I had got halfway through the Kill Chain a couple years back, and decided to pick it up again after the war in Ukraine started. It's fascinating just how much it seems to be written in a different world. Obviously the goal of the book was to jolt the US establishment out of complacency, so it presented some best-case scenarios for the Russian and Chinese militaries, but man comparing the Russian Army of that book to the one we saw in real life is extraordinary. A good reminder of how little we often know about others' capabilities, or our own as well - and how the error could really go in either direction.
"To me, this debate shows just how difficult it is to assess national leadership and capabilities in AI — really, the technology is just too new to make that kind of judgement. We don’t really know exactly how AI will be used on the battlefield of a future war, and thus we don’t know which measurements of AI capabilities are most important for assessing national power."
The fact that the U.S. government has let Taiwan, a small island nation of twenty four million people off the coast of China, a hostile power of 1.4 billion people, become the world's epicenter of microchip technology is shameful and a national security risk. We see the military risks of letting microchips be beholden to foreign powers, see Russia's desperate attempts to reclaim microchips from appliances to build military weapons systems in light of the sanctions, as a strategic risk.
All current AI or ML technologies are running on specialized microchips in the form of GPUs, of which the preponderance come from Taiwan, whether NVIDIA or AMD. On-shoring the manufacture of these designs, be it in Intel fabs in Arizona or the proposed fab in Ohio, should be of primary concern. The fact that the U.S. is spending $858,000,000,000 in the latest military budget yet is unconcerned about the sourcing of these processors is disheartening at best.
The US government didn't have much impact on that one way or another. The US has historically dominated in both chip design and manufacturing, TSMC only got a process edge quite recently and only due to some very complex mis-steps by Intel. There wasn't anything the US government could do about that one way or another - it can't ensure Intel never loses a tech bet.
"The fact that the U.S. is spending $858,000,000,000 in the latest military budget yet is unconcerned about the sourcing of these processors is disheartening at best."
There is one guy cited there as running the program. Also it looks like he was allocated $400,000,000 a year over five years to implement nine (9) technology research centers nationwide to work on the chip problem for the military. I am not saying this is nothing, but what can we expect from this program?
Calling Taiwan, "the world's epicenter of microchip technology," is absurd. They have the most advanced fabrication plant on the planet yes. But they lack all of the design expertise (mostly concentrated in the U.S.) and the lithography technology (Netherlands). They are a key player, yes. They are most assuredly not the epicenter of anything.
I heartily endorse your review of Chip Wars. I believe that I started on it after a prior mention of the book by you. The chapter on Intel's decline makes clear the flaw of depending only on financial metrics to guide firm decision-making. A manager who does not have operational experience cannot adequately weigh the merits of internal investment versus outsourcing. Investments in R&D are particularly vulnerable to cutting in the name of "profit maximization" which robs firms of their future in a technologically dynamic area like microelectronics.
„to spy on everybody who uses its equipment“. That’s not technically possible. If there were any ability in the 5G chips to „spy“ the technical details would be published. It’s not possible - the 5G chips in the cells just handle very low level bits and bytes and pass them through to the back haul - the wired connections. Anything that slows that down would make the chips inoperable.
The best chance Huawei have if spying would be a man in the middle attack on their routers, but there’s no evidence of that either and as far as I can see and no desire to ban them. The war against Huawei here is just protectionist.
It also says „ The Journal investigation didn’t turn up evidence of spying by or on behalf of Beijing in Africa. Nor did it find that Huawei executives in China knew of, directed or approved the activities described.***It also didn’t find that there was something particular about the technology in Huawei’s network that made such activities possible.***“
If there was such technology it couldn’t be hidden.
There are ways to spy using cell towers but it involves putting up fake cell equipment close to the people you hope to spy on. And taking the data away from the general internet to analyse.
A normal conversation or data share driven from a phone can swap cell towers multiple times particularly when on the move. There are easier ways.
Yeah, that's the big glaring issue with the narrative that Huawei is a threat. 12 years ago that would have been absolutely a genuine concern but then Snowden showed the world just how deeply and aggressively governments were hacking network infrastructure. There was a massive push across the tech industry starting from that exact moment to encrypt everything.
