This weekly roundup thread is intended for all culture war posts. 'Culture war' is vaguely defined, but it basically means controversial issues that fall along set tribal lines. Arguments over culture war issues generate a lot of heat and little light, and few deeply entrenched people ever change their minds. This thread is for voicing opinions and analyzing the state of the discussion while trying to optimize for light over heat.
Optimistically, we think that engaging with people you disagree with is worth your time, and so is being nice! Pessimistically, there are many dynamics that can lead discussions on Culture War topics to become unproductive. There's a human tendency to divide along tribal lines, praising your ingroup and vilifying your outgroup - and if you think you find it easy to criticize your ingroup, then it may be that your outgroup is not who you think it is. Extremists with opposing positions can feed off each other, highlighting each other's worst points to justify their own angry rhetoric, which becomes in turn a new example of bad behavior for the other side to highlight.
We would like to avoid these negative dynamics. Accordingly, we ask that you do not use this thread for waging the Culture War. Examples of waging the Culture War:
-
Shaming.
-
Attempting to 'build consensus' or enforce ideological conformity.
-
Making sweeping generalizations to vilify a group you dislike.
-
Recruiting for a cause.
-
Posting links that could be summarized as 'Boo outgroup!' Basically, if your content is 'Can you believe what Those People did this week?' then you should either refrain from posting, or do some very patient work to contextualize and/or steel-man the relevant viewpoint.
In general, you should argue to understand, not to win. This thread is not territory to be claimed by one group or another; indeed, the aim is to have many different viewpoints represented here. Thus, we also ask that you follow some guidelines:
-
Speak plainly. Avoid sarcasm and mockery. When disagreeing with someone, state your objections explicitly.
-
Be as precise and charitable as you can. Don't paraphrase unflatteringly.
-
Don't imply that someone said something they did not say, even if you think it follows from what they said.
-
Write like everyone is reading and you want them to be included in the discussion.
On an ad hoc basis, the mods will try to compile a list of the best posts/comments from the previous week, posted in Quality Contribution threads and archived at /r/TheThread. You may nominate a comment for this list by clicking on 'report' at the bottom of the post and typing 'Actually a quality contribution' as the report reason.
Jump in the discussion.
No email address required.
Notes -
This is just prolonged "nuh-uh". CUDA and CANN will continue to evolve divergently, Chinese models will likely be built around extreme sparsity and multi-tier memory (like ByteDance's UltraMem), scheduled Huawei systems are increasingly different from Nvidia's lineup (vaguely like Google's pods but Google doesn't yet sell those). We see that American providers took almost a year to implement DeepSeek and it's been just slightly unconventional, actually trained on H800s; SiliconFlow rolled it out on CloudMatrix 384s weeks after they were delivered. Lock-in happens on the software and hyperscaler level.
That said, all of this is beside the point because your idea is "hopefully we'll leave them so far behind their market won't matter much".
Industrial policy is about advancing domestic capability, protectionism and targeted subsidies – CHIPS act, banning Huawei in the US, not Nvidia in China. You're just calling any policy that has something to do with industry "industrial policy" I guess, but your argument is purely geostrategic and has no direct bearing on industry development in the US.
Incredible propaganda, yes. DeepSeek has maybe 300 people. And they weren't forced to do shit. If you mean that story, I've checked with the reporter and the report is basically unsourced rumor. This isn't happening, you're working purely from assumptions. There is now, indeed, effort to subsidize the adoption of domestic compute, but that's inevitable when Americans are deliberating on whether they can afford to sell even obsolete inference capacity.
You mean, it's a self-congratulatory, narcissistic Zeihanite myth? You're not protecting shit. You've just lost to Houthis, the first nontrivial challenge to sea lanes in forever. Your navy is designed around offensive operations against nation states and deterrence in nuclear war, not patrolling sea lanes, and its crown jewels are aircraft carriers and submarines. Maybe it would do great to block sea trade, at least that's the plan for Malacca. The global trade will certainly go on fine if it's scrapped.
Trump generally campaigns on real if exaggerated problems and popular frustrations, it's just his solutions are often hare-brained. China would definitely like to increase value-add, of course, but it wouldn't mean canceling "industrial policy" and shedding dominance in stuff like photovoltaics, they'll simply make factories more automated. Interestingly, in this case I even agree with Trump on selling GPUs to China, for once his mercantile instincts are appropriate.
Demand as such has zero value because it is easily produced at infinite scale and, for the purpose of this conversation, it's a malign concept. People don't sell to the US because the US pays back with some demandium, they just trade their work for a piece of liquid and appreciating American assets (insofar as those descriptions apply). Also, we've seen that as trade with the US fell due to tariffs, Chinese exports to ASEAN increased (and no it's not transshipping, the composition of goods is totally different) and fully canceled the drop in trade surplus. The world can produce plenty of "demand". You aren't that big anymore.
I think it's pretty stable (dysfunctional shitholes like Russia and even Iran stand strong), AI is likely to make it more so. Cybernetic superweapons are unlikely because hardening systems when you have unlimited time and root access is easier than pen testing; we'll get to verified kernels for everything much sooner than AIs become expert hackers. Material science and engineering advances promoting lasers, drones etc. are great for defense. Panopticon angle is obvious enough. I strongly doubt AI will enable some sort of super-nukes. This of course is a matter of opinion.
