> Who else has reduced the cost of solar panel manufacturing over the past 10 years? As far as I can tell it's entirely, 100%, Chinese companies. But maybe you know something I don't.
The solar cost reduction numbers come in the form of reduced cost to manufacture and improved panel performance. There is a lot of non-Chinese research into the latter, but even in the former case a lot of inventions and technology is developed then licensed. USA and Germany.
> I don't know enough about smartphones,
Smartphones, invented in the US.
> electric cars,
Obviously not recently invented, but the recent push to make these competitive and take market share has all come from US, Germany, Japan, and SK.
> conventional cars, reciprocating engines
> jet engines, airliners,
China has been trying to develop their own jet engines for 70 years and desperately needed a competitive jet engine since 40 years when they broke up with the Soviets. No dice. They also want civilian airliners rather than having to buy from the US or EU. There's always rumors this time they might have got something workable last year, but it's quite amazing how difficult it has been for them.
> reusable rockets,
Well the Chinese rocket program is basically repeating the rockets of the 60s. Meanwhile US companies are basically re-inventing commercial rocketry with innovations significantly reusable rockets. China has quickly tried to follow that naturally and they could well do a good job. My point is the innovation and the new ideas aren't coming from there. It's just the same story where ever I look.
> or AI
Just more same thing. The inventions around the recent AI boom ("deep learning" advances like neural networks on GPUs, GAN) has not come out of China.
> The most interesting innovations I've seen in CPUs in recent years are from Padauk (China)
What's interesting about Padauk?
> RISC-V (everywhere, especially Berkeley, but also including China).
RISC-V is another great example that's not come out of China. Despite the fact they desperately need an unencumbered ISA (which is part of the reason for their extraordinary hijacking of their ARM subsidiary), they had never taken the step to create their own (before that they were using MIPS). Using RISC-V is just more of the same thing: following not inventing or leading.
The ISA is not really where much innovation is going on in computing though. You just need something as an interface, and everyone is more or less the same (although RISC-V is not very good as ISAs go, sadly).
> NVIDIA seems to be leading in GPUs; like Lam Research, it's an American company with a Chinese-American founder.
Yes, and yet again not Chinese.
> But now the critical innovation bottleneck seems to be not designing a working CPU design but successfully manufacturing it, and the supreme giant in the field is TSMC, a Chinese company,
It's both, for high performance CPUs and GPUs. There are only a few companies on the planet who can do the logic and circuit design, Intel, AMD, Apple at the cutting edge, and say ARM and IBM after that.
Also we're talking about mainland China or the nation governed by the CCP here (the context was "China" catching up with "not China" in silicon manufacturing). So China does not have TSMC.
> In terms of crypto, it's an extremely international community, but two of the most significant achievements of the last 20 years were the collision attacks on MD5 and SHA1, which were found at Shandong University, which is in China. (Of course nobody knows where Satoshi is from, but China is near the bottom of the list of possibilities.)
Comparatively little crypto has come from China compared with, say, USA. Who developed the crypto math and algorithms that you're using right now?
> I agree with you about operating systems and programming languages, though of course if you `git annotate` Linux you will find a rather large slice of it is written by people with Chinese names, a lot of whom work at Chinese companies.
Not actually a lot compared with how much they rely on Linux (a non-Chinese invention). Recently it has picked up in number, but you actually find a lot of those are janitorial and automated linter type of patches. There are some good developments from China, but very few compared with many other countries.
> As for drugs, well, I explained in https://news.ycombinator.com/item?id=29792057 how far China is from "playing catch up" with respect to covid vaccines. But maybe in other drugs it's true. What are the big new innovations in drugs, and where do they originate?
Well since we're on the topic of COVID their vaccines have been a great achievement. Chinese vaccines use inactivated virus, adenovirus vectors, subunit/protein, which were all techniques developed by the west for a long time. The new mRNA vaccine they are developing is based on this new mRNA vaccine technology which has largely been developed and invented outside China.
> The solar cost reduction numbers come in the form of reduced cost to manufacture and improved panel performance. There is a lot of non-Chinese research into the latter, but even in the former case a lot of inventions and technology is developed then licensed. USA and Germany.
It's come entirely in the form of reduced cost to manufacture; mainstream solar panels have been about 21% efficient and low-cost ones about 16% efficient for decades. Thin-film panels have gotten a lot more efficient over the last 10 years, but they've also been completely eclipsed by traditional silicon panels.
Can you name one or two inventions that have been important in reducing silicon PV manufacturing costs in the last 10 years that originated in the US and Germany? I can't.
I'm confused about why you're bringing up the original invention of things like smartphones (arguably Finland and Japan in the 01990s, rather than Danger in the US, but there's been continuous development), and jet engines. I agree with you that 40 or 70 years ago the US was inventing a lot more things than China was; 40 years ago China was still largely Communist. 1000 years ago China was doing more; the Song had a paper-money-based fossil-fuel mass-production economy administered by a centralized bureaucracy. 2500 years ago the center of innovation was Greece and India.