Nowadays, what's the most Huawei could do, realistically? They can't spy on everyone globally all the time because all that data has to get back to China somehow and people would notice - if you're paying for a certain link speed and can't get it, you're going to investigate and notice all these packets flowing to Chinese IP addresses even though there shouldn't be any. So it'd have to be a very targeted, rarely used back door that lets them ... do what? See a graph of traffic patterns over a specific link? That's not useless but it's also far from a strategic capability on par with being able to manufacture high end components.
There are a few other things they can do: hide the origins of attacks, disconnect systems at strategic moments and so on. But the moment they actually *do* any of those things, people would take measures to prevent it through reconfigurations of non-Huawei equipment, so it seems like nothing very important.
They could also attempt to backdoor or otherwise compromise the encryption systems used in the west. But we're already well aware of that threat because the NSA was doing that for years.
A lot of the hype about Huawei feels like it's coming from senior civil servants who are remembering briefings they were getting 10-15 years ago, when they started subverting the largely unencrypted internet on a massive scale and this firehose of intel and capabilities opened up to them. Our best available understanding (in the tech industry) is that whilst endpoint hacking is still a serious threat the encryption systems we use are strong and governments don't know how to break them, nor will they any time soon. My strong suspicion is that the mid/lower management tiers inside the intelligence agencies haven't been entirely communicative about the steady slide into uselessness that a lot of their older network-based capabilities have experienced, and the absence of something is harder to spot than the presence. So governments are busy fretting about yesterday's threat when the battlefield has already changed around them.
It depends on what is meant by spying. A 5G tower station would have a hard time inspecting packets of encrypted TCP/IP data without being detected, but cataloging which devices connect to it at which times or inspecting SMS or even voice packets and then back-channeling that data seems technically possible without slowing down significantly. As far as hiding this capability it could be hidden during testing (like the VW diesel gate incident) or selectively enabled when needed. I don’t know how possibly this is, but feeling safe to depend on their equipment because you believe their technical publications seems risky. If a whole country uses only their equipment then you’re susceptible at least to them having a latent built in kill switch which could take out your whole network at a time of crisis.
In regards to CHIPS Ben Thompson at Stratechery makes a compelling case that China is capable of building trailing edge chips(7NM I believe) and dominating that market. While not as advanced they go into most of the electronics, cars, etc that people use every day. He posits that that is more important than leading edge chips and could cause a lot of problems if they end up becoming the source for trailing edge. Do you have any thoughts or opinions about that? Have you heard anything different?
I've seen this, too, and it brings an important point to mind: You don't need to have the best in tech in order to succeed at your goals.
If their trailing edge chips are widely adopted, even only in the context of non-Western nations outside the bounds of US & European allies and near-allies, China can create their own staggered hemisphere of influence. It's another avenue down the Belt & Road initiative that largely failed, as I understand it.
Edit: There's also the danger of intermixed technologies in mil-tech, say in a nation like Turkey or India that has ties to both the US and China. That opens us up that much more to spying on our bleeding edge chip tech (and more, I'd imagine).
Hi Noah and all, for the book Wireless War the critique is mentioned that the workings of the telecom industry is not explained to the reader and because of this it is more difficult to read for the people who are not familiar with this. Could any of you suggest a book similarly written as the other book Chip Wars, but for this topic? As a lay person I would be happy to get hold of a good book on the subject. Thanks very much!
Nilay Patel has a great interview with the guy running the biggest deployment of ORAN cellular technology (the openness you referred to in your cellular book review). It's fantastic - this dude is brilliant.
Well, interest in such military technologies would seem to be built upon an assumption that conflicts between the great powers will remain limited, such as we see in Ukraine. Hopefully that will remain true for as long as possible, but it's just not credible to assume that such conflicts will never spin out of control in to nuclear exchanges. Sooner or later, one way or another, some limited conflict is going to spin out of control, and then who has the best chips etc isn't going to matter much.