Selective quotation is a hell of a drug. On a single chip basis, even Huawei admits they can't compete and won't be able to in the foreseeable future (EUV breakthroughs may change that). They can make do with better systems integration and produce competitive (also due to more electric power, better grid) systems and that'd suffice to serve domestic demand, for lack of better alternative.. For the end product (AI), they'll be slowed down relative to the world of uncontested Nvidia dominance. I posit this is not critical. The critical thing is that this market will keep growing exponentially, and before too long you're forfeiting not tens but hundreds of billions, on not selling one of your few truly unparalleled products. Is the idea to make up for that with Singularity stuffa nd extorting allies in the meantime? This is a Hail Mary.
I guess this is the crux. In your world, where unipolarity is the default trajectory, it makes perfect sense to cling to Pax Americana and play negative-sum games hoping to outlast the opposition. Like, what is the alternative, capitulation, suicide? In my world, China is basically guaranteed to not only exist in 30 years but have comprehensively stronger economy than the US plus closest allies, no matter what you sell or don't sell, buy or don't buy. And the US will have to figure out how to exist, and exist well, without boons of global strategic superiority, in a bipolar world, and hopefully remaining a hegemon in its own backyard. That figuring out has got to begin now.
It's getting a little late on a Sunday to go point by point again but if memory serves correct, I briefly looked for the post can couldn't find it, you've acknowledged China's demographic issues and your solution to it was essentially ai enabled robotics to handle elderly care in order to keep the ratio of dependents to workers manageable correct? The happy case for China multipolar strength seems to rely on a pretty narrow outcome where AI is powerful enough to do a huge amount of menial labor but not powerful enough to where being ahead a couple years differentiates world power standings. And I just don't really see that as very likely. If AI doesn't get all that much better than it is now then I foresee china having another decade, maybe fifteen years of exceptional output before a lot of trouble handling a rapidly shrinking population. If AI does take off in a big way I think the couple years head start on chips will be pivotal and giving that up would be foolish. I can't really think of any reason to expect we'd fall between those two points.
Americans have invented themselves a lot of cope about China, to the extent they're not paying attention to similar domestic issues.
Right now Mainland Chinese are younger by over 4 years than White Americans (40.1 years in 2025 vs 44.5 in 2020 median). By 2040, they'll become about on par I think (46-47). I don't trust fertility projections after that, we've seen nations rapidly fall into the East Asian model, including throughout South America (Central America for now is holding up admittedly). American solution to American demographic issues is importing assorted Hispanics to take care of white boomers and hoping that Hispanics will somehow also become a replacement for the working population. I don't believe this is happening as far as O-ring economy goes (software, finance, deep tech jobs now ride on adding Indians and Chinese to local numbers; Indians don't have that deep a well of talent and they're having a demographic transition too, especially for higher castes; Chinese net flow is reversing), so in my view your productive working populations are shrinking at a similar pace. After seeing stories of elderly abuse by immigrants in the US, I am positive that robotic caretakers will at least reach parity soon for augmenting blue collar work and and caretaking. Americans are trying to add robots to industrial workforce and will likely begin to automate retirement facilities too, they're not that dumb and there are Western robotic projects clearly aimed at home labor. Robotic mobility is now at this level, and this is a blind policy. Progress in manipulation is similar, China reigns supreme in actuators market and casually make dexterous hands now, it's a very nice fit for their industrial model so they'll only increase their lead there. Robots will suffice for menial labor, both in China and in the US (probably marginally more so in China but it's not a crux). Finally I don't believe in the necessity of unproductive population to "provide consumption", rich people with robot slaves can consume as much and grow GDP as much as multiple poor people. On the whole, I am of the mind that demographic trend difference is a dumb and, again, Zeihanite red herring that ignores medium-term predictable AI progress.
So abstractly, it's not a narrow outcome, because as I've just said, the floor is basically established, and China won't have to make up for a large extra deficiency. The whole question is about those huge gains of productivity on the right tail, it's the US that will have to make up for having fewer and lower IQ people, NIMBYism, alienating allies, degraded supply chains and retarded and worsening political culture with geniuses in a datacenter, by gaining a couple years of edge in AI progress and not fumbling the application of gains (and starting at 7:44 here, Molson gives me a reason to suspect that China will also be better at applying what gains they make with AI throughout the period).
Now, I believe AI is going to be really useful. A review commissioned by DeepMind predicts that at a minimum, 2030 level AI will boost productivity for desk-based research by 10-20%.. That's a lot. All things considered, is this enough to "compound" your way to lasting hegemony after 2030? I wouldn't bet on it, but Americans are Winners by nature, so they might. The upside of hegemony is, in theory, near-infinite. The downside is just having a worse place in the eventual bipolar world. Whether to take this bet depends on the odds (and nuances of the value function). I'd say the US has maybe 25% chance of "winning the race" to hegemonic condition.
More options
Context Copy link
More options
Context Copy link
More options
Context Copy link