DaseindustriesLtd
late version of a small language model
Tell me about it.
User ID: 745
DaseindustriesLtd
late version of a small language model
3d ago
I seriously don't see how that addresses anything @SecureSignals has said. Like, yes, Bari Weiss is more than just «a Jew», she's a specific person, with individual connections, traits, skills and credentials that have differentiated her in the pool of admissible candidates. Certainly she couldn't be substituted with someone like Norman Finkelstein on the mere account of his ethnicity. The issue here isn't even Bari per se, it's the criteria defining the nature of the pool. It's the same issue as the negatively-defined criteria of DEI preferential treatment, which people are much more comfortable rejecting.
DaseindustriesLtd
late version of a small language model
3d ago
The problem is that this whole «human capital» philosophy that treats humans as interchangeable stuffing of different grades in the American pie is premised on some extremely degenerate assumptions about human nature at this point.
The most plausible explanation for the vast amount of cultural decline in our country over the past 15 years is NOT some secret cabal of Jews conspiring to destroy the countries that they share with us. It seems much more likely that outside actors who actually have a vested interest in watching the US and Europe collapse are to blame: the Qatari, Saudi, Emirati oil money are buying their way into influencing Western academic, political, and social capital in a way that undermines Western values and promotes Islamic ones. Likewise, the Chinese Communists are no doubt using all of the psychological warfare tools at their disposal to accelerate the collapse of the American Empire. The Chinese definitely want to see America be as multicultural as possible and promote identity politics to create more divisions within us.
Islam is at war with the West, and they have been for thousands of years. The difference is now the West is losing this war, and we are losing badly. If Western Civilization can muster the courage to actually declare war against Islam, as they have declared war on us, the Jews will be overrepresented in the political, military and cultural institutions that are fighting for western civilization. The Jews helped us beat Hitler. The Jews helped us beat the Soviet Union. The Jews can help us beat China. The Jews can help us beat Islam, too.
I notice you don't mention China except as something to beat, while extolling «Japanese», «Taiwanese» and «Asians». It's peculiar because of course Taiwanese are Chinese, most American [East] Asians are Chinese, and indeed, they're doing very well! Low crime, high SAT, sizable tax contribution, and as I've said in my last long post, when an American Jew (Zuckerberg) wants to build an American Superintelligence Lab, 20 out of 30 research scientists turn out to be Han Chinese. In fact, 18 of them even hold PRC citizenship. This is about how it looked at the absolute peak of Jewish dominance in the American cognitive elite.
And yet, and yet – the US is having a decade-long meltdown about strategic competition with China. Even these researchers are suspected as potential spies who'll leak our precious inventions (their own work, largely) to the Red Dragon. There's a lot of vitriol directed at the Chinese, smoothed over with unconvincing noises to the effect of «no no I don't hate Han people, love my Hapa children, much beautiful ancient culture, wow very friendly very nice, I hate the CCP [also nuke three gorges dam]». Well, but the CCP is made of 100 million Chinese people, it has a sky-high approval rating (no it's not fake), and it genuinely represents their collective will to be a successful race, a superpower with hegemonic potential, rather than an assembly floor and source of high-skilled labor for Americans (including Jews).
White people like (presumably) you, people who buy into this «human capital» doctrine, are simply people. Chinese and Jewish people are a people, and in their own cultural frame even the People – a distinction which is a bit better articulated than in many other cultures, but in no way an abnormal way of thinking. They are ethnocentric. Goys and barbarians are not part of the people, and the people will coordinate to achieve collective gain in zero-sum games with barbarians and goys. That's table stakes for a self-aware successful culture.
What kind of war against Islam are you envisioning? It's pretty funny because militarily, Islam is not a threat to the West at all and has no potential to become a threat. It is, of course, a moderate but real threat to Israel, which is why pro-Israeli actors will hype up the Islamic threat to try and have you fight their wars. The Saudi money works, if it does, because your ruling class is hilariously corrupt and disinterested in the long-term prosperity of the populace. The main danger scenario is illustrated by the case of the UK, with slow population replacement by a mix of different immigrant groups and the low-status people (low human capital, so much less interesting for «the UK» than Jews or Asians) converting to Islam out of desperation. To stop this, you don't really need high-IQ Jewish generals and nuclear scientists, you need to learn to think of yourselves as «a people» that have intrinsic worth not denominated in tax returns or exam scores. But that's at odds with those very «values» you hope to have the Jews protect for you.
All of this is very mush-headed. There's no need to antagonize any ethnic group or reject cooperation, but there is a necessity to acknowledge that major nations represent essentially ethnic interests that are partially shared by their diasporas, and there is not a single non-Western nation that is straight up invested in propping up the West for «values» or whatever. Including Israel. All alliances will be alliances of convenience.
DaseindustriesLtd
late version of a small language model
4d ago
I think that peaceful reunification is the base scenario at this point and Americans are a bit high on their own supply. It's not that there's more enthusiasm for it, but there is definitely less visceral rejection. The fundamental case against it has been not «freedoms and democracy» (Taiwanese adopted LGBT stuff largely to please their patrons) but the belief that the Mainland is a big, embarrassingly poor North Korea where you worship Mao, eat gutter oil, work like a dog at Foxconn and die.
It's still popular with the older Taiwanese, but facts change and speak for themselves. As a young graduate you can earn more in tier 1 city like Guangdong than in Taiwan, and Guangdong is plainly newer and bigger and cooler. The Mainland is increasingly seen as «awesome» by local influencers, DPP is unpopular (eg for shutting down their nuclear power on a Germany-tier green platform, which ironically makes reunification-via-blockade a lot easier, they'll run out of coal and gas in 2 weeks and their civilian society, nevermind those fabs, stops dead) and getting censorious in apparent desperation, KMT is likely to win this time, Ukraine as of 2025 serves as a warning rather than inspiration. The military buildup on China, including specialized assets like these zany barges that defeat the «few landing-worthy beaches» objection, seems very serious and increasingly impossible to deter. Basically, if you're against China, you can't rely on any shithole level deterrence like geography, they can simply engineer and build their way over it. You need to rely on hard military capability.
And that's the problem, because no matter how porcupine Taiwan gets, the real muscle has to come from the US. And they believe less and less that it will come. Lutnick-style opportunism is widely seen as dismantling their Silicon Shield, and I think they're right – the US that can make chips at home doesn't have an existential stake in Taiwan. China cares about the First Island Chain and about finishing the civil war. The US stopped caring about that back in 1979-1980, and using Taiwan as an opportunity to contain China is only worthwhile if that's the relatively cheap option. It doesn't look cheap. Only chips, then – and once chips are made in Arizona, not even that. There's broader logic about «our allies in the Indopacific» but at the end of the day that's hubris and imperial overextension, all of these arguments are downstream of the ambition to contain China and Win History, and as Trump's National Security Strategy demonstrates, ambitions can be downscaled in response to new circumstances. The US can keep Guam and Okinawa in a world where Taiwan has fallen, and will try to.
So I think that by default, China takes Taiwan within 5-20 years, either by a face-saving «1 Country 2 Systems» arrangement, or with a brief blockade followed by polite demonstration of overwhelming power. I believe China (Xi) has a similar theory and so won't rush into a hot conflict, which serves everyone for the moment just fine, even if me and Xi are in fact wrong.
P.S. People who argue about blockading China are not very familiar with the facts. They aren't dependent on imported food, these soybeans are for pigs. They don't actually biologically need to eat that many pork bellies. They have vast stockpiles too. They're electrifying very rapidly, from cars to trucks to ships now, and in a few short years their core logistics and power generation will be able to maintain wartime economy without any maritime fossil fuel supplies. Commodities like iron ore are harder but that's not even a blockade issue, the US can compel Australia/Brazil/Chile to stop exports. Even then, it's not going to be decisive. The Chinese can just do things, it's actually mesmerizing to see.
As promised, I'm responding in more detail. Much of the below I could nitpick at myself, but I'll leave you the pleasure. My confidence in being directionally right is underpinned by having been optimistic about Chinese trajectory for years of seeing «Xi bad» and «China collapse» narrative. I was even right in expecting their failures – such as no EUV breakthrough or quick solutions to fertility and real estate problems. I was, however, wrong in predicting a stronger American showing in response to their success.
