Huawei building chip equipment R&D center in Shanghai | Hacker News Reader

Huawei building chip equipment R&D center in Shanghai

asia.nikkei.com

38 pointsksec44 comments

It is quite an exciting era to be alive in - both because of the escalating conflicts between nuclear powers but also because we're welcoming another billion+ humans into living in an advanced industrial society. Hard to believe that 1970s China has come this far. Hopefully India follows close behind.

"may you live in interesting times..."

China is already very advanced, I’d say that when it comes to reasonably large countries (so ignoring city-states) it is one of the most tech/industrially advanced countries out there, if not the most advanced.

For example the US might still have some advances in tech but basic stuff like speed/rail systems are just not up to par, while countries like Japan or Germany might still be ahead of China in some specialized industrial stuff but their tech/IT feels like it’s a geological era behind China’s.

The next billion people that might get industrialized now live in Africa, but the situation is a little bit more complex there.

It's "easy" to go from nothing to 'the latest'. It's not easy to go from 'established' to 'the latest'. Ignoring the way China tramples over its peoples' rights, it was largely agrarian society. It could put factories, train lines, and roads anywhere it wants, build bridges and dams anywhere too.

Now? London spent over 40 years planning The Elizabeth Line before the first shovel was put in the ground. HS2 our second true high speed line look 20 years in planning, and they're still squabbling. Upgrading the country from an existing phone network of copper cables around since the 1900s will have taken 30 years to go fully fibre.

It might be interesting to imagine if China modernized in the 70s, how it might be now.

There's some truth to this, but in my opinion especially the line about not respecting their citizens rights is a cop out. In China there are plenty of silly situations like building a highway around a house, because the person didn't want to sell. The difference is that they then just build around it, instead of incessantly discussing back and forth for decades on different approaches. Many western countries have in effect neutered their ability to move through a) lack of long-term accountability in public administration and b) a motivated minority having greater influence than the silent majority.

For politicians it is usually prudent to just kick controversial stuff down the road a couple of years for the next person to pick up. While something the Elizabeth line is greatly beneficial to London as a whole, it is annoying for neighbors, at least while being built, and for the majority EL isn't a deciding factor for election, for those neighbors it might be.

Rail systems are not "basic", they are a superfluous technology for the US. There's simply no use-case for them outside of several population clusters.

Instead the US has the interstate freeway system that is unparalleled anywhere else in the world in size and capacity.

Rail is important in the US, it's just not for people. The highway system isn't that much better than some other places.

The US freeway (and highway) system is great because it's consistent across the entire US. I don't know any other comparable road system of that size.

E.g. German Autobahns are arguably better than Interstates, but the entire Germany could fit inside Oregon and a bit of Nevada. And nearby Polish roads are nowhere near as nice as Autobahns.

The bit caveat here is that most of the progress has been in a few areas along the coast. A majority of Chinese are still very poor and not really seeing the benefit of this technology.

Even the wealthiest provinces of China have lower GDP/capita than countries like Japan and Taiwan.

Can you point out anything specific where China is ahead in tech/IT? Are you sure your perspective is up to date?

From a European perspective there was a point where I was impressed how China was ahead in mobile payment. And I was impressed by Taiwans NFC payment tags. But today the QR code payment system seems quaint and outdated, and the mobile and NFC payment system we have seem as good if not better.

There's two kinds of technological progress: in depth and in breadth.

A country advances in depth by pouring lots of resources into very expensive facilities the likes of which might only exist a few times on Earth in order to make breakthroughs at the frontier of knowledge. I guess a "chip equipment R&D center" counts in this category.

A country advances in breadth by diffusing technology so widely it affects the life of basically everyone. When people talk about "welcoming another billion+ humans into living in an advanced industrial society" I think this is what they're imagining.

Advances in depth get easier with a larger population, because it increases the total resources available for megaprojects. Advances in breadth get harder because so many more people need to be covered.

I think comparisons that put China technologically ahead of Western industrialized nations are confusing depth for breadth.

For example, the International Federation of Robotics has a ranking of countries by ratio of robots to manufacturing workers https://ifr.org/ifr-press-releases/news/global-robotics-race... where China is in fifth place after Korea, Singapore, Germany and Japan. But then down below, they say that China has "a huge workforce of roughly 38 million people in the manufacturing industry" – in a country of more than a billion people, that number is clearly off by an order of magnitude. I guess they're excluding a lot of low-tech manufacturing that employs the vast majority of workers. So the comparison is only relevant for a small sliver of the population.

From the article:

But while Huawei's compensation package is generous, its working culture can be challenging, according to chip industry managers.

"Working with them is brutal. It's not 996 -- meaning working from 9 a.m. to 9 p.m., six days a week. ... It will literally be 007 -- from midnight to midnight, seven days a week. No days off at all," one Chinese chip engineer told Nikkei Asia. "The contract will be for three years, [but] the majority of people can't survive till renewal."

What does that even mean? Obviously people need to eat and sleep and can't literally work 24 hours a day.

You basically trade 3 years of your life for money. Just like the Ben Affleck's Paycheck.

