Hacker News Reader

Comment by wishart_washy | Hacker News Reader

Notes from the ground here. My undergrad was in EE at an "elite" school - I can say definitively that ~50% of the MS/Ph.d. students in EE were from China (with the other 50% largely from India). They TA'd many of the graduate level courses and were the majority of the students taking high level semiconductor design courses that centered on fabrication and nanotechnology.

China's short term goal is clearly to reach parity so these students can enter a mature semiconductor industry. Give these students ~5-15 years to gain expertise from the American academic system and industrial research base and I think we'll start seeing China innovate.

American expertise will naturally decline because we don't have students studying EE. To many good full-stack jobs that pay ~40% more.

View on HN

Too late.

China dominates in the 'jelly bean[1]' chip business. This was the bread an butter for the dozen semiconductor shops in the Bay Area during the 70's through the early naughts. As fabs shut down they became 'fabless' and had their work shipped out to TSMC. Now they exist as a genre in China in much greater numbers than they do here. Companies that will make a couple of wafers of USB + ADC + random logic kinds of things.

And the innovation never comes from the chip companies, it comes from companies that are building the products. They go to the chip companies and say "Can you make a chip that does this?" and then a new jelly bean chip is born and if that product gets traction everyone decides to make one with their own bit of spin.

What is unique to the Chinese scene is that there are thousands of minute variations rather than any sort of ordered discipline. Transistors are free at the jelly bean chip level so there is no incentive to re-use or save on chips.

ARM has become the defacto microprocessor architecture because, in part, ARM will license it to anyone.

In the US all of these folks have coalesced into a few mega companies (TI, NXP, Microchip, Maxim) which don't have much in the way of competition locally. So innovation stops because conservative business practices rule.

This really became clear to me when I started perusing datasheet sites for the Chinese market and realizing there are hundreds if not thousands of chips for which there is no English data sheet. This is a complete reverse from the last century when nearly every data sheet was in English and getting a Chinese version was a challenge.

[1] Jelly Bean chips are those that are made in batches of 1 - 10 million with a set of functions that are fairly specific to their application.

As chip manufacturing goes DRAM used to be at the bottom of the ladder. Many logic fabs did DRAM first to learn the rope because the fab technology is simpler for DRAM (less layers than logic). DRAM fabs bought used equipment from logic fabs, so there were always swing capacity available. DRAM business was tough and the main challenge was yield (therefore cost and margin). I don't know if any of this has changed more recently as it certainly seems people are not bringing DRAM capacity online as expected -- maybe they are scared by the prospect that China may bring in a large amount of capacity.

The Chinese company mentioned in the article is only incidental to the story so the article title is kind of a click bait. Sure they decided to buy technology from the Taiwanese fab UMC, which is perfectly legal. UMC as a very large fab itself, could also be expected to provide the technology. For whatever reason, UMC seemed unable to deliver and had allegedly resorted to underhanded tactics instead.

> And the innovation never comes from the chip companies, it comes from companies that are building the products.

I agree with you.

Eng has really got to stop thinking they are the source of ideas, most good ideas come from customers that have needs that can't be currently met. Eng balances design constraints, as companies grow, hierarchies get stronger and creativity drops. Maximum innovation comes from small companies solving customer (the ones building the products) problems.

The Jelly Bean chip business could get disrupted by the proper mix of analog and digital. The PSoC [1] was almost but not quite there

[1] https://en.wikipedia.org/wiki/PSoC

I think you are correct as well, its a matter of timing. Eventually there will be the equivalent of an FPGA that instead of having "LUT"s or "CLB"s they will have a bunch of useful bits of tech and an internal wiring bus that will be programmed by setting switches or blowing fuses. Then the variation will be how many pins on the package which will determine how many of these things you can bring out to pins.

I did a thought experiement of what it would take to make a 'universal' TTL chip, which was basically a 14 pin, 16 pin, and 18 pin device where all 74xx chips were actually inside of them and through a programming step you bonded one of them to pins on the lead frame. It required a 90nm or better feature size. The trick was you only have (number of pins) number of input stages and output stages (all the "TTL" part of TTL (these are fairly large by their nature). The logic was all internal 1.8v with a simple step down regulator to go from 5V down to 1.8V.