If you accept the premise that we can't keep nukes around forever and never use them, then there is a sense in which none of this matters today. As example, if we're headed towards a nuclear conflict between the great powers, does it really matter who wins in Ukraine?
I really don't disagree, and have had a large Ukrainian flag flying on the front of our home since the first week of the war. Please see my comment as a rhetorical flourish whose purpose is to inspire deeper questioning.
The technical details described above on this page will only matter until such time as some limited conflict spins out of control. The WHEN of a nuclear holocaust is hypothetical, agreed. But on the current course, the IF is not hypothetical. We have the option to not stay on the current course.
Noah, You need to learn about "Software Defined Network (SDN)" architectures. They separate network control from data forwarding. The former is the sensitive part, the latter the expensive part (it is what gets deployed everywhere), and they may be made by separate vendors. 5G uses a SDN architecture. I don't follow this any more, but I suspect that we need to worry about the security of the former but much less about the latter. Huawei's dominance of the non-Western market gives them economies of scale in the latter, much less so in the former. I think the GCHQ (Britain's NSA) had a non-classified analysis of this.
I would add on the AI point that you are right:AI is a technology in its infancy. The big area is AUGMENTED Cognition AI, the use of AI to do tasks that humans are not cognitively good at (cognitive including motor as well). We have just began to figure out ways to augment our world (a point I think Noah made previously).
I had got halfway through the Kill Chain a couple years back, and decided to pick it up again after the war in Ukraine started. It's fascinating just how much it seems to be written in a different world. Obviously the goal of the book was to jolt the US establishment out of complacency, so it presented some best-case scenarios for the Russian and Chinese militaries, but man comparing the Russian Army of that book to the one we saw in real life is extraordinary. A good reminder of how little we often know about others' capabilities, or our own as well - and how the error could really go in either direction.
Yeah. Everything written before 2020-21 sort of seems like it's from another age.
This is correct along so many dimensions that it's mind-boggling.
You never know when a black swan will appear.. but when it does, boy does it change things huh?
Russia is never as strong as it looks, or as weak as it looks.
"To me, this debate shows just how difficult it is to assess national leadership and capabilities in AI — really, the technology is just too new to make that kind of judgement. We don’t really know exactly how AI will be used on the battlefield of a future war, and thus we don’t know which measurements of AI capabilities are most important for assessing national power."
The fact that the U.S. government has let Taiwan, a small island nation of twenty four million people off the coast of China, a hostile power of 1.4 billion people, become the world's epicenter of microchip technology is shameful and a national security risk. We see the military risks of letting microchips be beholden to foreign powers, see Russia's desperate attempts to reclaim microchips from appliances to build military weapons systems in light of the sanctions, as a strategic risk.
All current AI or ML technologies are running on specialized microchips in the form of GPUs, of which the preponderance come from Taiwan, whether NVIDIA or AMD. On-shoring the manufacture of these designs, be it in Intel fabs in Arizona or the proposed fab in Ohio, should be of primary concern. The fact that the U.S. is spending $858,000,000,000 in the latest military budget yet is unconcerned about the sourcing of these processors is disheartening at best.
The US government didn't have much impact on that one way or another. The US has historically dominated in both chip design and manufacturing, TSMC only got a process edge quite recently and only due to some very complex mis-steps by Intel. There wasn't anything the US government could do about that one way or another - it can't ensure Intel never loses a tech bet.
"The fact that the U.S. is spending $858,000,000,000 in the latest military budget yet is unconcerned about the sourcing of these processors is disheartening at best."
That's just not the case. https://www.cto.mil/ct/microelectronics/
There is one guy cited there as running the program. Also it looks like he was allocated $400,000,000 a year over five years to implement nine (9) technology research centers nationwide to work on the chip problem for the military. I am not saying this is nothing, but what can we expect from this program?
This isn't the full extent of the interest or activity; its just the first hit I found on google.
