Building Skynet as the Hail-Mary move for late American imperialism
'The other day, some guy told me that Neo-China arrives from the future.'
I recently read in interesting article in the Guardian (no, I’m not referencing the L0m3z doxx, lol), and it got me thinking about a conversation I had last week with
for the Tooky’s Mag podcast (this isn’t out yet).Gabe and I had a very fun conversation about a great number of topics and I wanted to take a moment to pre-emptively expand on some of the ideas I was kicking around around regarding AGI, Landian technocapital acceleration, great-power competition, and the general shittiness of Asian-American literary fiction. These might feel somewhat tangential to the writings of an Chinese-American novelist, but I think they’re not, and I’m going to circuitously explain why.
Those of you who follow me on Twitter know that I am sometimes gripped by spasms of Sino-posting in which I am seized by manic waves of ‘Neo-China arrives from the future,’ which admittedly animates some degree of latent ancestral memory from the days when my forebears were building multi-stage rockets in the 14th century and just generally doing a bunch of cool shit with technology.
As a young man, I used to receive a vicarious egoic boost from looking at pictures of American military hardware: iconic photos of F-22’s or carrier strike groups sailing in blue ocean waters on the cover of Popular Mechanics. This was largely a feature of adolescent maledom but if you survey the natsec twittersphere you’ll see that this is as much a top-down phenomenon of imperial pride as it is a bottom-up phenomenon of young men thinking that weapons system are inherently cool, which is a facet of every society (for a particularly hilarious example of this, witness The Taliban aping Call-of-Duty in the mountains of Afghanistan while wearing tactical gear that was left over by US forces—now that’s a total American cultural victory!).
Ever since I stopped believing in American imperialism, I’ve always found ethnonarcissist gloating of any stripe to incredibly gauche and annoying and if there’s one thing I want to avoid doing it’s to encroach into any kind of psychological zone that is even remotely adjacent to this kind of mental masturbation.
I don’t for a second believe in any notion of Han Supremacy and I essentially agree with the reductionistic view that raw intelligence, not ethnos qua ethnos, is the main thing that matters for climbing the technology tree. With embryo selection and genetic engineering being around the corner, the nature of biological intelligence is going to change completely (like any other arms race, you won’t be able to stop the competition dynamics of human intelligence enhancement once it gets going). The only question is whether or not human intelligence enhancement is going to matter or not—if machine superintelligence is truly around the corner, then it won’t matter at all. If it’s much farther out—then it will matter a great deal.
The thing that’s most interesting to me right now is the juxtaposition between the increasing focus that American technofeudal Lords have on building killer robots and the increasing capabilities coming out of hardware production in the Shenzhen supercluster.
This video, for example, was quite something:
I spoke to my close friend who works in AI and he told me that as far as he can tell, the PRC may well be ahead on humanoid robotics development (they are, of course, going to wildly surpass our ability to make these robots).
So: we have American oligarchs pushing the killer robot industry in order to save freedom and democracy, and we have the greatest industrial power in the history of the species ready to launch itself all the way to the top of the tech tree.
It’s an interesting moment, isn’t it?
Reading through The Guardian piece on our enthusiastic sprint toward weaponizing AI, I’ve once again been validated in my belief that we’re going to orient the entire American technology industry toward a singular goal of achieving AGI before anyone else. The real (and stated) purpose of this project will be to win a Cold War with Eurasia so that we can maintain planetary control to the maximum possible degree.
The decaying American industrial base and the intrinsic profit-maximizing nature of our prime defense contractors means that GAE's global military hegemony is on a countdown timer. The system is optimized for grift and building small numbers of complex weapons systems with supply chains that almost certainly cannot survive a serious conflict with the Chinese (even the Raytheon CEO has said they can’t cut them out, lol). PRC naval shipbuilding capacity is roughly 200x ours at this point.