Disclaimer: the load-bearing phrase in your spiel is «cold bug person». Your perception of China is downstream of finding Han Chinese people, politely, uncharismatic, and bluntly – not hot, in terms of a vibe rather than mere looks. You are inclined to believe that «not hot» people are categorically lesser than hot ones, incapable of True Creativity/Courage/Honesty/Ambition/Valor/Decisiveness etc., and unconsciously reason backwards from that about object-level evidence. This is compounded by the historical and contemporary distortion that the Chinese themselves have done to their image, like painting a 1.9m tall cannibal commander from that famous "decisive Tang strategic victory" meme as a pompous literati with long nails, and today taking advantage of leftist grievance narratives. A PLA soldier can point a Dongfeng DF-17 missile in your face, and you'll laugh about how it's filled with water and compensating for something, Bugs Bunny style. It's similar to the blindspot the Kzinti had for human females. More controversially, this is a common bias that I think is genetically hardwired in peoples of Northwestern European extraction/WEIRD cluster, due to greater historical female mate choice and thus disproportionate returns to thinking about mate value and general charisma signaling on evolutionary timescales. This being basically a misgeneralized instinct, arguing about it is a waste of time. I will argue about downstream confusions, though.
As a Russian, I don't really mind – actually, I'll welcome it – if China becomes the preeminent power of this century (which seems more likely by the day) and you guys get knocked down a peg with your flabbergasting ill-earned provincial chutzpah. Given their traditional isolationism, I estimate the damage to be mostly confined to morale. I am not cruel, however, so these cowboy jeers are worrying me, because their popularity, coupled with populist incentives and low cultural level of American officials, implies you may wade into a serious war in the South China Sea and get mauled, with massive economic devastation and losses for you, them, and everyone else. You'll directly drag other nations into it, too, it has already happened on trade, and the escalation potential is pretty much uncapped. So we can't be too cautious and I will present my attempt at dismantling this theory of Chinese doom and American exceptionalism.
What I want to accomplish, however, is not just dunking on this particular laundry list of assertions. Ideally, I want you is see how your very frame of thinking, embedded in (the quite recently established version of) your civilization, is just one of possible frames. Your notions of how to evaluate success, what is hard and what is easy, what makes nations strong or weak – those are just opinions of a 21st century American (or mental American). China doesn't think in this way, it has a compelling claim to the priority of the very different Chinese stack, and it's not some vague Orientalist wisdom but a comprehensible, pragmatic product of millennia of social evolution.
One aspect of it is the theory of Mandate of Heaven, which amounts to a claim that the source of legitimacy is neither birthright nor opinion of some constituency, but undeniable object level performance, and that the people ought to obey performant rulers but topple those who had clearly lost their touch. It has its shortcomings, of course, but I am not sure if the justifications for representative democracy with universal suffrage are stronger than the case for Mandate. Another is focus on cultivating everything under the state's control, starting with human capital, which is the only truly irreplaceable resource. Mencius, 4th century BC: «The people are the most important element in a nation; the spirits of the land and grain come next; the sovereign counts for the least.» The clearest example of this is how they played the Rare Earths card. In 1992, during his Inner Mongolia tour, Deng Xiaoping had proclaimed «The Middle East has oil. China has rare earths». This wasn't because he had stumbled on the biggest pile of REEs known to man, like Americans in Bumfuck Indiana often boast of; it was because dominating REEs required long-term grit and investment that he could expect other nations to prove deficient in. In 1995, Deng's sons-in-law bought out Magnequench from General Motors. Over the next 30 years, China has indeed invested in relevant education, capex and R&D, evolved from mucking around in toxic sludge to 6N HREE refinement, consistently forced ≈everyone else out of business, and now can make Trump play ball with a single export controls announcement, even as it's building actuators for legions of robots of all kinds (flying, wheeled, legged, seaborne, industrial…) using a late evolution of this very Magnequench technology and others. All of that has been done more or less in plain sight, documented in official 5 Year Plans of the Communist Party of China and programs like MiC 2025, inciting decades of WTO-mediated outrage, «maybe we can substitute it» procrastination and «wake-up call» rhetoric on the West, and yet nothing has been done successfully to counteract it. That's one measure of each system.
On demographics, the problem is overstated. To return to the meta level, people often assume that if the Face-Saving Paper Dragon China admits a problem that's because the situation is too catastrophic to deny. No, they simply communicate clearly to coordinate their own policies. Yes births are plummeting, population contracting, median age and dependency ratio climbing, there are miserable men etc etc. It's rough. The same and far worse is happening in Korea, which despite some efforts hadn't had a comparably strict One-Child Policy (ironically enough, TFR had never reached 1.0 during 1CP, and only reached it once Xi abolished all limits and turned pro-natalist) and has no plausible policy response either, same as everyone else sans Israel. China is de facto running a massive research program, in parallel across provinces, counties and towns, with some promising results, so if anyone figures it out, I reckon it'll likely be them. But any success will take 20+ years to manifest so it's probably irrelevant in the grand scheme of things. And in the grand scheme of things they're retiring peasants and construction workers who've seen Cultural Revolution, and graduating more STEM cadres than the rest of the world combined, so even with the unfortunate pyramid shape the total productivity is expected to grow. The Party itself is becoming a hive of Tsinghua STEM Ph.Ds who (just a guess) might become even better technocrats than the current set of boomers in control, nevermind boomer lawyers or the US.
More to the point, the median age of a White American is ≈44 years vs 40 for a Han Chinese in the PRC, White American TFR too is well below replacement, and the population contraction has been offset via immigration, largely of people who aren't doing that well on MSAT and often don't speak English well. Around the time when average ages are projected to converge (2040s, age 49-51), non-Hispanic Whites will have become a minority at home, and equal to roughly 1/7th of the PRC's population. This community knows well what Trump thinks about such a strategy, and many agree. (Btw, contra racist stereotypes, South American fertility is in free fall and I expect Mexico to join soon, so this gig isn't going to work forever anyway). Given such factors as a) poverty level pensions for rural citizens of the PRC (citizens who were largely excluded from the 1CP, had TFR ≈2.5 in the 80s, and so actually have > 1 children today, unlike urbanites and party members – there's a fascinating aspect of an accidental inter-class contract), b) retirement age only ≈60 in the PRC, and c) 17% youth unemployment at 5% economic growth… I'd say they have plenty of gas left in the tank, are more demographically robust than the West and even the US specifically, and Zeihan-level doomposting is simply innumerate.
One last note on this. India also tried to implement population control, psyoped by the same Limits To Growth style western concern trolling, but India doesn't have Chinese state capacity, so after an atrocious sterilization campaign they gave up and now they're the most populous nation. They're also probably the only nation in modernity that has seen height decline without ethnic mix change, because they're unable to feed themselves and now their children are fucking stunted, exactly as had been predicted by the big bad Ehrlich; and now young Han Chinese men are like 10 cm taller than young Indian men, which no doubt adds to the cross-border seethe. On the economic growth side, both nations having started from subsistence agrarianism and Subsaharan poverty level in the 70s (with sizable Indian per capita lead), enough has been said. I am still ideologically quite disgusted by the 1CP, but like most things Chinese, it deserves deeper consideration.
Bullet point 2 is pure assertion. High tech surveillance is not very expensive when you make all the tech for it and are surveilling unarmed 40+ year old East Asians. By all accounts the primary goal is genuinely to improve public conduct and it had already been largely achieved, the people are more polite, they drive lawfully, there's less scam, corruption and squalor, and on the ground level China is looking more «Japanese» than ever since the Southern Song Dynasty probably. It's a pity that this is now mostly seen by Russians because Western tourists don't want to come after Covid. «Social credit score» as commonly imagined is a mythologization of ad hoc regional programs to do stuff like penalize public transport misbehavior (which you sorely need) and, well, actual credit scores you already have. And they do have a social safety net which is evidenced eg by virtual absence of homelessness, it's just implemented via in-kind transfers (6-7% GDP) such as community canteens, rather than direct cash redistribution, which makes sense given that they're a materially productive socialist society. It could be more robust but, again, the issue is blown out of proportion: «Social expenditures in China have more than doubled as a share of GDP since 2010 and are on par with Mexico and Turkey.»