Calling Taiwan, "the world's epicenter of microchip technology," is absurd. They have the most advanced fabrication plant on the planet yes. But they lack all of the design expertise (mostly concentrated in the U.S.) and the lithography technology (Netherlands). They are a key player, yes. They are most assuredly not the epicenter of anything.
Thank you so much! I've added all of these to my reading list, and I appreciate the recommendations and the breakdowns.!
My dad is reading chip wars now and really enjoying it. Anytime you highlight books, I perk up.
I learned a ton about the chip world (semiconductor fab stuff, in particular) in grad school; I'll definitely read Chip Wars and see how it resonates.
I heartily endorse your review of Chip Wars. I believe that I started on it after a prior mention of the book by you. The chapter on Intel's decline makes clear the flaw of depending only on financial metrics to guide firm decision-making. A manager who does not have operational experience cannot adequately weigh the merits of internal investment versus outsourcing. Investments in R&D are particularly vulnerable to cutting in the name of "profit maximization" which robs firms of their future in a technologically dynamic area like microelectronics.
„to spy on everybody who uses its equipment“. That’s not technically possible. If there were any ability in the 5G chips to „spy“ the technical details would be published. It’s not possible - the 5G chips in the cells just handle very low level bits and bytes and pass them through to the back haul - the wired connections. Anything that slows that down would make the chips inoperable.
The best chance Huawei have if spying would be a man in the middle attack on their routers, but there’s no evidence of that either and as far as I can see and no desire to ban them. The war against Huawei here is just protectionist.
Nah. In Wireless Wars, Pelson gives some actual cases where the Huawei equipment was used to listen in on African governments. It's real.
Looking for information on that I just see that in Uganda Huawei helped break into a WhatsApp group. From here.
https://www.wsj.com/articles/huawei-technicians-helped-african-governments-spy-on-political-opponents-11565793017
It also says „ The Journal investigation didn’t turn up evidence of spying by or on behalf of Beijing in Africa. Nor did it find that Huawei executives in China knew of, directed or approved the activities described.***It also didn’t find that there was something particular about the technology in Huawei’s network that made such activities possible.***“
If there was such technology it couldn’t be hidden.
There are ways to spy using cell towers but it involves putting up fake cell equipment close to the people you hope to spy on. And taking the data away from the general internet to analyse.
A normal conversation or data share driven from a phone can swap cell towers multiple times particularly when on the move. There are easier ways.
Yeah, that's the big glaring issue with the narrative that Huawei is a threat. 12 years ago that would have been absolutely a genuine concern but then Snowden showed the world just how deeply and aggressively governments were hacking network infrastructure. There was a massive push across the tech industry starting from that exact moment to encrypt everything.
Nowadays, what's the most Huawei could do, realistically? They can't spy on everyone globally all the time because all that data has to get back to China somehow and people would notice - if you're paying for a certain link speed and can't get it, you're going to investigate and notice all these packets flowing to Chinese IP addresses even though there shouldn't be any. So it'd have to be a very targeted, rarely used back door that lets them ... do what? See a graph of traffic patterns over a specific link? That's not useless but it's also far from a strategic capability on par with being able to manufacture high end components.
There are a few other things they can do: hide the origins of attacks, disconnect systems at strategic moments and so on. But the moment they actually *do* any of those things, people would take measures to prevent it through reconfigurations of non-Huawei equipment, so it seems like nothing very important.
They could also attempt to backdoor or otherwise compromise the encryption systems used in the west. But we're already well aware of that threat because the NSA was doing that for years.
A lot of the hype about Huawei feels like it's coming from senior civil servants who are remembering briefings they were getting 10-15 years ago, when they started subverting the largely unencrypted internet on a massive scale and this firehose of intel and capabilities opened up to them. Our best available understanding (in the tech industry) is that whilst endpoint hacking is still a serious threat the encryption systems we use are strong and governments don't know how to break them, nor will they any time soon. My strong suspicion is that the mid/lower management tiers inside the intelligence agencies haven't been entirely communicative about the steady slide into uselessness that a lot of their older network-based capabilities have experienced, and the absence of something is harder to spot than the presence. So governments are busy fretting about yesterday's threat when the battlefield has already changed around them.
