreased, but the material standard of living has decreased for most people in the working class in the developed world, while wealth has skyrocketed for a select lucky few.
It's more complex than saying "everyone is made richer" - it seems what has happened is that a few people have been made much richer, and everyone else is somehow equalizing somewhere closer to the bottom.
Who told you that? Both inflation adjusted median personal and household income have increased by double digit percentages over the past 50 years. In the case of household incomes that still happened even with the shrinking average household size (i.e. number of people in a household).
What he said isn't controversial for most academics studying this.
https://en.wikipedia.org/wiki/The_Two-Income_Trap
https://medium.com/@6pranavk/insight-of-the-day-cost-disease...
Want to compare to 1970? Then live with those restrictions and you'll probably have enough money to pay for insurance, college, etc.
Large TVs in 2020 cost just as much as small TVs did in the 70s (if not less), adding air bags to a car costs mere hundreds of dollars, and is ammortized over 20 years, and cancelling your phone plan is not going to let you support a family of four on a single blue-collar income. Shit, cancelling your phone plan won't even compensate for the annual growth in your health insurance premiums - which is something that compounds every year.
TVs are an example of the plethora of technology that most homes have. From microwaves to air conditioners to cell phones...
"Canceling one thing isn't going to support your family" Which is why I didn't mention one thing... I mentioned a handful of things. A host of "modern conveniences" that add up.
Education, health insurance and other things cost more? Those are also things that the government has stepped in to make more affordable as well... which is a massive topic in of itself.
My main point stands... quality of life and cost of life has risen with hundreds of costs that add up (a mere 100 for an air bag... a mere couple bucks a month for data on your phone... a mere 5 for text messages... a mere 100 for a microwave... etc.etc.etc.etc).
Remove all the "extra" stuff that's standard today compared to 50 years ago and even with outliers (IE Health Insurance), you'll still be much closer to being able to afford the same quality of life as "way back then".
AC & Dishwashers are also only a few thousand dollars of value, mostly on the AC side. The rest are relatively luxury items.
And low end AC & Dishwashers can be had for only a few thousand dollars NOW. Back in the 70s the cheapest ones were still expensive enough to require an upper middle class income.
They're not. Home ownership is currently at about the same level it was in the 90's, declining from a boom in the 2000's driven by sub-prime lending, but if you look at generational differences, everyone younger than baby boomers is basically owning homes at much lower rates.
https://journal.firsttuesday.us/bad-news-for-tomorrows-homeo...
https://www.census.gov/housing/hvs/files/currenthvspress.pdf
So why is the average home size nearly double the size in the 70s?
This may have something to do with the fact that the cost of housing has risen drastically as compared to median income over the past few decades[2]
While the cost of education has also more than doubled for men, and increased by 20% for women relative to median income (despite a 73% increase in median income for women) since the 70's[3]
Yes income is increasing, but the costs for the markers of a middle-class existence are increasing much faster.
[1]https://www.marketwatch.com/story/this-depressing-chart-show...
[2]https://www.longtermtrends.net/home-price-median-annual-inco...
[3]https://college-education.procon.org/median-incomes-v-averag...
So, inflation as calculated on a basket of goods divorced from living expenses gives inaccurate comparisons. Yes wholesale food prices for example got really cheap, but overall food is still a significant fraction of people’s budgets.
This is more of a feature of the transition from a Bretton Woods/Treasury/politics-based monetary policy to neoliberal/central bank/finance-based one no? At 10% inflation, the creditors just ate most of your mortgage debt regardless of whether you worked in manufacturing or not.
Your point stands tho.
Basically everywhere that wasn't the US and Canada (excluding Pearl Harbor).
For the US it was a huge shock, for a country like Poland it was Tuesday.
https://en.wikipedia.org/wiki/September_1939#September_1,_19...