But I thought we were talking about a much shorter timescale here: where are things being invented now, like, the last ten years. And the US definitely didn't invent jet engines or smartphones in the last ten years. What have been the crucial innovations in smartphones in the last ten years? I think you'll find that most of them came from China, though the M1 is designed in the US.
> Well the Chinese rocket program is basically repeating the rockets of the 60s.
Until a couple of years ago, the US couldn't even "repeat the rockets of the 60s"; it was relying on Soyuz to get its astronauts to the ISS. Reusable rockets themselves aren't a "new idea" on the timescales we're talking about here; the US tried that in the 70s and failed so badly it destroyed its entire manned spaceflight program. Presumably SpaceX has been able to succeed at this in part due to coming up with some genuinely new ideas; do you have any idea what they are or where they came from?
> Also we're talking about mainland China or the nation governed by the CCP here (the context was "China" catching up with "not China" in silicon manufacturing).
If you want to start talking about that, I guess you're free to, but I have no interest in starting to talk about that subject.
> What's interesting about Padauk?
Padauk's CPUs (another Taiwanese company, by the way, in case you're keeping track) have two particularly interesting innovations:
1. They use a second hardware thread on the same CPU (or, in theory, up to three of them) to do things like PWM, SPI, or quadrature decoding which are usually done in hardware (or, say, pioasm on the RP2040, or the tiny amount of programmable logic on some new Microchip AVRs). This means you aren't limited to the peripherals included on a particular chip; you can in theory do things like generate an NTSC signal while running application code, which is something I've sort of done on the ATMega328, but I had to settle for running application code during the HBI and VBI.
2. They cost less than many discrete signal transistors, so it's reasonable to use them where you would use a transistor. Harder to second-source tho.
They also invented their own programming language for the chips, an assembly language with C-like syntax, but I'm not that impressed with it.
> It's come entirely in the form of reduced cost to manufacture; mainstream solar panels have been about 21% efficient and low-cost ones about 16% efficient for decades. Thin-film panels have gotten a lot more efficient over the last 10 years, but they've also been completely eclipsed by traditional silicon panels.
> Can you name one or two inventions that have been important in reducing silicon PV manufacturing costs in the last 10 years that originated in the US and Germany? I can't.
Yes, https://www.sciencedirect.com/science/article/pii/S187661021... LeTID problems in PERC panels were comprehensively examined and solutions found by these researchers. The parent company is South Korean (also not Chinese), but PERC was one of the biggest jumps in efficiency of cells in the past 10 years.
> I'm confused about why you're bringing up the original invention of things like smartphones (arguably Finland and Japan in the 01990s, rather than Danger in the US, but there's been continuous development), and jet engines.
Everything is built on other things, but still you can see significant advances. By smartphone I'm not referring to cellular radios or batteries or computer chips (all of which themselves build on other things), but the invention that revolutionizes a technology. The iPhone. I'm aware there are things people claim are smartphones before that, but that's not what I'm talking about.
> I agree with you that 40 or 70 years ago the US was inventing a lot more things than China was; 40 years ago China was still largely Communist. 1000 years ago China was doing more; the Song had a paper-money-based fossil-fuel mass-production economy administered by a centralized bureaucracy. 2500 years ago the center of innovation was Greece and India.
And today US is still inventing a lot more things than China was.
> But I thought we were talking about a much shorter timescale here: where are things being invented now, like, the last ten years. And the US definitely didn't invent jet engines or smartphones in the last ten years.
My point with jet engines is that China is incapable of developing them independently after 70 years of effort and with the luxury of having samples of good jet engines to copy for the entire time.
> What have been the crucial innovations in smartphones in the last ten years?
iPhone is near enough (I never mentioned 10 years, just general modern).
> I think you'll find that most of them came from China, though the M1 is designed in the US.
Which of them of them came from China? Not software, not silicon manufacturing, not digital logic design, not LTE modems, not screens, not new NAND technologies, not cameras. So what is the "most of them" that exceeds those things?
> Until a couple of years ago, the US couldn't even "repeat the rockets of the 60s";
Of course it could.
> it was relying on Soyuz to get its astronauts to the ISS.
A matter of funding, the rocket technology did not go away obviously they still maintain ICBMs and SLBMs.
> Reusable rockets themselves aren't a "new idea" on the timescales we're talking about here;
No, I'm talking about inventing an actual working reusable rocket and all the technologies and inventions required to achieve that.
> the US tried that in the 70s and failed so badly it destroyed its entire manned spaceflight program.
I'm not sure what you're talking about, but a lot changes in 50 years. 70s was only a decade after the first manned space flight.