It depends on what is meant by spying. A 5G tower station would have a hard time inspecting packets of encrypted TCP/IP data without being detected, but cataloging which devices connect to it at which times or inspecting SMS or even voice packets and then back-channeling that data seems technically possible without slowing down significantly. As far as hiding this capability it could be hidden during testing (like the VW diesel gate incident) or selectively enabled when needed. I don’t know how possibly this is, but feeling safe to depend on their equipment because you believe their technical publications seems risky. If a whole country uses only their equipment then you’re susceptible at least to them having a latent built in kill switch which could take out your whole network at a time of crisis.
In regards to CHIPS Ben Thompson at Stratechery makes a compelling case that China is capable of building trailing edge chips(7NM I believe) and dominating that market. While not as advanced they go into most of the electronics, cars, etc that people use every day. He posits that that is more important than leading edge chips and could cause a lot of problems if they end up becoming the source for trailing edge. Do you have any thoughts or opinions about that? Have you heard anything different?
I've seen this, too, and it brings an important point to mind: You don't need to have the best in tech in order to succeed at your goals.
If their trailing edge chips are widely adopted, even only in the context of non-Western nations outside the bounds of US & European allies and near-allies, China can create their own staggered hemisphere of influence. It's another avenue down the Belt & Road initiative that largely failed, as I understand it.
Edit: There's also the danger of intermixed technologies in mil-tech, say in a nation like Turkey or India that has ties to both the US and China. That opens us up that much more to spying on our bleeding edge chip tech (and more, I'd imagine).
Chip Wars is excellent.
Hi Noah and all, for the book Wireless War the critique is mentioned that the workings of the telecom industry is not explained to the reader and because of this it is more difficult to read for the people who are not familiar with this. Could any of you suggest a book similarly written as the other book Chip Wars, but for this topic? As a lay person I would be happy to get hold of a good book on the subject. Thanks very much!
Nilay Patel has a great interview with the guy running the biggest deployment of ORAN cellular technology (the openness you referred to in your cellular book review). It's fantastic - this dude is brilliant.
https://www.theverge.com/23297756/5g-rakuten-mobile-ceo-oran-cloud-network-decoder
Well, interest in such military technologies would seem to be built upon an assumption that conflicts between the great powers will remain limited, such as we see in Ukraine. Hopefully that will remain true for as long as possible, but it's just not credible to assume that such conflicts will never spin out of control in to nuclear exchanges. Sooner or later, one way or another, some limited conflict is going to spin out of control, and then who has the best chips etc isn't going to matter much.
If you accept the premise that we can't keep nukes around forever and never use them, then there is a sense in which none of this matters today. As example, if we're headed towards a nuclear conflict between the great powers, does it really matter who wins in Ukraine?
Yes, it does matter what life is like for tens of millions of Ukrainians for the next decade or so until the hypothetical nuclear holocaust happens.
I really don't disagree, and have had a large Ukrainian flag flying on the front of our home since the first week of the war. Please see my comment as a rhetorical flourish whose purpose is to inspire deeper questioning.
The technical details described above on this page will only matter until such time as some limited conflict spins out of control. The WHEN of a nuclear holocaust is hypothetical, agreed. But on the current course, the IF is not hypothetical. We have the option to not stay on the current course.
Noah, You need to learn about "Software Defined Network (SDN)" architectures. They separate network control from data forwarding. The former is the sensitive part, the latter the expensive part (it is what gets deployed everywhere), and they may be made by separate vendors. 5G uses a SDN architecture. I don't follow this any more, but I suspect that we need to worry about the security of the former but much less about the latter. Huawei's dominance of the non-Western market gives them economies of scale in the latter, much less so in the former. I think the GCHQ (Britain's NSA) had a non-classified analysis of this.
I would add on the AI point that you are right:AI is a technology in its infancy. The big area is AUGMENTED Cognition AI, the use of AI to do tasks that humans are not cognitively good at (cognitive including motor as well). We have just began to figure out ways to augment our world (a point I think Noah made previously).