I'll be honest and say I didn't even know the origin of the quote :-))
Modern expectations mean that this isn't what people think it is. You also owned one car per family in the '70s; new houses often have to build two parking spaces. Housing prices were held down because public housing was still being built; today it's unmaintained and often torn down. Old houses were built on small lots; suburbs today require a quarter-acre. Colleges received way more public funding as a proportion of their income than they do today, and enrollment was lower.
Housing has gone from subsidized to restricted and college has gone from subsidized to free-market while demand keeps rising. It's not surprising that the price increased. (And the protectionism lobby is now also the anti-housing lobby!)
Since the 1970s, physical house size has increased from around 1600 sq.ft to around 2600 sq.ft. At the same time, house occupants has dropped from about 3 to about 2.5 [0]. In addition to that, the building requirements and safety features have increased costs significantly per square foot.
Average house price in 1975 was around 42k [1]. That is a little over 200k in 2020 dollars [2]. Doing some math with 200k/1600 vs 380k/2600 gives $125/sq.ft and $146/sq.ft or roughly a 14% increase in relative cost. In 1975, median income was around 12k [3] while median income in 2018 was around 63k [4]. (EDIT: I'd note that I'm being generous as home prices dropped significantly this year and I went with the higher numbers)
SHOCK!!!!
An average house in 1975 is 16.5x the yearly income while an average house in 2020 is only 6x yearly income despite being 40% bigger and 15% less occupied while being safer to live in as well.
I'll give you college debt, but for all the wrong reasons. The government artificially incentivizes bad loans. They insist that kids CAN'T get that 380k house loan (even though the house is collateral), but also insist that kids CAN get that same 380k for a piece of (often useless) paper without any collateral backing. Get rid of the stupid laws and you'll watch that problem go away all by itself. Get rid of the pressure to have a degree to sweep floors and you'll watch college prices fall even faster.
[0] https://www.aei.org/carpe-diem/new-us-homes-today-are-1000-s...
[1] https://fred.stlouisfed.org/series/ASPUS
[2] https://www.usinflationcalculator.com/
[3] https://www2.census.gov/prod2/popscan/p60-104.pdf]
[4] https://www.census.gov/library/stories/2019/09/us-median-hou...
As a side-note, lately I've been curious about what's actually been driving the cost of higher education. I know it's often cited as "administrative costs", but what does that actually mean? Where does the money actually go?
If you're asking about tuition? A few places.
Marketing, sports stadiums, student services (Counselling, mental health, etc), bureaucratic waste (Middle managers overseeing operations people), overbuilt facilities that serve no pedagogical value (Stained glass windows and large, pretty university grounds don't maintain themselves).
But that's just the half of it. In addition to tuition, most universities force you to pay for room and board, by living on-campus. Those fees go into paying for overpriced, poor-quality catered food, and overpriced dorm rooms.
I had 28k in margin interest last year, 100% deductible.
US is greatly skewed to the rich
Here is a Brookings report that explores whether global trade has directly contributed to worsening income inequality: https://www.brookings.edu/articles/worsening-american-income...
The problem isn't with the "Single Income" part.
It is with the "cost" of hosting and medical care.
The world in general has a huge incentive in driving up asset prices. And Housing is one of the biggest asset.( Correct me if it is not the biggest ) And rent is the biggest expenses in living cost.
What needs to be done is fixing housing prices / supply.
With free trade you put more workers in competition with each other pushing down salaries while capitalists can move their investment to a wider variety of places to get increased return.
Even for poor countries it is not always a good thing. China before WWII got screwed over by western capitalists owning all the industry. That is why they have the required local partners now.
I am not saying free trade is inherently bad. Just that you can have too much of a good thing. I think we are way past that in the current world and should be scaling back, not ramping up.
I find a lot of people in the US are quick to point out the problems with free trade, but tend not to look at examples of places that took strong protectionist positions and how that played out over the span of decades. Point being, there are examples to be made for either argument being taken towards a limit.
A career path in microelectronics is suicide for US citizens. It's going the way of nuclear engineering.