You could cover both the SSI and MSI catalogs. Of course there isn't anyone designing new stuff with TTL any more so it remains a thought experiment only.

Based on a link posted not long ago, FPGA's seem to already be changing this way:

http://semiengineering.com/fpgas-becoming-more-soc-like/

This would have been a godsend for managing inventory.

In a somewhat related idea, is that of unioning the footprint of alternate parts in a PCB design so that either part could be used. I have never used any high in EDA software, but if you had end to end design, schematic, pcb layout and parts databases with high level machine-readable symbolic pin descriptions, these could be inferred automatically (unioned footprints).

Using your idea, I wonder if you could automatically have "hot standby" components in the same package?

Well the vast majority of these kids would prefer to stay in the US.

Clearly the solution to prevent them from taking their knowledge back to their home countries is to denigrate them politically and make it harder and more inconvenient for them to live and work here.

I would not agree with this. I went to an elite business school and my Chinese classmates all felt there was far more opportunity for them in China with an American education.

They’re not stupid, they look at China and see an economy growing at 6-9% a year while the US is stuck at 1-2% and they know where the growth is for the medium-term. They also see a mature American economy that can be tough to crack even for native white citizens, so its not hard to see how their chances of long-term success are much higher in China than in the US.

They are stupid, if they didn't know about China's fake gdp for the past several years.

"U.S. GDP could top 4% in 2nd quarter, economists say"

https://www.marketwatch.com/story/us-gdp-could-top-4-in-2nd-...

"Baotou in China’s Inner Mongolia Autonomous Region revised its estimated fiscal revenue in 2017 lower by nearly 50% in an annual work report"

https://www.reuters.com/article/us-china-economy-data/anothe...

"China fake data mask economic rebound"

https://www.ft.com/content/a9889330-f51c-11e7-88f7-5465a6ce1...

Nobody can predict anything today, with president temper tantrum throwing weekly wrenches into the global economy. But over the last decade they would have been right; and keep in mind that even in China, most economic gains have been captured by the elite — which anyone educated in a top American university would be a part of almost by definition.

Also weighing heavily on people are family and cultural conflict issues. But anecdotally I can say that the majority of Chinese students don’t want to stay in the US for a variety of reasons; and that many feel their prospects are at least as good in China as the US.

My experience matches the parent's. Of a good 70 classmates the only two who went back had trouble finding work here. I suspect your business school experience might differ from an engineering department. A substantial personality difference wouldn't surprise me.

Business school and liberal arts kids will mostly go back to China, STEM is very different.

My Chinese STEP colleagues have or plan to go back. Too many cultural and political impediments to progress here now.

Is that new? Because the trend has been very stabke for the past 10 years: fuerdai kids (with rich parents) who come to america to study major in liberal arts and business but will almost always go back because that is where their connections are. These are relatively new foreign students in the USA and constitute much of the recent boom.

The more tradition Chinese students who have been coming here since the 1980s (and don’t have very rich parents) mostly pick STEM majors. On graduating, they would rather work for Google in Mountain View than Tencent in Shenzhen.

More like US 4% and China 7%.

A lot of them would love to stay, trading better life/career prospect with the inconveniences of living in a foreign country which only partially accepts them.

That's the old trade-off, and how America got the worlds' talent. But as lives in home countries get better, US would have to do more to make those immigrants feel like home. Or that trade-off would slowly stop making sense, and more people will go back to their home countries. This is not just happening with the Chinese.

>Well the vast majority of these kids would prefer to stay in the US.

I'm afraid you're making assumptions here.

America is not where opportunity is these days, especially in the hardware manufacturing sector.

>America is not where opportunity is these days

Yes, but the living standards in the United States are far higher. Sure you may never become a billionaire here like you can in China, but you can breathe the air and the traffic isn't that bad. The schools are better, the houses are bigger. China may get there, but I know where I'd rather be over the next 20-30 years. And they can always move back.

Not that far, and shrinking fast, depends on who and where you are in China, just like the US. Trump is fixing the air quality issue[1]. China has excellent high speed trains and very innovative and modern shopping malls and housing estates. Google around. Chinese primary and secondary schools are way better than American high schools. Chinese attend international universities to gain experience and network opportunities. They are not here to because here is better.