On technology, I have pretty strong opinions. The catch-up growth with IP «theft» (overwhelmingly, not theft but joint ventures, M&A and other above-board mechanics that foreign entities myopically signed on to) is tale as old as time, from Britain-Germany to USA-Japan, only made special by their sheer scale. At this point they have everything, invention AND innovation AND tinkering. They fucking license anti-cancer drugs to you, Intel is testing their wet etch equipment, they run circles around you in hypersonic warfare and EMALS and radars and power electronics… but certainly they are doing better in the «lots of compounding, happy accidents shared across a large population» department. That's the whole story of Shenzhen – a giant singularity of small shops with narrow expertise, doing swarm process innovation. That's what they build dozens of technoparks and industrial zones for. That's why they rapidly churn through zany concepts like here– semi-autonomous truck caravans, the exact sort of product-oriented tinkering innovation Americans were doing before they had assembled their own centers of basic research (mostly from European human capital fleeing Apocalypse). Their ever-growing trade surplus, driven by expansion of exports and contraction of non-commodity imports, is not about «weak RMB» or «subsidies» (you also do subsidies, you're just bad at subsidizing structural growth factors) – it's about trivial, brutal and unceasing productivity increases. Per unit of labor, a Chinese worker produces 2-3 times as much equal quality physical output as an American one:
Electric vehicles provide a clear comparison, as Tesla’s Shanghai and California Gigafactories produce identical Model 3 and Model Y vehicles. In 2024, Shanghai produced one million vehicles with 20,000 workers, while California produced 464,000 with 22,000 workers. Even in nominal value-added terms, Tesla’s Shanghai workers were twice as productive as their US counterparts. This despite the fact that Model 3 prices are 24-32% lower in China due to competition from numerous other EV makers, whereas Chinese EVs are practically barred from the US market.
That's not because they're superhumans who work 996, that's decades of learning and coming up with ideas, on a level that Americans have forgotten how to think about, and now cannot even conceptualize as a dimension of innovation. You have like one guy who's thinking about this and screams DESIGN IS OVERRATED MANUFACTURING IS UNDERRATED, and half your country wants to cancel him for being an asshole. Well, Elon Musk is venerated in China, because he's the apex of what every Chinese industrialist wants to be, and there are thousands of these guys. Accordingly every Chinese EV company is now an AI company and a humanoid robotics company, and other companies try to catch up and differentiate, and XPeng IRON 2 is in my humble opinion a more impressive piece of hardware than Optimus 3 (to the extent that people were suspicious it's amputee in a suit), and UBTech will be the first company to ship thousands of factory-grade units with Walker S2 while FigureAI CEO Brett Adcock is hyping on social media and trying to deboonk their video as CGI, and so on and so forth, in every single industry. What MENA now.
On domestic market side and weak demand, this is again a problem acknowledged as a big one by the CPC (Xi's entire «dual circulation» agenda) and thus overrated as a catastrophe. They are still heavily dependent on exports but that isn't an existential problem, seeing as they've cultivated an enormous and growing market in the Global South/ASEAN, for high-margin capital goods (that these industrializing nations can't afford to tariff heavily) rather than Walmart trinkets, while the US and EU have stagnant demand for material exports, tariffs or not. They're eating up German capital goods market share even in Germany, anyway. Historically, raw superiority in productivity wins against contingent trickery, it's what defeated China the last time, shattering their self-conception as the most productive civilization (indeed the only real civilization, because for them the civilization is largely about productivity), and they have a very clear «never again» position here. As with other points, I don't seek to deny real problems, but the doom narrative gets somewhat embarrassing when put in proper context. This is what is supposed to collapse them? They're «merely» consuming about the same volume of goods as the US, while producing more electricity than the next three biggest entities combined? They're a much bigger economic power than the Soviet Union ever was, and you had a decades-long meltdown about the genuinely poor and dysfunctional Soviets, but China is a nothingburger? Sure I'll grant that us Ruskies are more charismatic. And more handsome. Gagarin, Korolev, Zhukov, Kasparov, Baryshnikov… But please, guys, that's getting weird. Be real.
On «cultural aptitude for tinkering, exploring, developing in a truly creative sense», maybe you need to install RedNote and see what they're actually doing, because they absolutely tinker like mad, as they have been tinkering for centuries, it's the same energy as young Palmer Luckey but on an incomparable scale. They are natural shape rotators and tinkerers, what they lacked was precisely systematic scientific inquiry (and risk-tolerant patient capital). Do you realize that this is the people who've developed treadmill-driven paddleships in the 5th century and mechanical landmines in 15th? That «four great inventions» is just a PR term and they had invented vastly more stuff along the way? When Meta is desperately trying to build a Superintelligence Lab at any cost, paying $100M sign-on bonuses to highest-alpha talent, this is what the list ends up looking like, 20 out of 30 «research scientists» are Han Chinese (18 of them holding PRC citizenship). VAUK is a furry weirdo who tinkered with wolf-head exosuits and now works with the PLA to build actual exoskeletons for «wolf warriors». Here's a more Reddit-coded example of their tinkering (considered lame in the Mainland because it's «not hardcore»). But more to the point, I'll just repeat what I've already quoted:
Impression 1: In our sophomore year, while we were obediently attending classes, doing homework, and preparing for exams, Liang Wenfeng was already self-studying digital circuits and analog circuits and had begun his own engineering practice. What left a deep impression was that he personally handled everything from circuit design, PCB layout, microcontroller programming, to UI design, creating something like a miniplayer software (doing Software UI in 2004 was a high-skill endeavor). He modified an ordinary guitar into an electric guitar, where the guitar's string sounds could be controlled via a UI on the computer. This project seemed incredibly impressive at the time; we all looked at it in awe. He humbly said the guitar's tuning wasn't great and it would be better if it could tune itself automatically. This can be considered a testament to the seed of his ideas about AI intelligence back then.
Liang Wenfeng proceeded to tinker his way to National First Prize in the Electronic Design Competition, then to tinker full stack marine navigation systems, then he tinkered away 3 years in isolation as he developed quantitative trading strategies, then he tinkered into building one of the 4 biggest quant funds in China, then he tinkered and somehow now we have DeepSeek in the open source, from their in-house file system to weights and methods published in Nature (he's on top 10 this year). All that said, Liang agrees with you that the default Chinese mindset is suboptimal:
"DarkWaves": Do you think curiosity-driven madness can last forever?
Liang Wenfeng: Not everyone can be mad for the rest of their lives, but most people, in their youth, can devote fully into something, with no utilitarian concerns at all.
"DarkWaves": Why is it so easy for Chinese companies -- including big companies that don't lack money -- to prioritize rapid commercialization?
Liang Wenfeng: Over the past 30 years, we have emphasized making money and neglected innovation. Innovation is not entirely business-driven, but also requires curiosity and creativity. We're just bound by the inertia of the past, but it's just a phase. […] "DarkWaves": But it's hard to form an absolute advantage in a large model, simply by being ahead in technology, so what's the bigger thing you're betting on?
Liang Wenfeng: What we see is that Chinese AI can't stay a follower forever. We often say that there is a gap of one or two years between Chinese AI and the US, but the real gap is the difference between originality and imitation. If this doesn't change, China will always be a follower, so there's no escaping of doing exploration. […]
“DarkWaves": Back to the topic about original style innovation. Now that the economy is trending down, and capital is entering the cold phase of the cycle, will it put more of a damper on original innovation?
Liang Wenfeng: I don't think so. The restructuring of China's industry will rely more on hard-core technology innovation. When many people realize that the fast money they made in the past probably came from the luck of the draw, they will be more willing to bend over backwards to do real innovation. "DarkWaves": So you're optimistic about this too?
Liang Wenfeng: I grew up in a fifth-tier city in Guangdong in the 1980s. My father was an elementary school teacher, and in the 90s, there were a lot of opportunities to make money in Guangdong, and many parents came to my house at that time, basically because they thought education was useless. But when I go back to look at it now, the ideas have all changed. Because money is not easy to make anymore, even the chance to drive a cab may be gone. It has changed in one generation.
There will be more and more hardcore innovation in the future. It may not be yet easily understood now, because the whole society still needs to be educated by the facts. After this society lets the hardcore innovators make a name for themselves, the groupthink will change. All we still need are some facts and a process.
Their slogan is «Unravel the mystery of AGI with curiosity. Answer the essential question with long-termism» and they're the highest-prestige lab in China, with 5 IOI Gold winners on a team of ≈200 and a crapload of comparable talent. Liang is a hereditary teacher and I think he'll succeed in teaching China that his way is the right way, as I've previously argued. He's already impressed Xi and Li Qiang a year ago, he's just matched GPT-5 on a shoestring budget after a year of pundits saying that 100x compute advantage will erode his competitiveness, I think that's a big part of the reason Trump has abandoned Biden era containment and greenlit H200 exports. That's a very compelling set of facts, much more compelling than Jack Ma's (btw Jack Ma was at the same meeting with Xi and he's back in control, only now Alibaba is building AGI rather than shady fintech; that said, Ant Group also has an AGI division doing DeepSeek-style MoEs) and… it's pretty clear which way the process is going.