Knowing this, in light of our decaying power, I predict that we will go all-in on pushing toward a militarized Artificial Superintelligence that can be instrumentalized for hegemonic purposes.
The two questions are: (1) Is such a thing (fast takeoff) possible?, and (2) if it is indeed possible, can we get to ASI first ? Elites like Eric Schmidt and Vinod Khosla are on record as essentially believing this to be the case.
This is the underlying dream you see behind defense tech companies funded by the neo-nationalist Silicon Right—that foundational models + the Military Industrial Complex (e.g. Anduril) will converge on the 'God-Weapon' and we will restore the world to 1990 in perpetuity.
What is the ‘God-Weapon,’ more precisely?
Let’s borrow a Landian term borrowed from science-fiction: it’s Skynet, an agentic artificial superintelligence (the God) that commands power in the real world through an army of killer drones (the weapon). The idea is that this would be an "instant-win" type of scenario where we effectively rule the planet once again and can impose our will by way of fiat.
The reason we need to do this is because our existing military industrial base isn’t statically behind the PRC, it’s falling more and more behind as time goes on. PRC military industrial capacity is a subroutine of PRC industrial capacity more broadly. The Shenzhen hardware suplercluster and similar industrial zones are the physical equivalent of Silicon Valley but for the world of atoms and for myriad reasons related to deindustrialization we are in a position of deep relative weakness on this front. One of the few big name tech commentators who isn’t coping on this is
, but if you watch the language, more and more people are coming around to this idea.Thus, as we continue to fall more behind, it will become increasingly clear that our defense industrial base cannot compete with the Chinese defense industrial base in terms of quantity (which is particularly important for drones) or even quality, we are going to go "all-in" on racing to ASI.
The first risk of our whole-nation effort in chasing after the 'God-Weapon' is, naturally, that it will end up killing us. However, the consensus I’m seeing is that experts like Yann Lecun are increasingly persuasive in the argument that fast-takeoff and misalignment are interwoven problems and neither is likely.
Even if an intelligence explosion is possible (i.e. if fast takeoff is possible) and we 'solve' alignment (i.e. the ASI obeys humans and does what the CIA + deep state wants it to do), there's still the risk that we won't get there first.
We've seen the pace at which semiconductor fabs like SMIC in the PRC are catching up to TSMC, and we know that their AI engineers will have enough energy + compute to 'fast-follow' US foundational model researchers.
One reason they may reach ASI (or AGI) first is if further advancements are compute constrained and compute becomes energy constrained—their nuclear power buildout and capacity for grid-building far exceeds ours (there's that pesky industrial asymmetry again).
Therefore the Silicon Right (people like Andreesen, Khosla, Thiel, Schmidt, and probably eventually Altman) will end up selling this idea of an American race to ASI being the decisive technological race—another Manhattan project, except this time, we are the ones in need of a ‘decisive victory,’ and desperately playing the role of a losing imperial Japan facing the industrial might of a larger manufacturing power.
If you want to see how we are likely going to deploy this kind of technology, you can look to our most favored client state in the Middle East—the most innovative “startup nation” of the world:
The technofeudal Lord, sitting on his prestigious panel in California, adored by the press: “American dynamism will create the God-weapon, and we will use it to dominate and defeat the entire Eurasian continent once again—the 'new axis' of China, Russia, and Iran. It’s 1939, and we’ve got to defend democracy.”
To put it simply: we're going to build Skynet as the final Hail-Mary to save American imperialism.
One small problem: it's potentially looking likely that we may have already hit the approximate upper bound of what LLM's are capable of and that we are not close to AGI in the way that it seemed like we were roughly one year ago. We may well be several significant breakthroughs away from the type of AGI that we’re positing as being possible when we speculate about intelligence explosions/ASI.
If that's the indeed the case, then it really is curtains on global American hegemony and the end of the unipolar moment. In this scenario, restoring an equal power balance between the West and the Middle Kingdom has effectively already been overdetermined by their technocapital escape velocity, which further sanctions can no longer constrain.