[1] https://www.nytimes.com/interactive/2017/10/05/climate/trump...

> Chinese primary and secondary schools are way better than American high schools. Chinese attend international universities to gain experience and network opportunities. They are not here to because here is better.

You heard of the (infamous) Gaokao? It's gruelling, and necessitates nothing but continuous study day in and day out to get into a decent Chinese university.

If they could, a lot of Chinese would prefer to study in the US (or any other western country for that matter).

I'd be wary of thinking that it'll take that long. Live in Shenzhen and it is crazy the amount of progress here every year. Public transportation is way better than most American cities and housing is more expensive than most American cities. Apartments here regularly go for a couple million USD, the ads are plastered on the windows of every real estate agency on every street corner. The air here is quite breathable too, and from what I hear, other cities are getting better as well.

Yeah, I think US living standards are higher for now. But I can totally see China seeing parity in less than 20 years. For one thing, they have made large strides in infrastructure investment, and for another thing, they've been realizing and acting on their need to clean up their environmental act over the last few years.

Chinese cities will open about 15 subway networks in the next 5 years. Let that sink in :)

By choosing to live in China, you take on the opportunity cost of a shortened lifespan, likely a decade or so give or take. As long as you can accept that, then China is perfectly fine place to live.

>The schools are better

At the undergrad and graduate level, perhaps, and as another poster commented, probably not for very long.

But for the school years before undergrad, I'm sorry to say, America is in a very bad way.

I don't know how it compares to China, but American public primary and secondary school is definitely not where I'd have my children receive an education

"Well the vast majority of these kids would prefer to stay in the US."

For now.

That doesn't seem to be the current administration's goal. H-1B Visa, America's secret weapon: https://www.youtube.com/watch?v=NK0Y9j_CGgM

Reminded me of comment under recent Cornell dormrooms clip uploaded by Bruce Land (ece.cornell.edu) https://www.youtube.com/watch?v=qthb6taW_9U

"Just dropped in by accident. I see 80% of them are asian girls, is this recording from Shenzhen?"

> To many good full-stack jobs that pay ~40% more.

The gap is even greater in China, since China has a booming Internet industry, second only to the US, while its semiconductor industry is still in infancy and cannot survive without subsidies. My friend, who graduated with a degree in EE from Tsinghua University, lamented that almost none of her classmates stayed in EE fields. Note that Tsinghua University is the undisputed best Chinese university in engineering schools.

Many semiconductor advances were made by Chinese for the last 60-70 years. They just weren’t residents of mainland china.

Throughout the 20th century, the United States benefited greatly from this sort of knowledge drain, from war-torn Europe, and from the Soviet sphere of influence.

The trend is just reverting to the mean.

Many scientists and academics immigrated to the US because of its values: it was a free country and a multicultural country where immigrants are welcomed. The knowledge drain moved smart people to a free country to build a knowledge economy. (Economists too: the Bretton Woods system was built in part by smart people who fled the autocratic or fascist countries). That system of free exchange between stable democracies with social safety nets kept peace in the world for 50 years.

Not sure this brain drain has equal intentions for rule of law and democratic societies.

> The trend is just reverting to the mean.

That's a lie. United States get 1M immigrants a year since 2000. There's over 37,000,000 legal immigrants in US. United States gets the creme of the crop from countries around the world. On the other hand, there were approximately 594,000 immigrants living in China in 2010, most from Asia.

If you can live in US and make 300k-400k, you would do that in a heartbeat, vs living in China where the pollution could shorten you and your family's life, a brush with someone in CCP could destroy your life, no way to get the money you make in China out, no ownership of properties, censorship prevents you from doing research in the real internet, etc.

Oh yeah, what do you think will happen to China after additional $400 billion tariffs, when XiJingPing dies, when the massive shadow debt ($12Trillion and counting) causes widespread defaults and 30% devaluation of yuan?

The current US government approach to immigrants, legal or otherwise, would seem designed to convince immigrants to go elsewhere.