Can the US change its process after seeing new facts? Or do you lack the self-awareness to notice its defects?
On self-awareness. I think one of the greatest Chinese strengths – in addition to their sheer intelligence, endurance, their concept of the performance-based Mandate and the legitimacy (in fact, moral imperative) of rising against a degraded power, their notion of cultivation of the realm on all scales – is simply that they're an old, continuous, cultured civilization. They're jaded. They have seen it all, usually centuries before us; before you. Military overextension, proto-capitalism, proto-fascism, proto-Nietzscheanism, currency debasement, religious zealotry, multiethnic assimilation politics, infrastructure boondoogles, whatever. They can draw on an immense wealth of examples, positive and negative, and unlike modern Americans it's not restricted to the 95th percentile of cultural elites. Along the way they have developed a language to talk practically of how societies on the scale of modern Western powers fail and collapse, what moves people, how to make them act better, what is the «superior person» and how he differs from a «thief of virtue»… They have all these terms like «mianzi» and «guanxi», and Westerners look at it with condescension, Orientalist excitement or cold zoological curiosity, not realizing that these are human univerals, that Westerners themselves are driven by fairly similar if not cruder mechanisms. From what I can tell, the US today is a lot like late Qing, and its people are trying to save face before a more productive and well-ran civilization that they have grown used to regard as a barbarian shithole (and also, as I've said – «not hot»). And there's a whole lot of Guanxi going on in this Administration, too. Trump is trying to do a Self-Strenghtening Movement. Godspeed, I guess.
P.S. On cope. One of the most popular copes about China is «we've already heard all that about Japan». To me, that alone exposes Americans as an intellectually washed people. You heard that about Japan because Japan was seriously challenging you in trade, like it had previously challenged you in the war. In both cases, Japan was doomed. It's simply too small, too resource-poor, has too few people, and after the war it was under occupation and dependent on your security umbrella. Still it became enough of a problem to prompt that unhinged prophesying of the sexy scary neon-illiminated Japan Inc. turning the US into one big Detroit, and force Congressmen to smash a Toshiba radio on the Capitol lawn, like a bunch of rabid chimps. No other nation, save the USSR, had merited the distinction of such fear. So you won again. So what? The correct takeaway is not this self-congratulatory «America fuck yeah baby, our values!» attitude, but understanding that a structurally disadvantaged nation of 120 million East Asians can push you to the ropes; that in some very relevant ways they are more productive than you. Then, you ought to have noticed that China is 12 times larger in population, 25 times larger in territory and resources, is fully sovereign, and (contrary to the very strange stereotype divorced from what we see in everyday individual performance) Chinese people are not racially inferior to the Japanese. It's more or less Japan times 12, with unsurprising extra benefits of scale, clustering and on top of that a more competent industrial policy (MIIT>METI, as evidenced by the rigid embarrassment of Basic Hydrogen Strategy vs the civilization-scale triumph of the «New Three»).
To predict that they'll somehow fizzle out before eclipsing your global power (ie, more than 25% of your «per capita power») is to assert that either you have some near-supernatural tricks up your sleeve or that they're, bluntly, subhumans. Bugmen. An euphemism like «barbarians», «commies» or something also works. You're free to say that, but I think it's a bit too bold of a strategy to bet your nation's future on.
DaseindustriesLtd
late version of a small language model
6d ago
Consider John Derbyshire’s essay in infamy The Talk: Nonblack Version, focusing initially on its relentless obnoxiousness, and attentive to the negative correlation between sociability and objective reason. As Derbyshire notes elsewhere, people are generally incapable of differentiating themselves from group identities, or properly applying statistical generalizations about groups to individual cases, including their own. A rationally indefensible, but socially inevitable, reification of group profiles is psychologically normal – even ‘human’ – with the result that noisy, non-specific, statistical information is erroneously accepted as a contribution to self-understanding, even when specific information is available.
From the perspective of socially autistic, low-EQ, rational analysis, this is simply mistaken. If an individual has certain characteristics, the fact of belonging to a group that has similar or dissimilar average characteristics is of no relevance whatsoever. Direct and determinate information about the individual is not to any degree enriched by indirect and indeterminate (probabilistic) information about the groups to which the individual belongs. If an individual’s test results are known, for instance, no additional insight is provided by statistical inferences about the test results that might have been expected based on group profiling. An Ashkenazi Jewish moron is no less moronic because he is an Ashkenazi Jew. Elderly Chinese nuns are unlikely to be murderers, but a murderer who happens to be an elderly Chinese nun is neither more nor less murderous than one who is not. This is all extremely obvious, to obnoxious people.
To normal people, however, it is not obvious at all. In part this is because rational intelligence is scarce and abnormal among humans, and in part because social ‘intelligence’ works with what everyone else is thinking, which is to say, with irrational groupish sentiment, meager information, prejudices, stereotypes, and heuristics. Since (almost) everybody else is taking short-cuts, or ‘economizing’ on reason, it is only rational to react defensively to generalizations that are likely to be reified or inappropriately applied — over-riding or substituting for specific perceptions. Anybody who anticipates being pre-defined through a group identity has an expanded ego-investment in that group and the way it is perceived. A generic assessment, however objectively arrived at, will immediately become personal, under (even quite remotely) normal conditions.
Obnoxious reason can stubbornly insist that anything average cannot be about you, but the message will not be generally received. Human social ‘intelligence’ is not built that way. Even supposedly sophisticated commentators blunder repeatedly into the most jarring exhibitions of basic statistical incomprehension without the slightest embarrassment, because embarrassment was designed for something else (and for almost exactly the opposite). The failure to understand stereotypes in their scientific, or probabilistic application, is a functional prerequisite of sociability, since the sole alternative to idiocy in this respect is obnoxiousness.
DaseindustriesLtd
late version of a small language model
7d ago
Out of curiosity, do you speak/read Mandarin?
I don't. I'm trying to learn though. The point isn't that it literally describes a nation as it is. The point (more than a bit sentimental one) is that the ambition is deeper and more interesting than "Warsaw Pact shithole, Asia, really big", and openly stated across infinity of Party Nomenklatura documents that ≈nobody is willing to read seriously. My proposal to look at the literal characters is an attempt to break through the cognitive barrier this negative charisma duckspeak creates.
The Chinese are mostly petty men and women, like elsewhere (arguably more than elsewhere). That's fine. They have a (compelled) respect for hierarchy, and enough thinkers with enough influence, who can make meaningful nudges. It's hard to notice for cultural reasons, and the Chinese themselves are very cynical about what they're doing. But the CPC, at least in some eras including Xi's one, is a sincere ideological-civilizational project with unironic Chinese characteristics.
I recommend reading this, was pretty surprising to me. https://x.com/kyleichan/status/1992405985626124744
The point of that argument is not that China is incapable of building datacenters, but that America hasn't lost it's ability to build - it's just very focused on profit.
To restate my point, I think «lost ability to build» is melodramatic, but what is definitely true is that even «datacenters» are not a very impressive building project by Chinese standards, even adjusted for population. China could do that trivially but mom won't let them have the chips. It's ≈assembly and construction, Chinese «building» is at this point profoundly wider and deeper, they run VAST supply chains from mines to refineries/smelters to factories to shipyards. Americans are already running into constraints like having to ship transformers (physical parts, not LLMs) for their coveted Manhattan Project datacenters (eg Stargate) from China. The grid upgrade is a horrible slog. You are probably well aware of the REE context by now (read this for more if you haven't https://www.notboring.co/p/the-electric-slide).
Americans are good at building McMansions and installing HVACs, they have the workforce for that and in theory it's fungible. It remains to be seen if they can do better.
I don't know if I should trust the economists the way I trust psychologists and social science majors, the way I trust engineers or somewhere in between.
I would say that some economists are very correct but even they can be frustratingly dogmatic or outright deceptive, which nudges me towards «social science» field. For example here https://research.gavekal.com/article/unraveling-chinas-productivity-paradox/ a very fair argument is being made, except the point about FGMs is false and I bought it at face value. Lost face, very sad. Popular Total Factor Productivity stats are just gibberish. And so on. You have to scrutinize everything.
