4 ms·
The sudden leaps forward are the result of how technical development fundamentally works. As you master the foundations, you suddenly make huge leaps forward be
by FooBarWidget 3y ago
The sudden leaps forward are the result of how technical development fundamentally works. As you master the foundations, you suddenly make huge leaps forward because better foundations enable all sorts of new paths. For example, nuclear technology emerged all of a sudden, in multiple countries, because the foundations were there. People who don't understand this attribute way too much to stuff like "spying".
China made huge leaps lately not because they suddenly poured a lot of money in this effort. They poured money in this effort for multiple decades, but progress was slow because the foundations were not there. However, investments in previous decades built such a foundation, which is why they are now making a lot of progress.
- alephnerd 3y ago> However, investments in previous decades built such a foundation To add on to this, China entered the Fab space in the late 90s/early 2000s when vendors in MY, SG, SK, JP, TH, and TW began outsourcing low margins sectors such as Memory, Photovoltaic, Packaging, and OSAT to Guangdong+Fujian in the early 2000s due to falling margins It takes a generation to build the supply chain needed for precision fabrication, and in certain sectors of the Chip industry (eg. Design) China is still somewhat behind (though it is changing rapidly due to investments in the 2010s). I'm no wumao, but a lot of the criticism against China's R&D capacity on HN is just plain stupid and show a ridiculous lack of understanding of EECS, business, or East Asia.
- FooBarWidget 3y agoI think you shouldn't even need to add a disclaimer like "I'm no wumao" for stating things like that. You're not even being especially supportive of the Chinese government, just merely stating that they have a long history development and that there exist market- and power dynamics that favor China.
- alephnerd 3y agoI don't want to be downvoted to oblivion for stating facts, so I have to preface this. HN has severely degraded, and people are getting downvoted to oblivion for bringing this up.
- seanmcdirmid 3y agoThey aren’t talking about yields, so I assume they can make fast chips but suffer on yields. The same is true with jet turbines and why they are still using Russian ones in the military and western ones for their civilian planes. They can make high performance, but the fly time between overhaul is too low to be economical. But China can afford to subsidize $1000 chips, especially if that means not having any chips at all.
- alephnerd 3y ago^^^ why is this downvoted?!?! Nothing is wrong in this statement.
- seanmcdirmid 3y agoThat is just about mianzi, no worries. Anyways, after reading the articles, it sounds like they are using older ASML lithography equipment, rather than homemade equipment from SMEE. Which is strange, because they announced they can do 3nm, but they are only putting 28nm into production this year. https://abachy.com/news/chinas-28nm-lithography-machine-expected-2024 https://abachy.com/news/chinas-28nm-lithography-machine-expe...
- alephnerd 3y agoI mean it makes sense. Most vendors can do 3nm, but probably not at production scale other than ASML. 28nm is very worrying from a NatSec standpoint though - way more than leading edge nodes - as that enough to produce a Sandy Bridge or Maxwell type architecture, which meets the needs for most American defense applications (eg. drones, precision missiles, C-RAM, etc) Russia manufactures their Elbrus-8s architecture using 28nm lithography as well, and that's the backbone for most of their newer precision, electronics, and avionics weapons systems.
- araes 3y agoThe opposite view on that state, is that spying is effective, because many technologies require huge amounts of work by the developer, yet can be stolen relatively easily with a document overview theft. The same reason there's all those jokes in "Silicon Valley" (the TV show) about the "brain rape" episodes and End Frame of Season 2, where all they do is lure Richard into a meeting and then get him to put the whole algorithm up on the board. [2] https://en.wikipedia.org/wiki/Silicon_Valley_(season_2)#Episodes https://en.wikipedia.org/wiki/Silicon_Valley_(season_2)#Epis... ("Runaway Devaluation")
- alephnerd 3y ago> stolen relatively easily with a document overview theft Nope. There are too many moving parts to make this realistic. You'd have to steal the IP of hundreds of vendors across an entire supply chain AND THEN finance. It's a hurculean and unrealistic task.
- FooBarWidget 3y agoLet's do a thought experiment in an area the HN crowd is more familiar with. Gitlab is open source. Heck even all their internal documents are open. Can you become an effective Gitlab competitor just by copying their stuff? Let's say Slack's source code got leaked. Can you become an effective Slack competitor just by copying them? Now imagine that the semiconductor industry is many times more complex. Also look at what the Chinese semiconductor industry is actually doing. Anybody who pays attention sees that they're trying out new things, choosing solution directions that are free from foreign IP. Chinese fabs are replacing foreign equipment with Chinese (incompatible) equipment. Why would they do all that instead of doubling down on hacking and stealing? Why would Chinese equipment not be drop-in-compatible if copying and stealing is so lucrative? Why would Huawei drop all Android support, instead opting to design their entire OS stack from the ground up in an incompatible manner, instead of leeching from open source Android forever?