I think China will be fine given that it can simply shut down exports to the US if Trump's tariff pettiness starts to sting. Where, precisely, do you think your phones and computers are manufactured?

lol right, China is going to stop exporting $100B a year of electronics to US. Even though Xi Jing Ping begged Trump not to kill ZTE.

China is not fine. China is trying to balance its shadow debts (12T, 170% corporate debt to gdp), prop up stock market (lost 15% this year), prop up real estate (decline in 1st tiered cities, 25.96x House price to income ratio), and prop up yuan (4% loss this year).

Xi Jing Ping bribed Trump not to kill ZTE.

We've banned this account for using HN primarily for political and nationalistic battle. That's not ok here, as the site guidelines make clear: https://news.ycombinator.com/newsguidelines.html.

Where are the guidelines on usernames?

“I think China will be fine given that it can simply stop accepting money from the US if Trump's tariff pettiness starts to sting. Who do you think precisely, phones and computers are sold to?”

Fixed it for ya. (I tease)

> The current US government approach to immigrants, legal or otherwise, would seem designed to convince immigrants to go elsewhere.

I'm curious where you live and wether or not you can actually, personally, witness the effects of legal immigration there. In SoCal, where I've spent the last decade+, what you're saying rings completely hollow, to me. I have seen literal tour buses of mainland Chinese folks touring new developments, inside of my own new development, in fact, as recently as this spring. If what you're saying were true, why aren't these folks getting the message? That they're touring these homes this way is a discussion for another day. But, as far as I can tell, the only morale panic happening about legal immigration to the United States is happening with American citizens.

> I think China will be fine given that it can simply shut down exports to the US if Trump's tariff pettiness starts to sting. Where, precisely, do you think your phones and computers are manufactured?

What does that have to do with anything?

Are they touring those developments in order to find a place to live?

Or are they investors who will stay in China?

As far as I've seen from the last 2 years of living in this neighborhood, they're moving in and staying here. At least enough time out of the year that you don't notice if they're gone for prolonged periods of time (and it would be hard not to notice, the neighborhood is blanketed by solicitors and their ridiculous flyers). A few homes are known to be empty, but only a few, and this development has > 2k homes in it. There are a couple of rentals, but from what the neighbors and I have gathered there really are only a few. It's Irvine, CA, so it's just sort of like this, I gather.

edit/ Having said that, it's not very different from what you'd see in the San Gabriel valley in recent years, either. If they're living here full-time, or not, is any ones guess. I suppose I could find out if I really wanted to, but it'd probably take weeks.

I live in New York City. We have all sorts of immigrants here.

> Where, precisely, do you think your phones and computers are manufactured?

Vietnam, which now produces more Samsung phones than China does.

"Samsung has invested $17.3 billion in eight factories and one research and development center in Vietnam, turning the country into its largest smart phone production base, the government said. ... Exports from Samsung Electronics’ factories in Vietnam totaled $54 billion last year" [1]

US trade with Vietnam has massively exploded higher over the last decade. The US now imports as much from Vietnam as it does France (an economy 12 times larger). That trend will continue.

China's low value manufacturing is too easy to replace.

The US can move its manufacturing to any number of about two dozen other low cost countries. Just as US capital and the vast US consumption machine built China (and Japan + South Korea before that), it can next build up numerous other nations. The US has no permanently fixed need for China, there's nothing they contribute that can't be manufactured elsewhere (as witnessed by the way clothing manufacturing rapidly fled China on cost, and now eg Samsung's smart phone manufacturing).

Most of what China is looking to do is copy and then compete with what the US already does, from Boeing planes to military tech to semiconductor tech to software to biotech, medtech & pharma. China is a giant clone machine, nothing has changed about that at all. Accordingly, China has nothing novel that the US doesn't already have access to elsewhere or otherwise.

[1] https://www.reuters.com/article/us-samsung-elec-vietnam/sams...

OTOH, China doesn't seem to want immigrants either. It's close to impossible to get Chinese citizenship if one would want to, for example.

I would assume there are lots of expats living and working in China.

Yeah, certainly, especially in places like Shekou. It just doesn't seem possible to fully assimilate, you will always be a laowai.

Just curious, have you been to China?

The $$ is in ASICs. Most code is sorting, hashing, encoding/decoding, matrix-matrix multiply. Bitmain is hitting it out of the park.

