Live data from Hacker News

How China built its ‘Manhattan Project’ to rival the West in AI chips

japantimes.co.jp

61–70 of 645 posts

Re: How China built its ‘Manhattan Project’ to rival the West in AI chips

#61

A "Manhattan Project" would be building some shocking new technology that didn't previously exist. If they're cobbling together old parts, it sounds more like something you'd to to keep things running in case a conflict erupts: > The availability of parts from older ASML machines on secondary markets has allowed China to build a domestic prototype

They are acquiring parts to reverse engineer them and build their own

Re: How China built its ‘Manhattan Project’ to rival the West in AI chips

#62
post #55

I'd argue ASML's moat isn't the machine itself but the ecosystem: Carl Zeiss optics, decades of supplier relationships, institutional knowledge. This is clearly a significant achievement, but does anyone with semiconductor experience have a sense of how far "generates EUV light" is from "production-ready tool"?

They are nowhere close to beat ASML. This isn't a moat ASML can keep for long though. There can be alternatove technologies to achieve the same goal. So far only China has that incentive. The real problem is process scaling is slowing down. How many more generations of lithography machines will ASML design? Probably not many. This means there will be no edge left in 5 or 10 years, as eventually brute force will work…

No, they won't beat ASML but they'll be good enough and most importantly cheap. And they'll catch up eventually.

Re: How China built its ‘Manhattan Project’ to rival the West in AI chips

#64
post #36

With Nvidia scaling down their consumer GPU production [0] I wonder if we will see consumer GPUs shipping from China in the future. Western companies seem to be abandoning the consumer/prosumer market which will have bad implications for hobbyists and aspiring professionals down the line. [0] https://www.pcmag.com/news/nvidia-might-cut-rtx-50-gpu-suppl...

It’s a good thing that Chinese companies have zero expertise in leveraging consumer demand for lower-end tech to develop know-how and catch up with the state of the art from Western-aligned companies and then economies of scale to surpass them in distribution.

Exactly. That's where this is heading, and the West-- as usual-- is pursuing quarterly profits and forgetting to look up.

Re: How China built its ‘Manhattan Project’ to rival the West in AI chips

#65

[flagged]

China seems to be doing well on supply chain integration (with the exception of the trust part).

Being how strategic this is, I imagine that the investment won't be entirely laissez faire and there will be lower tolerance for cheating in this endeavor. I think that ultimately they'll do quite well with their efforts.

Re: How China built its ‘Manhattan Project’ to rival the West in AI chips

#66

The knowledge came from former ASML employees. I wonder if countries will sanction these individuals given the geopolitical implications of their assistance.

Well real question is how much would that limit PRC talent from working abroad. PRC will be producing plurality of STEM / high skilled talent for decades. They're going to be the only country with project intergrated circuit talent glut in next 10 years, every other semi power projected to have 100,000s shortage. No PRC talent, and you cap western semi talent pool.

Ultimately a lot western innovation run on brain drained PRC talent. There is bamboo ceiling in western tech for east asians, specifically to restrict reverse knowledge transfer. Side effect is once PRC talent hits this ceiling they know big title and fat paychecks and upward mobility is back home, where frankly QoL is off the charts. Ultimately PRC wealthy enough to reverse brain drain aka brain recirculation and PRC talent aren't retarded enough to limit their career aspirations because west decides to cap their career trajectory and try to lock their future behind noncompetes, especially in cold war vs their birth country. Worse, PRC wealthy enough even if there's no bamboo ceiling they can afford to reverse brain drain top 1%, hence current equilibirum. West needs PRC talent, west cannot afford PRC talent to climb too high, PRC can afford to take them off west's hands.

Until west figures out another source of talent, they're stuck in this talent trap. And IMO India ain't it, they don't have the integrated industrial chains and academic structure to produce same kind industrial ready workers yet.

Re: How China built its ‘Manhattan Project’ to rival the West in AI chips

#67

The knowledge came from former ASML employees. I wonder if countries will sanction these individuals given the geopolitical implications of their assistance.

Handing out sanctions without at least a plausible legal cover, sounds like a recipe for disaster that would come back to bite. I wonder what could be used here, non-compete? IP infringement? Or doing it "for all mankind"? As for knowledge, the YouTube channel Branch Education explained EUV lithography in great detail, sponsored by ASML itself. My impression is that the knowledge is not that secretive, the precision…

Yeah it reminds me of the Smyth report, published in August 1945 about atomic bombs, commissioned by the director of the real Manhattan Project. It’s fine to reveal knowledge in detail, if it doesn’t reveal anything related to constructing the apparatuses (the chemistry and the metallurgy) needed.

Re: How China built its ‘Manhattan Project’ to rival the West in AI chips

#68

With Nvidia scaling down their consumer GPU production [0] I wonder if we will see consumer GPUs shipping from China in the future. Western companies seem to be abandoning the consumer/prosumer market which will have bad implications for hobbyists and aspiring professionals down the line. [0] https://www.pcmag.com/news/nvidia-might-cut-rtx-50-gpu-suppl...

You don't need CUDA for gaming but software is still just as big of a moat. Gaming GPU drivers are complex and have tons of game-specific patches. With their new Radeon/RDNA architecture it took AMD years to overcome their reputation for having shitty drivers on their consumer GPUs (and that reputation was indeed deserved early on). And I bet if you go read GPU discussion online today you'll still find people who avo…

On the other hand, all it would take would be one successful Steam Deck/Steam Machine-style console to get all the developers of the world making sure that their games work on that hypothetical GPU.

I don't think that it will happen in the next 5 years, but who knows?

Re: How China built its ‘Manhattan Project’ to rival the West in AI chips

#69

Earlier quoted context omitted.

How will that work, exactly? The chip makers are going to have a list of approved "cloud providers" and they will refuse to sell to anyone else?

Cloud providers will use cheap investment capital to buy chips at increasing prices, while the public will be economically forced to get computational services from these cloud providers. After a few years, most software will work only when connected to cloud infrastructure, either for performance or for "safety" reasons. We're already seeing this with AI.

aka "return of the Minitel"

Re: How China built its ‘Manhattan Project’ to rival the West in AI chips

#70

With Nvidia scaling down their consumer GPU production [0] I wonder if we will see consumer GPUs shipping from China in the future. Western companies seem to be abandoning the consumer/prosumer market which will have bad implications for hobbyists and aspiring professionals down the line. [0] https://www.pcmag.com/news/nvidia-might-cut-rtx-50-gpu-suppl...

You don't need CUDA for gaming but software is still just as big of a moat. Gaming GPU drivers are complex and have tons of game-specific patches. With their new Radeon/RDNA architecture it took AMD years to overcome their reputation for having shitty drivers on their consumer GPUs (and that reputation was indeed deserved early on). And I bet if you go read GPU discussion online today you'll still find people who avo…

Cuda is 20 years old and it shows. Time for a new language that fixes the 20 years of rough edges. The Guy (Lattner) who made LLVM is working on this: https://www.modular.com/mojo

Good podcast on him: https://newsletter.pragmaticengineer.com/p/from-swift-to-moj...

Post reply on HN