DaseindustriesLtd
late version of a small language model
7d ago
Secondly, most of their people don't have a cultural aptitude for tinkering, exploring, developing in a truly creative sense. When the high esteem strategy in life is to grind 996, be an excellent bureaucrat, and a cold bug person, you're not going to get weirdo genius inventors and you're especially not going to get weird genuis asshole corporate leaders --- Jobs, Musk, Bezos, Gates et al would've been hammered into shit in childhood in China because they're non-conformity is off the charts. Jack Ma is the only homegrown Chinese "visionary" who didn't leave, and he was either some sort of plant by the CCP or became a permanent political prisoner because he did a good job and built a compelling tech company.
I'll respond in more detail later, but for now I'll share a translation from a Zhihu post because it addresses several points here in an amusing way.
Who is Liang Wenfeng, the Founder of DeepSeek?
Qingfeng Xuezha (The Breeze Academic Underachiever) North American Computer Science Professor; Technological Innovation and Entrepreneurship; Providing Value, Emotional, and Knowledge-Based Services. Navigator Duan Xiaocao and 4173 others agree.
I've seen a lot of discussions about Liang Wenfeng online. Yesterday, I happened to have a phone call with a close friend from the same university year, and we also talked about Liang Wenfeng. So here I am, brazenly invoking my university classmate Liang Wenfeng. Some netizens want to know what Liang Wenfeng was like during his undergraduate days before he ventured into investment and the AI industry. This answer is meant to satisfy a bit of everyone's curiosity. I hope these "revelations" won't affect Liang's privacy. If they do, please remind me promptly, and I will modify or delete the answer.
The answerer and Liang Wenfeng were both in the 2002 cohort of Electronic Information Engineering at Zhejiang University (ZJU), not in the same class, but participated in the same Electronic Design Competition. Although we had some contact during the four years of university, because we weren't in the same dormitory or class, my impressions of Liang Wenfeng are limited and fragmented.
Impression 1: In our sophomore year, while we were obediently attending classes, doing homework, and preparing for exams, Liang Wenfeng was already self-studying digital circuits and analog circuits and had begun his own engineering practice. What left a deep impression was that he personally handled everything from circuit design, PCB layout, microcontroller programming, to UI design, creating something like a miniplayer software (doing Software UI in 2004 was a high-skill endeavor). He modified an ordinary guitar into an electric guitar, where the guitar's string sounds could be controlled via a UI on the computer. This project seemed incredibly impressive at the time; we all looked at it in awe. He humbly said the guitar's tuning wasn't great and it would be better if it could tune itself automatically. This can be considered a testament to the seed of his ideas about AI intelligence back then.
Impression 2: He rarely attended classes; most courses were self-taught. The answerer speculates the reason was he felt the teachers' pace was slow, a waste of time, and self-learning was faster. The downside was not following the teacher's emphasis on key points, which could hurt during exams. Liang Wenfeng's GPA in the major back then wasn't outstanding; it was upper-middle, not reaching the line for guaranteed postgraduate admission (保研线) (at ZJU back then, the proportion for guaranteed admission to the university's own postgraduate programs for ordinary majors was the top 5%). He later secured guaranteed postgraduate admission through winning the National First Prize in the Electronic Design Competition. This will be mentioned below.
Impression 3: During university, Liang Wenfeng traveled around several provinces in East China on his bicycle. Surprisingly, he often spent nights finding a spot in the wild to sleep on the ground, completing the trip without spending much money. This matter hasn't been verified; the answerer learned about it from the hot post "Liang Wenfeng, the Pride of 02 Telecommunications" on the 88 forum during graduation. The poster back then was also one of his teammates from the Electronic Design Competition, so the credibility should be quite high.
Impression 4: Liang Wenfeng and two other classmates from the same department signed up for the National Undergraduate Electronic Design Contest during the summer of their junior year. None of the three were top students in terms of academic grades, but their competition performance was outstanding. Naturally, Liang was the main force of the team. During ZJU's internal training camp, he single-handedly completed many design tasks. In the final competition, their team won first place in the province and the National First Prize. All three earned the qualification for guaranteed admission to ZJU's postgraduate programs without examination (免试推荐). However, because the national award announcement for the Electronic Design Contest that year was in October, they missed ZJU's guaranteed admission timeline for that year. Therefore, Liang could only start his postgraduate studies one year later. This explains the one-year gap between his undergraduate (2002-2006) and postgraduate (2007-2010) studies. It is said that during this gap year, he continued working on electronic sensing system design and products, something related to marine navigation, handling hardware, software, and algorithms all by himself. Every electronic system he built during his undergraduate years could easily suffice as a master's thesis for an electronics major.
Impression 5: Liang Wenfeng has always been low-key, just like he was during undergraduate days, so much so that many classmates in the same major weren't very familiar with him. Many heard of him through the National First Prize he won in his senior year. Therefore, it's not surprising to us, his university classmates, that he didn't come out to publish an article, say a word, or record a video amidst the overwhelming popularity of DeepSeek earlier. Ordinary people don't possess such composure and steadiness. (Addendum: Thinking back now, Liang Wenfeng isn't deliberately low-key; rather, his incredibly strong focus on his work makes him appear low-key - like Huang Yaoshi's final evaluation of Zhou Botong: "Old Urchin, Old Urchin, you are truly remarkable. I, Huang Laoxie, am indifferent to 'fame.' Master Yideng sees 'fame' as illusory. But you, with a mind empty and vacant, never had the notion of 'fame' in the first place, which puts you a step above us.")
Conclusion: Liang Wenfeng created his own success in his own way. He didn't live his university life according to the traditional standards of a "good student," nor did he study worldly social skills. He is a classic case of "Be Yourself" among Chinese university students and an example of contemporary intellectual youth entrepreneurship changing their own destiny (even the nation's destiny). Huanfang (幻方) was just the appetizer; DeepSeek is only the beginning. As an old classmate, I'm very happy to see him making outstanding contributions to the world's technological development and also honored to have seen the fledgling eagle before it soared across thousands of miles.
I hope the above sharing can provide some inspiration and motivation for China's tech-savvy youth. Chase your dream, and be yourself!
Yixiao Daxia (Smiling Hero) History and Current Affairs Enthusiast, Secretly Observing the World. 4358 people agree with this answer.
The answers are very fragmented. I carefully collected some information to try and organize it.
1. Birth Background and Early Experience
Liang Wenfeng was born in 1985 in Mili Ling Village, Qinba Town, Wuchuan City, Zhanjiang, Guangdong. His family circumstances were indeed ordinary; both parents were primary school Chinese language teachers, basically with no significant background. Liang Wenfeng made it mainly through studying.
Liang Wenfeng attended Meiling Primary School near his hometown in Wuchuan for elementary school. Both his junior and senior high school were at Wuchuan No.1 Middle School. He had some talent in mathematics; during junior high, he had already self-taught high school mathematics and started reading university-level math textbooks. In the 2002 college entrance exam (Gaokao), Liang Wenfeng scored 806 points, ranking first in Wuchuan No.1 Middle School, 14th in Zhanjiang City, and around 100th in Guangdong Province that year.
His first-choice application was for the Electronic Information Engineering major at Zhejiang University, graduating with a bachelor's degree in 2006. The year after graduation, 2007, he entered ZJU's Communication Engineering postgraduate program, graduating with a master's degree in 2010 (if it were a continuous bachelor's-master's program, graduation should have been 2009. Whether it was because he took the exam twice or something else is currently unknown).
2. Liang Wenfeng's Stock Market Life
As mentioned above, Liang Wenfeng had some talent in mathematics, and his undergraduate major was Electronic Engineering. Combining these two, the best application field he discovered was undoubtedly stock market trading. Therefore, during university, he developed a strong interest in financial trading. In 2008, Liang Wenfeng was 23, likely in his second year of master's studies, and began experimenting with automated trading in the A-share market with a principal of 80,000 RMB.
In 2010, the year he graduated, the stock market was in a downturn. However, it is said that Liang Wenfeng, through partly automated trading strategies, made 1 million RMB, gaining significant fame at the university and being called the "Campus Stock God."
After graduation, Liang Wenfeng did not seek employment nor start a business. He remained a retail investor, tinkering in the A-share market, continuously trying to write quantitative, automated strategies, testing them in the market to see if they could generate returns. It is said he once left Hangzhou and rented a place in Chengdu, closing himself off to trade stocks for three years.