This happened with Japan too, anyone that wants to know how they dominated the electronics industry in the 70's/80's should checkout the book "We were burning"

Well, if you EE guys would finally got off your high horses and dropped Verilog in favor of JavaScript, we wouldn’t have this kind of problem!!!

Sure, the circuits might catch fire once in a while, but imagine the ease of development when you can get a bunch of underqualified people to develop stuff!

I left EE and became a full-stack developer! No complaints about JavaScript here. Verilog should just switch to lambdas and get it over with.

Same here. Switched to CS from EE and never looked back, although my respect for EE majors increased a lot.

(I had trouble with the continuous math once we reached steady-state analysis. Besides, I wanted to just get through all that to the fun stuff: discrete math, logic design, and programming.)

On that path after 12 years of blood and sweat as a HW engineer in various FAANG companies. Self teaching myself Algos so that I can eventually leetcode the day light out of my life and take a crack at the same FAANG this time as a SW engineer. The problem with the whole EE landscape is much beyond the Mandarin syndrome. It’s systemic. In fact the canary in the coal mine is Berkeley where the enrollment in graduate level EE courses in analog circuit design and RFIC have been falling rapidly as more and more grads prefer the “deep learning” bandwagon much more than the traditional popcorn chip design and manufacturing which was the mainstay of the local Silicon Valley economy until I was in grad school. Back to my BFS study from clrs :-)

I know you're joking, but Verilog isn't really the best language...

I mean, it wasn't even made for chip synthesis.

I know you’re joking, but look at Chisel or Clash.

Funny, and beneath the joke lies a nice nugget of truth: the landscape of hardware design languages is a truly sad thing to behold and deeply hinders innovation in the space.

There are numerous attempts at fixing the problem, but all of them ring hollow as the big ASIC design shops basically are all relying on ancient languages.

Another thing that IMO has dealt a grave blow to the whole scene is a cultural one: people in the hardware business still haven't understood and digested the lessons of open source and sharing. The whole industry still operates on notions of "industrial secrets" and spends tremendous amounts of energy trying to hide how what they build works.

I mean if you need a example of how bad it still is, the dominant word still used to describe a chunk of functionality in that world is "IP".

Who in their right mind in the software industry would call a piece of code they wrote "intellectual property", as in "just download my IP and use it in your project" ?

If you think Verilog is bad, you should see the languages used by heart surgeons, civil engineers and rocket scientists. The language of human biology is ridiculously verbose without even a hint of polymorphism. ;-)

My point is that it's a little silly to evaluate hardware engineering through the lense of software engineering.

In hardware engineering the part that looks like "software" (what a HW engineer may refer to as the behavioral RTL) is a small component of the total design effort. The tradeoffs are all completely different.

Computer hardware engineering is much closer to designing a bridge than it is to writing a software application. Debating the best hardware-descriptor language is similar to debating whether bridge designers should use No. 1 or No. 2 pencils.

Sorry, but that comment makes it sound like you've either never used pencils, or you have never used verilog.

25 years of high performance chip design.

Perhaps you’re an FPGA guy? The tradeoffs are different in that world.

I am not claiming in any way that the problem we face in HW design are the same as the SW folks do. I most certainly agree that RTL is but a small chunk of the overall effort of designing a chip. But I am afraid my point that this chunk is in a terrible place still stands.

You have also entirely ignored my second point :)

Your second point was that HW engineers haven’t learned about sharing—as evidenced by the fact that they use a different term for it than SW does.

Again this is evaluating HW through a SW prism. Where’s github for bridge designers? rocket scientists? brain surgeons?

HW engineers share info in pretty much the same way that every other industry in the world does (except SW). The concepts are made available through patents, conferences, journals, and press releases. More detailed info is often made available through purchaseable IP.

Maybe it could be better, but there are good reasons that HW IP can’t be shared as easily as SW source code.

> American expertise will naturally decline because we don't have students studying EE.

We won't decline in absolute numbers but comparatively. We will never produce as many engineers as china because they outnumber us by more than a billion people.

That near 100% foreign is the norm everywhere, not just “elite” schools. It has been that way for at least 30 years.

Reply on news.ycombinator.com