I have no idea. I should have added a smiley.

The whole setup is out of this world: A vilified Chinese company at war with the full American empire. Its leader mumbling military clichees like "surround the cities from the country side". Its headquarters being a Disney-land of copied classical European architecture. And at the same time it is delivering breakthroughs that two years ago would have been thought to be impossible.

In that working culture it is often okay to sleep (at work, at your desk) for 2-4 hours. It is brutal, but people are willing to do it, so it persists.

I assume that's a mistake and it means "twelve hours a day, seven days a week"?

Either way, I have nothing but bad wishes for their new endeavors if that's the case - this is obviously not an acceptable way to treat people.

If they state it up front and people know what they are getting into. What’s the issue? If it’s so horrible, it won’t work right?

007 is a humorous mockery of the younger generation of Chinese workers who work overtime desperately in their companies.

The incompetence of Russian communists has given people the impression that only capitalism works, but the reality is that the Russian communists were simply too stupid and too hatred driven by their ideology. The simple truth is, it doesn't matter how you organize your economy, as long as it produces what people need and want. Economists see the failure of central planning as evidence that government meddling can never improve the economy, which they then project onto the next closest thing (e.g. China's market economy with extremely severe government intervention) and then they act surprised that it works. Winners are chosen by their performance and luck, not by (western or eastern) ideology.

The US handed China economic power and now is upset about it. In a tantrum we're now forcing them to develop some of the most advanced technology in the world. In both cases it seems to be a result of underestimating them. Elites thinking everyone else is just labor.

In the 90's a friend of mine from business school said outsource all production was the future and people in the US would just manage everything. I laughed. We've seen how engineering and manufacturing are linked. China had to learn engineering to do production better. And management? Did anyone seriously think they couldn't figure that out?

This result from cutting off fab equipment was blatantly obvious. What's less obvious is that they will do it cheaper than ASML does today by jumping straight to next generation technology.

I’m really curious if they can turn the investments into competitive process nodes. My understanding is that semiconductor technology has a very wide moat.

the moat is wide if the market was free. However, when it becomes a matter of national security, the economic viability becomes a far lower priority. China is obviously going to get there, just a matter of time.

Yes - as they say once the possible is known it is always easier to replicate

Indeed. Second mover advantage should be well known by now

Sure. But sanctions are fundamentally a tool to incur costs. China doesn't have an unlimited budget. If chips becomes a priority, something else will lose some funding, be it military or e. g. other industry subsidies.

Sure. Opportunity cost is there, I am sure China would much prefer to freely buy Western equipment. However, in case of escalating export controls, it clearly makes sense to further invest in domestic capacity. The 'opportunity gain' is more jobs for Chinese workers, more inward investment that would otherwise have gone into foreign countries.

There were intense arguments regarding whether to buy or to research in China. The "buying is better" party won and hence China lost ~30 years on chips R&D. But the embargo snapped right on the faces of the "buying is better" guys and leaving China no choice but to bite the bullet. So what those Chinese pioneers wanted but failed to achieve has now firmly implemented by the embargo...

US policy reflexive and unreflective when it comes to China.

Export controls on simply redirect investment inward to domestic R&D and subsequent production capacity. Effectively, this is defunding US and US-controlled companies, who otherwise would have hoovered up Chinese money, fed into its own R&D and maintained the lead on Chinese companies.

The policy doesn't even succeed on its own terms. But no iteration, no reflection

It's again the same opportunity cost, though. China has to invest a lot of money just get access at scale to the same class of chips as the rest of the world. That money could be used for actual innovation instead.

> The policy doesn't even succeed on its own terms.

The policy doesn't succeed only if you limit your thinking on the chip industry and ignore the overall strategic context. Reducing China's access to latest & greatest is only the proximate reason for the export controls, with the overall strategic competition being the ultimate reason. It's simply a mean to an end.

As you said in your previous post, China would likely prefer to continue buying western chips, because being forced to develop their own will not improve their strategic posture, at least in the next years/decades.

Indeed, China preferred buying from external sources as merchants won the debate on whether to buy or to research and it's indeed cheaper in short term.

Dependency on critical parts from external could be the only viable solution for smaller countries, but for countries like China is a intolerable risk to national security and has to be addressed at highest priority. The embargo just demonstrated how bad it could be, and it actually can be much worse in different contexts. So, obviously R&D is the only option left on the table. Cost is no longer a problem.

The reason behind is that US lost its confidence to compete. So if it cannot compete, it might be easier to remove the competitors. Such move would work on smaller countries such as Japan etc. But China is another story.

China does not have a choice, does it? The funny thing is that, whatever the west tried to prevent China from acquiring, in the end China managed to crack it and eventually made a better version.

I think it is already competitive. They are selling millions of the Huawei Mate 60 pro.

Feels like somehwere between not paying 4000% nvidia premium, and spamming gigawatts of cheap renewables to power data centres might be a competitive business plan if they get 12nm tools out in time. Strategic bonus of not sending 100s of billions to trillions to US for IP fees.

https://archive.md/vWHOs

Reply on news.ycombinator.com