If this stock trading venture hadn't succeeded, Liang Wenfeng would have been a typical negative example criticized by many. Imagine, a graduate from a prestigious university, not pursuing a proper career, and stock trading easily criticized as having a gambling addiction.
It wasn't until 2013, presumably after making considerable money from the stock market, that he began to end his status as an unemployed retail investor and started institutionalizing himself.
That year, he and his classmate Xu Jin established Hangzhou Yakebi (雅克比) Investment Management Co., Ltd. Generally, such asset management companies issue private fund products, get registered, and then raise money for investment. However, I guess it's likely that during the Yakebi phase, Liang Wenfeng and Xu Jin were similar to their previous retail investor status, probably lacking the qualifications and fundraising ability to issue products. The difference was having a company identity; their main work still focused on continuously researching, refining, validating, and improving their quantitative trading strategies.
After two years, the Hangzhou Yakebi company might have encountered issues, or perhaps they wanted to become a sunshine private fund (阳光私募), and the company didn't meet certain requirements, so it was abandoned. In 2015, he and Xu Jin together established a new company, Huanfang (幻方) Technology, and began the process of becoming a private fund manager (奔私).
In 2015, a recruitment post by Huanfang on Tsinghua University's Shuimu Community stated that Liang Wenfeng personally grew his 80,000 RMB principal from 2008 to 100 million RMB in profits over 7 years. It's unknown if this is true. If true, that's 1250 times in 7 years, basically tying with "Beijing Trader" as one of the fastest money-makers among retail investors in A-shares, and should be the domestic stock market's return champion. If the 100 million was accumulated through profit sharing during the Yakebi phase by raising significant external funds, then it involved substantial external leverage.
The period 2015-2017 was likely the most critical phase for Liang Wenfeng's stock trading. During this stage, all the quantitative trading explorations accumulated earlier finally bore fruit, and he successfully transitioned to a private fund institution, with asset management reaching a certain scale.
In 2016, Huanfang launched its first complete AI strategy. In 2017, they fully AI-ized their investment strategies. Presumably, their high-frequency trading AI strategy fit the characteristics of the A-share market very well, performing excellently. In 2017, Huanfang Quantitative's assets under management (AUM) broke through 30 billion RMB, and in 2018, they won the Private Fund Golden Bull Award (私募金牛奖).
Then things took off uncontrollably. In 2019, their managed funds exceeded 100 billion RMB. In 2021, they broke through 1 trillion RMB. However, by the end of 2021, perhaps due to the sheer size, over 100 products under Huanfang Quantitative saw declines exceeding 10%, causing investor losses. Subsequently, Huanfang Quantitative gradually reduced its funds under management.
By the end of 2024, Huanfang Quantitative's AUM was 45 billion RMB, with 63 fund products under its umbrella. However, performance differentiation is noticeable; 29 stock quantitative long-only products mostly maintained slight profits, while all 36 quantitative hedge-type products incurred losses. Of course, this is also related to the 2024 market conditions and policies. In 2024, the China Securities Regulatory Commission (CSRC) imposed significant restrictions on quantitative trading, likely preventing their high-frequency products from functioning normally.
So, here I must also advise my fellow A-share investors: you must work harder, diligently study the market every day, analyze companies, and focus on operations. Only then can you better compete on the same stage with Liang Wenfeng in the A-share market and defeat him.
3. Liang Wenfeng's AI Breakneck Advance
On October 21, 2016, Huanfang's first stock position generated by a deep learning algorithm model went live for real trading. They began using GPUs for computation. Before this, algorithms mainly relied on linear models and traditional machine learning algorithms, with model computation primarily depending on CPUs.
Since then, his breakneck advance in AI began. In 2019, Liang Wenfeng started large-scale procurement of GPUs, self-developing the Huanfang "Firefly One" (萤火一号) AI cluster, equipped with 500 graphics cards, interconnected with a 200Gbps high-speed network. In 2020, "Firefly One" had a total investment of nearly 200 million RMB, equipped with 1100 accelerator cards, and was officially put into use that year, providing computing power support for Huanfang's AI research. In 2021, presumably having really made money, Huanfang invested 1 billion RMB to build "Firefly Two" (萤火二号), equipped with about 10,000 NVIDIA A100 GPUs, breaking through the physical limits of the first phase and doubling computing capacity expansion.
After Huanfang's hardware and funding scale expanded, quantitative trading likely encountered some difficulties. Firstly, making money isn't as easy when the volume is too large. Secondly, the A-share market in 2023-2024 experienced a "Northern Myanmar"-like trend (a metaphor for a difficult/unpredictable market), with investors complaining bitterly, and regulators began supervising quantitative trading. Huanfang started reducing its funds under management from 2021, almost halving it. So, the hardware and computing power prepared for quantitative trading became idle and needed a new direction.
In 2023, Liang Wenfeng recognized the prospects in the field of Artificial General Intelligence (AGI). In July, he officially founded Hangzhou DeepSeek Artificial Intelligence Basic Technology Research Co., Ltd. (深度求索), focusing on the research and development of AI large models. In less than a year, in May 2024, DeepSeek released the mixture-of-experts language model DeepSeek-V2. On December 26, they launched and open-sourced the DeepSeek-V3 model, the version most of us used during the Spring Festival. The entire training process used less than 2.8 million GPU hours, costing about 40 million RMB.
On the evening of January 20, 2025, they released DeepSeek-R1. Its performance in mathematics, coding, and natural language reasoning tasks is comparable to OpenAI's o1 official version. They simultaneously open-sourced the model weights and training techniques, causing a huge stir worldwide.
DeepSeek directly shattered the American plan to monopolize cutting-edge AI technology and computing resources because it is both free and open-source. Anyway, I just made it casually; anyone who wants to use it can take it. I'm not making money from this, just for exploration and fun.
OpenAI: I'm getting a headache. I charge $150/month, $1800/year, and you're giving it away for free. What am I supposed to do? You're not charging either; what are you after?
Liang Wenfeng: It doesn't matter if I make money or not. What's important is that you can't make money!
DaseindustriesLtd
late version of a small language model
7d ago
I find this topic very irksome, for reasons that can be gleaned from some jingoistic comments below. "Breakneck" is also an annoying gimmick. They're not engineers. They've got a ton of engineers but the CPC is still heavy on lawyers and straight up apparatchiks.
China is completely non-mysterious. Any specific domain, like HSR, is pointlessly nitpicked at and debated but matters little in isolation, and the big picture is very straightforward and expressed in their official messaging. Maybe the thickest layer of obscurity is official translation. For example: «中华人民共和国». «People's Republic of China». Character by character: something like "The Middle Splendid Land's People's Common-Harmony State". That's what they intend to be.
It is a modern (as opposed to postmodern) state, with Leninist ideology, built on top of Chinese Confucian civilization, with enormous, high-IQ population, led by a man who's passed through hardship and one of the most competitive and cutthroat filters in any system ever, a product of Party-arranged marriage between two other Communist zealots. I think Xi's character is actually misunderstood and important, especially given what the Superpower Number One has got. To get a feel for it, I recommend reading this interview (1, 2) on a book about his father Xi Zhongxun (btw, he's the guy who invented Special Economic Zones among other things). An excerpt:
Jon Sine: This one is from 1976, the very late years, near the finale. At this point, Deng is being widely denounced and criticized. Yang Ping goes into Xi Zhongxun’s apartment at 8 p.m. one night:
“Yang was surprised to find Xi drinking strong cheap liquor and crying alone in the dark. Xi explained that it was his son Jinping’s birthday. Xi said, ‘Your father is better than I am; he took such good care of you. I am also a father, but because of me . . . Jinping only narrowly escaped death!’ Xi then proceeded to tell Yang about Jinping’s experiences during the Cultural Revolution. Yang later wrote, ‘That night, Old Xi spoke to me, and at the same time, he cried. He kept saying he had let down everyone in his family. He said that in terms of taking care of his entire family, his behavior had been criminal and so on. One could say that his emotional state was approaching a total lack of control. It made me feel extremely sad. Normally, his words would be very concise. He wasn’t verbose, and he didn’t repeat himself. He definitely was never incoherent.’”
At the end of this same paragraph is the thing that stuck with me the most — Xi Jinping comes to visit him a few days later. They’re both sweltering because it’s summer and they’re both sitting in their underwear smoking as Jinping recited Mao speeches from memory while Xi Zhongxun watched. At some point near the end of the book, you say that we shouldn’t necessarily think of Xi Jinping as thinking, “How could I be loyal to a party that treated my father so badly?,” but rather the inverse — “My father sacrificed so much for the party, yet still is this loyal, and still wants me to be reciting Mao speeches. How could I ever transgress that party?” In some ways, this underwear incident actually helped make that stick a little bit more for me.
