5 ms·

And thanks to all that effort the Chinese now likely have all the blueprints. The only reason why they still don't have EUV is because this stuff is difficult t

by m0zg 5y ago

And thanks to all that effort the Chinese now likely have all the blueprints. The only reason why they still don't have EUV is because this stuff is difficult to bring up even if you have all the documentation - screw something up even a little, and it's not going to work. But it certainly isn't going to take them 20 years to replicate it.

jillesvangurp 5y ago
The nature of patents is that it is mostly public information anyway and free to use after the patents expire. That works both ways of course, the Chinese have some pretty fancy technology in other fields that manufacturers in the EU or US might benefit from. Of course with chips, patents and know how that e.g. Intel or ASML have are two things. Ultimately this is good as it drives companies to continue innovating. Intel's current problems are related to them forgetting to do that.
- m0zg 5y ago
  Precisely. A billion ways to get things wrong, and if you get it wrong it doesn't work. And the info to get it to work is unlikely to be in the patents. You have to have folks who worked with this stuff before, which is difficult, but doable. And as far as "innovating" in this field, we're almost at the end of the run - there are, as far as I'm aware, no processes planned beyond 3nm, and even 3nm is astonomically complex and expensive.
- 988747 5y ago
  >> patents and know how that e.g. Intel or ASML have are two things. Shouldn't that make the patent invalid? I thought the requirement is that the patent should give enough info so that "person skilled in the art" can reproduce the invention.
  - swiftcoder 5y ago
    "skilled in the art" is somewhat irrelevant if that skill takes a couple of decades and a $120 million machine to develop...
gostsamo 5y ago
At the time I worked for an ASML contractor, the joke was that the chinese already have all the ASML code, just they couldn't compile it. This sounded so true, because the ASML code is the worst pile of c built over almost 40 years that the world has ever seen and because the chinese were rather blatant in their attempts to steal whatever they can from ASML.
- mhh__ 5y ago
  > the chinese were rather blatant in their attempts to steal whatever they can from ASML. Any stories you have (and are allowed to) tell?
  - gostsamo 5y ago
    Not personal, but chinese agents would wait in front of the ASML buildings and target ethnic chinese workers offering them money and honor if they come back to the motherland and work there. Stories about stolen laptops and so on.
scyzoryk_xyz 5y ago
This reminds me. I have an anecdote from my 90-something year old uncle who coordinated and managed construction efforts for the power grid in a post-WWII socialist-bloc country between 1940-1990. At first they relied on the blueprints left behind by Germany, and then they sourced their power generators and components from the USSR. But within several decades, it became clear that the USSR stuff is just bad. Like straight up unreliable and faulty bad. So the question became about whether to build their own, which the engineers were very confident they could do. Although it seemed feasible in principle, after many years of failures it just turned out that manufacturing these things was way too difficult to do reliably domestically. They had underestimated the entire intellectual investment necessary. Buying from the West, from manufacturers with decades of experience became the only option. And this was a long time before the iron curtain even started to collapse. And this was a process managed by socialist/communist loyalists. They were buying tech from the West while still projecting an image of strength and control. Regimes trying to copy something strongly reminds me of that anecdote. They usually manage to waste a lot of resources only just barely catching up, copying, and then they still end up buying the stuff that works from the companies that have the decades and the experience doing it. Making this chip tech sounds orders of magnitude more complex. Reading the article I only think about what a major fumble it was geo-strategically that the people with good ideas went to Taiwan.
- m0zg 5y ago
  China is in a _much_ different place than the USSR was: for one thing hundreds of thousands of Chinese have advanced degrees from Western universities. A lot of them have industrial experience as well (as well as probably copies of trade secrets they could divulge if the CCP really insists - we even catch the dumbest of them in the process of doing so from time to time). As if this wasn't enough, the West has outsourced darn near everything manufacturing related to China. If they want 5nm, they'll have 5nm, within 5 years or less. And if they can't have it, they'll just take Taiwan.

by m0zg 5y ago

jillesvangurp 5y ago

The nature of patents is that it is mostly public information anyway and free to use after the patents expire. That works both ways of course, the Chinese have some pretty fancy technology in other fields that manufacturers in the EU or US might benefit from. Of course with chips, patents and know how that e.g. Intel or ASML have are two things. Ultimately this is good as it drives companies to continue innovating. Intel's current problems are related to them forgetting to do that.

m0zg 5y ago
Precisely. A billion ways to get things wrong, and if you get it wrong it doesn't work. And the info to get it to work is unlikely to be in the patents. You have to have folks who worked with this stuff before, which is difficult, but doable. And as far as "innovating" in this field, we're almost at the end of the run - there are, as far as I'm aware, no processes planned beyond 3nm, and even 3nm is astonomically complex and expensive.
988747 5y ago
>> patents and know how that e.g. Intel or ASML have are two things. Shouldn't that make the patent invalid? I thought the requirement is that the patent should give enough info so that "person skilled in the art" can reproduce the invention.
- swiftcoder 5y ago
  "skilled in the art" is somewhat irrelevant if that skill takes a couple of decades and a $120 million machine to develop...

gostsamo 5y ago

At the time I worked for an ASML contractor, the joke was that the chinese already have all the ASML code, just they couldn't compile it. This sounded so true, because the ASML code is the worst pile of c built over almost 40 years that the world has ever seen and because the chinese were rather blatant in their attempts to steal whatever they can from ASML.

mhh__ 5y ago
> the chinese were rather blatant in their attempts to steal whatever they can from ASML. Any stories you have (and are allowed to) tell?
- gostsamo 5y ago
  Not personal, but chinese agents would wait in front of the ASML buildings and target ethnic chinese workers offering them money and honor if they come back to the motherland and work there. Stories about stolen laptops and so on.

scyzoryk_xyz 5y ago

This reminds me. I have an anecdote from my 90-something year old uncle who coordinated and managed construction efforts for the power grid in a post-WWII socialist-bloc country between 1940-1990. At first they relied on the blueprints left behind by Germany, and then they sourced their power generators and components from the USSR. But within several decades, it became clear that the USSR stuff is just bad. Like straight up unreliable and faulty bad. So the question became about whether to build their own, which the engineers were very confident they could do. Although it seemed feasible in principle, after many years of failures it just turned out that manufacturing these things was way too difficult to do reliably domestically. They had underestimated the entire intellectual investment necessary. Buying from the West, from manufacturers with decades of experience became the only option. And this was a long time before the iron curtain even started to collapse. And this was a process managed by socialist/communist loyalists. They were buying tech from the West while still projecting an image of strength and control. Regimes trying to copy something strongly reminds me of that anecdote. They usually manage to waste a lot of resources only just barely catching up, copying, and then they still end up buying the stuff that works from the companies that have the decades and the experience doing it. Making this chip tech sounds orders of magnitude more complex. Reading the article I only think about what a major fumble it was geo-strategically that the people with good ideas went to Taiwan.

m0zg 5y ago
China is in a _much_ different place than the USSR was: for one thing hundreds of thousands of Chinese have advanced degrees from Western universities. A lot of them have industrial experience as well (as well as probably copies of trade secrets they could divulge if the CCP really insists - we even catch the dumbest of them in the process of doing so from time to time). As if this wasn't enough, the West has outsourced darn near everything manufacturing related to China. If they want 5nm, they'll have 5nm, within 5 years or less. And if they can't have it, they'll just take Taiwan.