The usual argument made against the strategy of regulation is that the economic pressures would drive production underground, possibly literally, frustrating any attempt at regulation or enforcement.
My point is that other economic pressures force consolidation into a few efficient and extremely expensive & vulnerable facilities, which can be affected by ordinary military mechanisms—hence, regulation + military-level enforcement may well work in contrast to the usual pessimism.
I have got to say that when you’re talking about capital requirements in the billions, and construction of the sort where you use the world’s largest cranes, and a need for clean-room facilities, I have a really hard time seeing how production could go underground. Explosives, drugs, and weaponised bacteria, yes; these are all technologies that were known well before 1800, you’ll note. Chips? I really don’t see it.
Did you perhaps mean that the fabricators will go for unregulated markets, but build openly there? Possible, but they still have to sell their products. I suggest that the usual smuggling paths are not going to be very useful here, because where is the demand for forbidden hardware? Drugs go through porous borders because they’ll make money on the other side. But high-performance chips? Are you suggesting a black market of AI researchers, willing to deal with criminal syndicates if only they can get their fix of ever higher performance? And even if you are, note again the capital requirements. Any dirt farmer can set up to grow opium, and many do. If drugs required as much capital, and all in one place at that, as a modern fab does, I suggest that the War On Drugs would be over pretty quickly.
Really, hardening against nukes seems like a completely wrong approach unless you’re suggesting that the likes of China would be hiding this fab from the US. For private actors the problem is in finding, not blowing up. If you have military-level enforcement, nukes (!) are just total overkill; send in a couple of platoons of infantry and be done. What are the illegal fabbers going to do, hire the Taliban as plant security? (I mean, maybe they would, and we all know how that fight would go, right?)
I think you’ve got, not the wrong end of the stick, but the wrong stick entirely, when you start talking about nuclear hardening.
It does seem implausible that any non-state actor could setup an underground chip fab; but someone is going to suggest just that, despite the idiocy of the suggestion when you look at what cutting-edge chip fabs entail, so I have to argue against it. With that argument knocked out of the way, the next issue is whether a state actor could set up a cutting edge chip fab hardened or underground somehow, which is not quite so moronic and also needs to be argued against.
Ok, but then why not point out what I just did, which is that anyone but a state building such a thing is quite implausible? Pointing out that it takes a lot of hardening to protect against nukes just seems like a mis-step.
I think perhaps you need to make clearer who you envision as wanting to stop Moore’s Law. Are we talking about private actors, ie basically terrorists with the resulting budget limitations; or states with actual armies?
The usual argument made against the strategy of regulation is that the economic pressures would drive production underground, possibly literally, frustrating any attempt at regulation or enforcement.
My point is that other economic pressures force consolidation into a few efficient and extremely expensive & vulnerable facilities, which can be affected by ordinary military mechanisms—hence, regulation + military-level enforcement may well work in contrast to the usual pessimism.
I have got to say that when you’re talking about capital requirements in the billions, and construction of the sort where you use the world’s largest cranes, and a need for clean-room facilities, I have a really hard time seeing how production could go underground. Explosives, drugs, and weaponised bacteria, yes; these are all technologies that were known well before 1800, you’ll note. Chips? I really don’t see it.
Did you perhaps mean that the fabricators will go for unregulated markets, but build openly there? Possible, but they still have to sell their products. I suggest that the usual smuggling paths are not going to be very useful here, because where is the demand for forbidden hardware? Drugs go through porous borders because they’ll make money on the other side. But high-performance chips? Are you suggesting a black market of AI researchers, willing to deal with criminal syndicates if only they can get their fix of ever higher performance? And even if you are, note again the capital requirements. Any dirt farmer can set up to grow opium, and many do. If drugs required as much capital, and all in one place at that, as a modern fab does, I suggest that the War On Drugs would be over pretty quickly.
Really, hardening against nukes seems like a completely wrong approach unless you’re suggesting that the likes of China would be hiding this fab from the US. For private actors the problem is in finding, not blowing up. If you have military-level enforcement, nukes (!) are just total overkill; send in a couple of platoons of infantry and be done. What are the illegal fabbers going to do, hire the Taliban as plant security? (I mean, maybe they would, and we all know how that fight would go, right?)
I think you’ve got, not the wrong end of the stick, but the wrong stick entirely, when you start talking about nuclear hardening.
It does seem implausible that any non-state actor could setup an underground chip fab; but someone is going to suggest just that, despite the idiocy of the suggestion when you look at what cutting-edge chip fabs entail, so I have to argue against it. With that argument knocked out of the way, the next issue is whether a state actor could set up a cutting edge chip fab hardened or underground somehow, which is not quite so moronic and also needs to be argued against.
Ok, but then why not point out what I just did, which is that anyone but a state building such a thing is quite implausible? Pointing out that it takes a lot of hardening to protect against nukes just seems like a mis-step.
I think perhaps you need to make clearer who you envision as wanting to stop Moore’s Law. Are we talking about private actors, ie basically terrorists with the resulting budget limitations; or states with actual armies?