American Economics Thread

[manqiangrexue]

Interesting stats on how CHIPS act is forecasted to impact US semiconductor production.

Please, Log in or Register to view URLs content!

Eversince the COVID readiness map, I have never ceased to be amazed at America's propensity to imagine a future or scenario that is wildly favorable to itself and present it as if it were real and serious data as opposed to the amorphous random fantasies that they actually are. It's like the adult version of a kid drawing himself in a Superman costume with 30 inch arms a 70 inch chest standing next to the kid he doesn't get along with, who is actually taller, stronger and smarter than him in real life, as some hobbit-sized black and brown scribble scrabble ball of dirt with lines for arms and legs. The real data has China at 365% and the US at 11%. The instant we exit the realm of reality and cross into the realm of American imagination, America shoots up to 203% and China drops to 86%.

 

[Airhouse]

Eversince the COVID readiness map, I have never ceased to be amazed at America's propensity to imagine a future or scenario and present it as if it were real and serious data as opposed to some made up shit. It's like the adult version of a kid drawing himself in a Superman costume with 30 inch arms a 70 inch chest standing next to the kid he doesn't get along with, who is actually bigger, stronger and smarter than him in real life, as some hobbit-sized black and brown scribble scrabble ball of dirt with lines for arms and legs. The real data has China at 365% and the US at 11%. The instant we exit the realm of reality and cross into the realm of American imagination, America shoots up to 203% and China drops to 86%.

I would even go a step further by saying anyone that predicts events 10 years, or even 5 years beyond current date is performing voodoo magic or astrology, regardless of whether the result is positive/negative or not.

 

[Serb]

Going from 0% at <10 nm to 28% within a year.

Lol. Lmao even. Are they assuming TSMC up and running or something? Or counting Intel 10 nm rebrand to 7 nm?

Essentially, the only way TSMC sets up a facility there is if the gov. directly pays for it all through subsidies (taxes - inflation).

For now, everything the US got out of TSMC (new fab) was a "non-binding preliminary memorandum of terms", as current subsidies won't be enough, then there will be standard delays, yet they act like it's already all over lmao, Jai Hind logic.

Under normal circumstances and market forces, there is no way that they could produce anything profitably there, zero economic sense.

The US is simply not a manufacturing country. Engineers in shortage are one thing, but this is just the tip of the iceberg there.

So what if they produce more chips? Would that be worth it, knowing how much they had to bleed out for that? I doubt.

 

[gelgoog]

Analyzing these predictions a bit.

China has at least three companies working on DRAM. Right now only CXMT has fabs. But I would not expect the situation to remain like this over a whole decade.

In memory the incumbents CXMT (DRAM) and YMTC (NAND) will likely continue to expand their production. They just doubled their fab floor space last year. And I expect this to continue. The adjacent plots of land to where they have their fabs allow for another expansion. So production will increase by at least a further 50% after the doubling this year.

China is just a mere few steps away from having its own tools for making the whole DUV immersion lithography process. Once that happens they will have the technology to compete with the market leaders in memory over the next decade. NAND is limited mainly by improvements in etching and China has good tool makers in that segment. DRAM will likely move towards 3D structures this decade and this means the need for EUV in DRAM will likely be reduced.

Getting DUVi tools which work at 7nm or 5nm is just a couple iterations away once they get the Chinese 28nm tools. I expect SMSC to continue to increase its <10nm logic production after the recent doubling of floor space they did.

You could say the chart is already wrong since it claims that China will only have 3% of <10nm logic by next decade. I think SMSC is probably already at a higher 7nm market share than that.

 

[chgough34]

If you go to the bare baseline of "competent" STEM professionals, ie software engineering at a FAANG, you will find that indeed 60-70% of the engineers are Indian/Chinese. Not that you'd ever be able to witness this for yourself.

Even taking this as true with all of the requisite assumptions that it requires, are the FAANGs the only successful software publishers in the United States?