Basically, to understand China the easiest strategy is to stop coping, take them at their word about what they are and what they're doing, and watch as things become predictable. How cutting-edge capabilities are deployed faster and at larger scale, how air is getting cleaner, how problems just get solved (except profound structural ones no society knows how to solve – like fertility or real estate bubble, which they are deflating), how in 7 years of «slowing down» or «collapsing» they go from taking American export controls lying down to retaliating so severely that Trump is pressed to concede. How we go from «haha Huawei will die» to «please buy H200s». These are a people and a system that is very good at completing tasks. It's how a state should be. Its values may be alien, but operationally, all serious modern states were similar. Some mix of dirigisme and free market, competent leadership with skin in the game, investment into human capital, infrastructure buildout. The US was this. "Datacenter buildout" is not this. Does anyone seriously think they will have trouble building sheds with lots of cooling and grid connection. They have the world's best HVDC system, they ate several major markets in the last 5 years, their heavy machinery is penetrating German/Japanese markets already. They'll be fine.
What merits explanation is not China but the dysfunction of Western societies, the decline of civilization really.
DaseindustriesLtd
late version of a small language model
1mo ago
We filter the pretraining datasets. Extremist materials are among the things routinely filtered out, and often not targeted in scraping at all. LLMs actually learn about 4chan from ADL.
DaseindustriesLtd
late version of a small language model
1mo ago
That's what you find noteworthy?
DaseindustriesLtd
late version of a small language model
1mo ago
People consistently underestimate Musk. In particular his ability to pick a niche with apparent incumbents, say «that could be done better», bulldoze through cringe and come out on the other side with a product that redefines and expands the market. Sometimes he fails and abandons the effort. I think this could happen here, too. But not necessarily, nor even likely. He wants to do to Wikipedia what he did to Roskosmos and other legacy launch providers. He has emotional stake in this, he has the resources and allies for this, and he has the flow of Grok interactions on X to lean on. He can make it work.
There will be a Grok 5, and Grok 6, and they'll be vastly more powerful, not just as modern-day LLMs, they'll have continuous learning and strong multimodality. The main feature you need for good article generation is aggregating tens to hundreds of data points and deeply processing it, meaning context in the millions of tokens and probably weight updates or something functionally close; Grok will be there. Layout, flow etc. are easily solved if you apply work to it, it's trivial compared to general coding and we've come very far with coding LLMs (people who say they're terrible lack the sense of perspective, 2 years ago they were ≈unusable). Even if currently many higher-quality pages are handcrafted, that'll be useful data.
Judge this thing by its strong points, not by its slop and cringe.
Compare:
DaseindustriesLtd
late version of a small language model
2mo ago
Should I read the book or can you briefly explain the motives of Willoughby here?
DaseindustriesLtd
late version of a small language model
2mo ago
I mean, economic Armageddon doesn't imply that Chinese assets will do great. And in fact they're doing poorly, this is a lose-lose move. I'd be investing into… idk, India?
DaseindustriesLtd
late version of a small language model
2mo ago
American reactions are pretty bizarre, I don't understand how these people take themselves seriously. Take this «House Select Committee on the […] CCP»:
“China’s action today is an economic declaration of war against the United States and a slap in the face to President Trump amid his efforts to fight for a level-playing field. China has fired a loaded gun at the American economy, seeking to cut off critical minerals used to make the semiconductors that power the American military, economy, and devices we use every day including cars, phones, computers, and TVs. Every American will be negatively affected by China’s action, and that’s why we must address America’s vulnerabilities and build our own leverage against China. We should immediately pass my legislation to end preferential trade treatment for China, build a resilient resource reserve of critical minerals, secure American research and campuses from Chinese influence, and strangle China’s technology sector with export controls instead of selling it advanced chips. Xi Jinping only respects strength and I am ready to work with patriotic business leaders, our congressional leadership, and the Trump Administration to show China that its belligerent trade actions will be met with serious efforts to protect the American people, secure our supply chains, and cut off the flow of U.S. capital and technology into China."
Man, why the pearl-clutching? Why the shocked Pikachu face? It's unbecoming. This is explicitly a retaliation for your ongoing and already very impactful attempts to throttle their tech sector, you've declared this war many years ago when you sentenced Huawei to death. Moreover the fact that they'll eventually be in a position to retaliate via REE dependence has been known for well over a decade, and there's been a warning shot in Japan over Senkaku. You chose this route, your team was consistently rejecting all offers of deescalation because you believed to have escalation dominance. So own it. The extent of third-worldist hypocrisy is breathtaking:
China’s new rare earth export controls prove one thing: Beijing will weaponize any leverage it has.
America needs a new normal—not tit-for-tat.
We must strategically decouple and throttle the PRC tech sector before it’s too late. 🧵
Chinese export controls are obviously tit-for-tat for US export controls, so by new normal they basically mean unrestricted always-defect economic warfare. It seems that the US isn't really capable of negotiating, the notion that you can't always bully your way to an objective is alien to these folks (we've already seen this with Bessent's "China has revealed itself as a bad actor"). There's the assumption that the US (or at least "with allies") necessarily possesses some hidden strength that can be activated to indignantly reject the adversary's offer.
Well, maybe there is. I'm optimistic that out-of-China production can be scaled up in a matter of years, if no deal is made. Primarily China intends to cripple defense applications, and frankly how can you object to that, this idea «greedy communists will sell us the materiel to shoot them with» has always been risible. They are also likely going to suppress the planned reindustrialization and (very dubious) robotic labor revolution in the US. All of that «just» reduces CAGR in a wide range of industries for the next 3-7 years, while Chinese physical productive capacity keeps growing exponentially. The demand for chips, though, will definitely be met. In the meantime, the US will have to capitalize even harder on its software/AI advantage. We'll see which is more important.
DaseindustriesLtd
late version of a small language model
2mo ago
The first reasoning model for which we have decent documentation, and one which basically defined the research field (as OpenAI/Google/Anthropic hide their recipes) is DeepSeek R1. They've finally had it published in Nature, too. The supplementary has been very informative, because people kept speculating about implementation details.
…we aim to explore the potential of LLMs for developing reasoning abilities through self-evolution in a RL framework, with minimal reliance on human labelling efforts. Specifically, we build on DeepSeek-V3 Base8 and use Group Relative Policy Optimization (GRPO)9 as our RL framework. The reward signal is only based on the correctness of final predictions against ground-truth answers, without imposing constraints on the reasoning process itself. Notably, we bypass the conventional supervised fine-tuning (SFT) phase before RL training. This design choice originates from our hypothesis that human-defined reasoning patterns may limit model exploration, whereas unrestricted RL training can better incentivize the emergence of new reasoning capabilities in LLMs. Through this process, detailed in the next section, our model (referred to as DeepSeek-R1-Zero) naturally developed diverse and sophisticated reasoning behaviours. To solve reasoning problems, the model exhibits a tendency to generate longer responses, incorporating verification, reflection and the exploration of alternative approaches within each response. […] During training, we design a straightforward template to require DeepSeek-R1-Zero to first produce a reasoning process, followed by the final answer. The prompt template is written as below. “A conversation between User and Assistant. The User asks a question and the Assistant solves it. The Assistant first thinks about the reasoning process in the mind and then provides the User with the answer. The reasoning process and answer are enclosed within ... and ... tags, respectively, that is, reasoning process here answer here . User: prompt. Assistant:”, in which the prompt is replaced with the specific reasoning question during training.
Then they do a large number of rollouts for questions, automatically verify, GRPO creates gradients based on answer correctness within each rollout, model is updated, continue for thousands of steps and that's basically it.
As well as the progressive enhancement of reasoning capabilities during training, DeepSeek-R1-Zero also demonstrates self-evolutionary behaviour with RL training. As shown in Fig. 1b, DeepSeek-R1-Zero exhibits a steady increase in thinking time throughout training, driven only by intrinsic adaptation rather than external modifications. Mak- ing use of long CoT, the model progressively refines its reasoning, generating hundreds to thousands of tokens to explore and improve its problem-solving strategies. The increase in thinking time helps with the autonomous develop- ment of sophisticated behaviours. Specifically, DeepSeek-R1-Zero increasingly exhibits advanced reasoning strategies such as reflective reasoning and systematic exploration of alternative solutions provided in Extended Data Fig. 1a, substantially boosting its performance on verifiable tasks such as maths and coding. Notably, during training, DeepSeek-R1-Zero exhibits an ‘aha moment’, shown in Table 1, character- ized by a sudden increase in the use of the word ‘wait’ during reflections, provided in Extended Data Fig. 1b. This moment marks a distinct change in reasoning patterns and clearly shows the self-evolution process of DeepSeek-R1-Zero.