Soon, the PRC will hit parity across the board of high-end technology.
If you're wondering why I, a Chinese-American novelist, find any of this relevant or interesting, it's simply because great power competition is inherently interesting, and is almost certainly about to become increasingly salient to the entire Chinese-American diaspora, artists or otherwise.
Asian-American artists spend a great deal of time looking to the past but have failed almost entirely to understand how quickly we are accelerating into the future. My view is that millennial Asian-American writers (and older generations of writers) in the literary fiction sphere generally do not see themselves are being truly equal to the West, and this feeling of inferiority is downstream from imperial domination. This creates a certain lack of vitalism, or to borrow a bastardized Orientalist metaphor, a certain lack of Chi.
If and when the Anglo-American military occupation and domination of East Asia comes to an end, this will fundamentally alter the psychology of the Asian-American citizen as an imperial subject. By extension, this will alter the type of art that we produce in all forms, not merely the literary.
My underlying hypothesis is that technology, whether it be a robot or an aircraft carrier, isn’t just a machine or even an expression of geopolitical power. It’s also a psychological symbol that produces some degree of racial self-identification.
My problem with Asian-American writers in general is my feeling that they desperately need to cultivate their chi—they need to construct a persistent identity outside of Liberalism in which they have the essential capacity of seeing themselves non-passively: as an agentic force in the world. There are few works of Asian-American literary fiction that feel truly sui generis, and I see this as downstream from a broader cultural malaise related to the identity shredder I wrote about earlier.
There is no reason our cohort of writers need to be low-agency weepy losers whining about white people not wanting to fuck them (or about white people fucking them but not sufficiently loving them). Let’s move on, for God’s sake, to a completely different topic. Let’s have some fucking imagination and step outside of the assumptive automatisms that have been imprinted from last century+ of colonial (and neocolonial) domination.
Due to domestic repression of speech and culture in the PRC, it’s on the Chinese diaspora, wherever they may be, to chart a new course for literary fiction and hopefully one-day reinject that into the mainland if and when conditions around speech ease. A similar-but-not-identical mandate falls upon other Asian-Americans: the Korean-American or the Japanese-American will see themselves in a completely different way if and when the US military is extricated from occupying their homelands.
Wherever we may be, we do not have to live as culturally conquered people.
While it does feel that cultural Liberalism is inherently contingent on the success of Liberal Imperialism, it may well be that cultural Liberalism wins the day even as the balance of power between East and West shifts back to its historical norm. Regardless, I am keenly interested, as always, in the world and its manifold changes.
The Orientalist futurism of cyberpunk Japan may well have proven to directionally correct, but it seems that history has chosen to grant us the Sinicized version instead.
I absolutely agree that, whether it will be due to an AGI race or not, the American Empire is falling behind increasingly quickly, and I would bet that that's been apparent to backroom power for quite a long time. I remember writing about falling American capital asset prices and the potential for their acquisition by China in something like the early 2000s, and I don't think I was clairvoyant. The writing was on the wall. And of course, whenever you're number one you spend a lot of time being paranoid about your impending fall...
I think of the past 20 years and likely the next 10 or so (depending on the speed of the fall) as the French idiom "sauver les meubles". Saving the furniture, generally in the context of your house burning down. Every global actor whose power is founded on the American status quo will be under increasing pressure to get as much out of the arrangement as they can before it doesn't exist, which I think is the biggest reason that Israel is engaging in general omnidirectional violence. If you know you're going to slowly drown otherwise, now is the time to do everything you can, as aggressively as you can; the more you wait the less strength you'll have. Israel is just the lowest-lying Imperial island.
China wins in every scenario aside from some kind of divine inspiration striking an American genius I suppose ... since most technologies are actually built slowly and incrementally on decades of work, this seems almost impossible to me. 新中国已经到了...