Today’s newsletter is on a topic that deserves far more attention than I’m about to give it, and so I’ll be returning it in the future. This installment is just me organizing some thoughts and asking for feedback and pointers so that I can take a real swing at it in a later update.
There’s a viral video clip going around in which Fareed Zakaria tries to debunk the idea, lately promoted by the US military, that China is any kind of serious military threat to the US or peer competitor in that realm:
To make his case, Zakaria recites a litany of stats on US vs. Chinese military spending, warship tonnage, number of nukes and fighter jets, and so on. The numbers for the US are all dramatically higher than those for China, and the overall effect is quite convincing, which is why this clip went viral.
Zakaria may have marshaled a lot of numbers, but I don’t think they support the argument he wants to make with them. Furthermore, I think his use of these stats is wrong in a very classic, gear-centric American military mindset kind of way.
The problem: his clip is an inventory of equipment and sites, along with some budget numbers; but when a battle between an incumbent and an upstart takes place in the domain of networked software, inventory lists and budget numbers are not necessarily correlated with success.
This analogy may seem a bit of a stretch, but bear with me for a moment: Imagine if way back in 2000 you had totaled up the number of computers and square footage of office space owned by both Google and Yahoo!, and then used those numbers to make a prediction about which company would out-compete the other. That would’ve seemed as silly then as it does now, yet I think Zakaria is making a similar mistake in this clip.
I say this because I recently read a very good book in a possible future US vs. China conflict: 2034: A Novel of the Next World War. In this book by Elliot Ackerman, which boasts as a co-author a retired four-star U.S. Navy Admiral, China dominates the US in a localized naval conflict (it eventually goes nuclear), and it does so with cyberattacks.
Because our ships, fighter jets, satellites, drones, bombers, and spy planes are all designed to work together and feed each other information in real-time, they’re all on the network. This makes the network a central point of failure so that an adversary who gets access to it can use it to “blind” or otherwise disable everything attached to it.
Weak links in the kill chain
I’m also currently working my way through The Kill Chain: Defending America in the Future of High-Tech Warfare, by Christian Brose. And so far, one of the lessons of this book is similar to the lesson of 2034, which is that the military’s past few decades’ push to a networked warfare model has made it vulnerable to networked disruption.
Another lesson of The Kill Chain is that in the orgy of whiz-bang tech spending, we haven’t even thought through clearly how all the parts of these systems we’re buying fit together. As with my Google vs. Yahoo! analogy, it’s not just the total cost and number of gadgets you have, but how you use them in concert with each other to achieve your goals. And in that latter respect, Brose suggests we’re in big trouble.
For instance, in the book’s intro, Brose begins by describing how our stealth fighters have very limited ranges, so in order to deploy them from an aircraft carrier that’s any distance away from a threat, you need large in-air refueling tankers. But the in-air tankers aren’t stealthy, so they make great targets. The tankers’ lack of stealthiness means that the fighters’ range is limited to a few hundred miles, which would put any platform they’re launched from within range of most adversaries’ land-based missiles.
The point: while the stealth fighters themselves may be super awesome and high-tech, the larger system they’re deployed from and maintained by essentially neutralizes their advantages. The different parts, as high-tech as they are, just don’t work together very well. It’s a sweet collection of gear, but something went wrong in the planning and assembly of that collection.
Zakaria’s thinking in the clip, then, strikes me as the same kind of thinking that seems to have contributed to our present predicament, i.e., we’ll spend a ginormous amount on to high-tech, software-driven, networked gear and facilities, and that’ll be better because more, fancier gear is better?
If our military is now a weaponized Internet of Things (IoT), then… well, we’ve all read the news stories — TVs spying on people, thermostats going offline, appliances doing bitcoin mining, etc. This seems like a big mistake.
Really, the take-home point of this post is summarized in this tweet, but add “warfare” to the list alongside voting, news, entertainment, etc.:
jonstokes.com @jonst0kes2. To the extent that a model is too big for them to govern via some kind of root access, outlaw it. Literally ban its use, because it's "harmful." To boil it down: woke fights over language are now fights over who has root in the ~5 systems that shape speech & thought online.
It’s profoundly depressing to think that over the past few decades, we’ve spent trillions of dollars on a massive inventory of IoT devices that, like all IoT devices, likely will not work when we need them most and are probably spying on us for the Chinese.
These books have so far left me with the feeling that America’s apparent military dominance is an illusion created and propped up by our unconstrained spending on inventory and facilities. I.e., it’s probably all an even bigger scam than WeWork. Maybe this is overstating, though, and as I learn more the picture will become less dire. I’d be interested to hear thoughts on this from informed readers.
I also expect that battlefield AI, which is the newest new thing, will not necessarily fix any of this. Sure, you can reduce a machine’s reliance on the network by giving it more on-board smarts, but if the machine itself is not secure then you’re still screwed.
Once I get all the way through The Kill Chain, I’ll try to score an interview with the author for this newsletter. I’m curious to learn more about proposed solutions for this problem.