> Presumably SpaceX has been able to succeed at this in part due to coming up with some genuinely new ideas; do you have any idea what they are or where they came from?
Certainly not China, which (yet again) is following and copying.
It's an American company though, with very strong restrictions on who can work for them (non-citizens / permanent residents I thin are not allowed).
> If you want to start talking about that, I guess you're free to, but I have no interest in starting to talk about that subject.
This is what the conversation is about. Read the thread, it started from wondering about whether China will catch up in silicon manufacturing. That does not make any sense if you count China as including Taiwan does it?
All my comments apply to mainland China.
> Padauk's CPUs (another Taiwanese company, by the way, in case you're keeping track) have two particularly interesting innovations:
I'm not sure what is innovative about what you're describing as hardware multi threading or software DSP but I'll take your word for it there's probably more to it that I'm not understanding. Still, they're also not mainland China, not surprisingly.
Also, it is not very generous to describe something like this as innovation but at the same time say that iPhone was not innovative.
You also don't address any of the rebuttals I made to other things (e.g., RISC-V, crypto, AI, GPUs). Does that mean you concede on those points?
The solar cost reduction numbers come in the form of reduced cost to manufacture and improved panel performance. There is a lot of non-Chinese research into the latter, but even in the former case a lot of inventions and technology is developed then licensed. USA and Germany.
> I don't know enough about smartphones,
Smartphones, invented in the US.
> electric cars,
Obviously not recently invented, but the recent push to make these competitive and take market share has all come from US, Germany, Japan, and SK.
> conventional cars, reciprocating engines
> jet engines, airliners,
China has been trying to develop their own jet engines for 70 years and desperately needed a competitive jet engine since 40 years when they broke up with the Soviets. No dice. They also want civilian airliners rather than having to buy from the US or EU. There's always rumors this time they might have got something workable last year, but it's quite amazing how difficult it has been for them.
> reusable rockets,
Well the Chinese rocket program is basically repeating the rockets of the 60s. Meanwhile US companies are basically re-inventing commercial rocketry with innovations significantly reusable rockets. China has quickly tried to follow that naturally and they could well do a good job. My point is the innovation and the new ideas aren't coming from there. It's just the same story where ever I look.
> or AI
Just more same thing. The inventions around the recent AI boom ("deep learning" advances like neural networks on GPUs, GAN) has not come out of China.
> The most interesting innovations I've seen in CPUs in recent years are from Padauk (China)
What's interesting about Padauk?
> RISC-V (everywhere, especially Berkeley, but also including China).
RISC-V is another great example that's not come out of China. Despite the fact they desperately need an unencumbered ISA (which is part of the reason for their extraordinary hijacking of their ARM subsidiary), they had never taken the step to create their own (before that they were using MIPS). Using RISC-V is just more of the same thing: following not inventing or leading.
The ISA is not really where much innovation is going on in computing though. You just need something as an interface, and everyone is more or less the same (although RISC-V is not very good as ISAs go, sadly).
> NVIDIA seems to be leading in GPUs; like Lam Research, it's an American company with a Chinese-American founder.
Yes, and yet again not Chinese.
> But now the critical innovation bottleneck seems to be not designing a working CPU design but successfully manufacturing it, and the supreme giant in the field is TSMC, a Chinese company,
It's both, for high performance CPUs and GPUs. There are only a few companies on the planet who can do the logic and circuit design, Intel, AMD, Apple at the cutting edge, and say ARM and IBM after that.
Also we're talking about mainland China or the nation governed by the CCP here (the context was "China" catching up with "not China" in silicon manufacturing). So China does not have TSMC.
> In terms of crypto, it's an extremely international community, but two of the most significant achievements of the last 20 years were the collision attacks on MD5 and SHA1, which were found at Shandong University, which is in China. (Of course nobody knows where Satoshi is from, but China is near the bottom of the list of possibilities.)
Comparatively little crypto has come from China compared with, say, USA. Who developed the crypto math and algorithms that you're using right now?
> I agree with you about operating systems and programming languages, though of course if you `git annotate` Linux you will find a rather large slice of it is written by people with Chinese names, a lot of whom work at Chinese companies.
Not actually a lot compared with how much they rely on Linux (a non-Chinese invention). Recently it has picked up in number, but you actually find a lot of those are janitorial and automated linter type of patches. There are some good developments from China, but very few compared with many other countries.
> As for drugs, well, I explained in https://news.ycombinator.com/item?id=29792057 how far China is from "playing catch up" with respect to covid vaccines. But maybe in other drugs it's true. What are the big new innovations in drugs, and where do they originate?
Well since we're on the topic of COVID their vaccines have been a great achievement. Chinese vaccines use inactivated virus, adenovirus vectors, subunit/protein, which were all techniques developed by the west for a long time. The new mRNA vaccine they are developing is based on this new mRNA vaccine technology which has largely been developed and invented outside China.