 

[chgough34]

I would even go a step further by saying anyone that predicts events 10 years, or even 5 years beyond current date is performing voodoo magic or astrology, regardless of whether the result is positive/negative or not.

Predicting 5 years beyond current date in this is perfectly fine - foundries take multiple years to construct so CapEx and capacity coming online are going to be from projects already announced and under construction

 

[FriedRiceNSpice]

Even taking this as true with all of the requisite assumptions that it requires, are the FAANGs the only successful software publishers in the United States?

Looking at more competitive/selective companies than FANG, generally an even larger proportion of the engineers will be East Asian. Overall majority of software engineers in US (~90%) are employed at less competitive/selective companies than FANG, and you see the proportion of East Asians and Indians gradually decline the further down you go. FANG is rough proxy for 90-95 percentile of STEM talent.

 

[chgough34]

Looking at more competitive/selective companies than FANG, generally an even larger proportion of the engineers will be East Asian. Overall majority of software engineers in US (~90%) are employed at less competitive/selective companies than FANG, and you see the proportion of East Asians and Indians gradually decline the further down you go. FANG is rough proxy for 90-95 percentile of STEM talent.

Even ignoring the massive selection bias associated with this type of analysis (with restrictions associated with which foreign-born individuals get employment authorization + US-born Asians are overrepresented in San Francisco/San Jose and U.S. born children generally stay in the metro areas where they grew up) - this effectively concedes the point - US firms (including thousands of successful software publishers) are primarily staffed by U.S.-born residents educated in the United States - and thus by implication, U.S. education is thorough, efficient, and effective. What’s more - the towering global successes of less notable U.S. software publishers - such as Oracle, Adobe, PayPal, Salesforce, VMWare, ServiceNow, NetApp, ADP, Citrix, Ansys, among others - are made by individuals in mid-career and beyond (those educated in the 1990s/early 2000s). Every indicia of educational quality shows education is much better than it was 2 decades ago.

 

[Serb]

Where is the magic invisible hand or whatever that bullshit is it called to fix this, lol, haha?

https://twitter.com/i/web/status/1788141087132950620

Small businesses are struggling, but plutocracy-owned sheeple there are taught to only look at the stock market, which is large caps territory.

https://twitter.com/i/web/status/1788603336205472099

https://twitter.com/i/web/status/1788679155841892720

 

[manqiangrexue]

Even ignoring the massive selection bias associated with this type of analysis (with restrictions associated with which foreign-born individuals get employment authorization + US-born Asians are overrepresented in San Francisco/San Jose and U.S. born children generally stay in the metro areas where they grew up) -

The selection bias is that the smartest available to the US make it to those positions and those are East Asians, both from abroad and those who have dominated the best education that American can offer.

this effectively concedes the point - US firms (including thousands of successful software publishers) are primarily staffed by U.S.-born residents educated in the United States - and thus by implication, U.S. education is thorough, efficient, and effective.

The only point conceded is that US firms are heavily dependent on imported and/or Asian minority input at the upper and most critical rungs and thus by implication, US education is thoroughly, efficiently, and effectively dominated by the Chinese and other East Asians at home and defeated by Chinese education abroad.

What’s more - the towering global successes of less notable U.S. software publishers - such as Oracle, Adobe, PayPal, Salesforce, VMWare, ServiceNow, NetApp, ADP, Citrix, Ansys, among others - are made by individuals in mid-career and beyond (those educated in the 1990s/early 2000s). Every indicia of educational quality shows education is much better than it was 2 decades ago.

What's more, American successes rely on the foundation that it has built in the past. In regards to technologies of the future, the US is completely unable to compete with China, and that remains the case even with the willing and unwilling recruitment of Western Europeans and vassalized Asian nations. That is why today, an innovatively-defeated US can only resort to political means to attempt to slow down China's inevitable victory in the tech war and dominance of future technologies.