- x86-CPUs from AMD with better efficiency in Desktop and Datacenter
- Apple-ARM Macbooks showing that it doesnt have to be x86 at all
- x86-APUs from AMD with Navi-GPUs versus Intel Xe
- Intel Xe as Compute vs. NVidia Cards.
All of these pivot around Intel at least matching the process of their competitiors. Also, Intel has optimized 14nm that much, that their initial yields and performance gains on 10nm and 7nm will probably piss-poor initially.
I suspect this is why Ampere is gonna run hot and consume a lot of power relative to AMD's GPU showing this cycle.
1. Intel has purchased capacity from TSMC which will presumably be used for Xe HPC. 2. TSMC does not have enough capacity for AMD to take significant market share 3. Intel still has a lock on high margin datacenter parts
TSMC 5nm uses EUV by default, and very likely has no quad patterning layers, and possibly they threw out double patterning layers too.
Not having to give any considerations to multiple patterning should allow for much freer design rules, and let recover density from sacrificial lines.
“Part of what makes 5nm yield slightly better is perhaps down to the increasing use of Extreme UltraViolet (EUV) technology, which reduces the total number of manufacturing steps. Each step is a potential chance to decrease yield, so by replacing 4 steps of DUV for 1 step of EUV, it eliminates some of that defect rate.”
For example, can I get double the logic density if I'm happy to get defects from 0.1 per square centimeter to 1000 per square centimeter? If I can, I'd very much like to for some usecases. Think of things like SRAM, where you can simply have a "bad address handling unit", which keeps a list of the bad addresses and maps them elsewhere. As long as my design is fine with adding a few cycles latency on 1 in a million ram accesses, I'm fine.
The exception is wafer-scale chips, which are chips that span multiple "reticles", i.e. the size of the masks used to manufacture normal chips.
I have a Ph.D. thesis linked to building a wafer scale chip (ala cerebras, although 10 years ago). And cerebras AFAIK is the FIRST commercial product using wafer-scale approaches. Maybe we will see more products ? I have doubts...
Note also that defect density varies depending on location of the chip on the wafer. There are areas with more defects on average, so it's not as simple as 1 number.
Amdahl?
Binning is any kind of sorting by silicon quality. It can refer to sorting by the number of "good cores" you get, but it has usually meant sorting by speed grade or other characteristics.
> and how all manufacturers build their line of products.
Having identical functional units you disable is a great yield maximization strategy if it fits your product. Most semiconductor products are not in this category-- it's mostly just multicore CPUs and GPUs that have a lot of identical cores and relatively small uncore.
> I have all but mobile, or about 70% of the market.
In dollars, maybe, but there a whole lot of units of things that you cannot build a meaningful deactivation strategy on.
But there are also other tricks, like redundant vias, or SRAM spares, which do make 'yield' less listed by defects than it might first appear.
https://www.eetimes.com/altera-uses-redundancy-to-boost-yiel...
There are multiple "bins" between "good core" and "failed core". The first one runs flawlessly at the nominal speed while the second one can't run within the envelope defined. There are cores that can run overclocked, cores that will fail above a certain frequency, and cores that are so broken they won't work on any clock.
That's why, for instance, some Cell processors left the factory with 6 SPUs and Sun's first Niagaras had SKUs with only a small number of cores.
Not any semiconductor defect can be "fused off," or worked around in software, and provisions to do so cost space too.
Super high yields are transformative, as they allow for much simpler test, and binning strategies. Throwing out 1 chip out of 1000 is incomparably cheaper than having to salvage "every n-th chip."
It is amazing to me that we can achieve such low defect rates on projects so complicated as highly brain-insecure sticks of meat that evolved to hunt and forage. Under the principle of radical skepticism the true defect rates could be much higher.
I guess the complacency mindset was already at work...
I doubt they'll pass on the savings to consumers however. Certainly not Apple or Nvidia, and so the competitors Qualcomm and AMD have no reason to drop prices either.