At the time there's been a lot of suspicion about it but now it's pretty solidly settled and replicated science.
Not enough to make a really useful model (I've tested R1-Zero, it was somewhat unhinged), but for reasoning as such. Everything else is basically a quality of life or speed-increasing implementation detail. Using reward models to grade qualitative queries, branching and merging, iterative dataset augmentation and dozens of other tricks add little to the core principle.
DaseindustriesLtd
late version of a small language model
2mo ago
I mean my entire post, of course. If you want to see how biased they are in less clear-cut cases, feed the wiki page of Kobi Kambon and ask if this is science.
DaseindustriesLtd
late version of a small language model
2mo ago
Tanner Greer is extremely blackpilled on this:
One theme that Xi Jinping repeats eternal is that his cadres must have “calamity consciousness” — real awareness that if they get things wrong they will be responsible for historical disaster. The country is only ever a few steps away from catastrophe.
If you believe that the single largest priority of the US Navy is physical fitness and hazing then you just don’t have that consciousness.
This reform program is decadent. it is superficial. It has no respect for the depth of the Navy’s problems or the catastrophe they might result in.
We are in a very bad place. Culturally, even, the Navy is in a bad place. Had they done something crazy but real on that stage—like promote half a dozen men from the submarine service in a desperate bid to fix the Navy’s operational culture—I would not be writing this.
But Hesgeth did not do anything of the sort. Even on the terrain that he chose—service culture and readiness—what he had to say fell woefully short of the problems we now face.
We are running out of time.
I am being hard on these guys, yes.
I am on hard on them because we are running out of time. We do not have the time to squander attention or resources.
There is need for a calamity consciousness.
Greer is a major China hawk, though. And Hegseth isn't.
I believe this is part of the broader strategic posture adjustment, or at least a hedging bet (not clear if it'll be maintained). The US defense/war department, in Hegseth's vision (I don't think he's intelligent enough to have a coherent strategic vision like, say, Elbridge A. Colby, but there probably are people behind him making this functionally true), is going to implement Monroe Doctrine 2.0, focus on the Western hemisphere. Tough, masculine, no-nonsense, scary bunch, unencumbered by rules of engagement, to more easily topple regimes in South America and pressure neighbors into resource and trade concessions. Death squads eliminating suspected narcos, National Guard prepared to pacify Portland. More like Russian Airborne Forces that exist to terrorize the domestic audience. This is all, of course, noise in the context of conflicts with peer powers, which realistically mean just China and require far more logistical and industrial competence than warrior ethos or indeed individual warriors (only so many guys you can fit on an aircraft carrier). But on that level, the US will rely on strategic deterrence and the hope of transformative results from AGI.
DaseindustriesLtd
late version of a small language model
2mo ago
This is the part of Uncomfortable HBD Facts that is very underdiscussed: Blacks More Likely to have paranoid schizophrenia.
African Americans were four times more likely than Euro-Americans to receive a Schizophrenia diagnosis … race was the strongest predictor of an admission diagnosis of Schizophrenia after controlling for the influence of other demographic variables. Interestingly, Barnes showed that Schizophrenia subtypes were not equally distributed by race, with African American consumers significantly more likely to receive a diagnosis of Schizophrenia-paranoid subtype or Schizophrenia-undifferentiated subtype than Euro-American consumers.
There's a ton of results in this vein, and his Arabic name makes me suspect that his parents were into Nation of Islam stuff, Yakub doctrine and all that, which is of course schizophrenic throughout (very funny though), so it might be hereditary for him, on top of direct schizophrenia-promoting nurture.
But I don't think the beancounting method really explains the problem. Many, many, many more Blacks who don't get any diagnosis and generally function well buy into this zany conspiracist bullshit. In the limit, the whole narrative of Black existence under the yoke of Systemic Racism and mass slaughter of unarmed black men by the police in the US is a giant conspiracy theory with clear paranoid motives, and spotlight cases like Nick Cannon going off about melanin, people buying into Jussie Smollett hoax, Candace Owens spiraling into even crazier hypotheses about secret societies behind Macron's wife being a man, "black scholars" with mind-bogglingly idiotic and racist doctrines propped up as real intellectuals (I like Kobi Kazembe Kambon a.k.a. Joseph A. Baldwin, read it up), etc. etc. all add up to a general schizophrenic-paranoid tone in the Black community. Except – wasn't it Whites (and Jews, but in any case, the well-off, educated non-Black demographics) who championed the doctrine of systemic racism? Aren't Whites also buying into many of these hoaxes and libels today?
These are accomplishments he is very proud of, that frequently come up during his expositions. He will start by telling me how the people at the Social Security office are stealing from him (AFAICT, that was either taxes or a garnishment of some sort), then veer into reciting all the vegetables he eats because he knows how to eat healthy, he cooks for himself, but these people they not eatin' right and it causes problems, mental problems in they head they be havin' mental problems because they don't eat right, not like him because he eats his green beans, real food that he cooks for himself because he knows how to eat right, act right because he learned it in school, third grade, food pyramid, he learned that here in Jersey in school, third grade, and these other people should have learned it but they not acting right, that’s just Jersey, lotta bad people in Jersey, obsessed with money, takin’ from you, takin’ your money.
Unless you exaggerate greatly, Hassan sounds like he has quite a low IQ on top of his "schizophrenia". Mentally, he's close to an 8-10 year old White or Asian kid.
And I think this is the elephant in the room. There is simply not enough capacity to suppress delusions induced and exacerbated by the information environment. There is no clear separation between delusions and sincere confusions, Hassan won't have a mature enough epistemology to distinguish things that he pathologically overfixates on from things that just kinda legit sound right, even if you pump him full of antipsychotics. This is the ground zero of Black American Madness, along with smarter people than Hassan who also have a more profound illness and/or higher ego strength so they keep it together and spread their nonsense around. Then there are borderline cases who are relatively gullible and either propagate the message or simply do not push back, like that smart and non-insane Black guy once did for another celebrated Guggenheim scholar, Ibram X. Kendi:
"They are aliens," I told Clarence, confidently resting on the doorframe, arms crossed. "I just saw this documentary that laid out the evidence. That's why they are so intent on White supremacy. That's why they seem to not have a conscience. They are aliens."
Clarence listened, face expressionless. "You can't be serious."
"I'm dead serious. This explains slavery and colonization. This explains why the Bush family is so evil. This explains why Whites don't give a damn. This explains why they hate us so damn much. They are aliens!" I'd lifted off the doorframe and was in full argumentative mode.
"You really are serious about this," Clarence said with a chuckle. "If you're serious, then that has got to be the dumbest thing I ever heard in my life! I mean, seriously, I can't believe you are that gullible." The chuckle turned to a grimace. "Why do you spend so much time trying to figure out White people?" he asked after a long pause.
Clarence had asked this question before. I always answered the same way. "Because figuring them out is the key! Black people need to figure out what we are dealing with!"
"If you say so. But answer me this: If Whites are aliens, why is it that Whites and Blacks can reproduce? Humans can't reproduce with animals on this planet, but Black people can reproduce with aliens from another planet? Come on, man, let's get real."
"I am being real," I replied. But I really had no comeback. I stood and turned around awkwardly, walked to my room, plopped down on my bed, and returned to staring at the ceiling. Maybe White people were not aliens. Maybe they became this way on earth. Maybe I needed to read more Frances Cress Welsing. I looked over at The Isis Papers on my nightstand.
Well, Clarence then probably got a normal job and never achieved prominence as a thought leader for his people. He wasn't exuberant enough to be made a champion.
…and then comes everyone else, who is simply overwhelmed by moral blackmail, the volume of second-hand "corroborating evidence" and plain emotional confidence. And here we are.
In short my thesis is that a small (large relatively, but still amounting to fractions of a percent of the total population) difference in the rates of paranoid schizophrenia, compounded by significantly lower IQ, creates a critical mass of self-propagating cultural madness in the Black community, which lowers general epistemic standard to rock bottom and then spills over to the broader society, warping the entire default narrative of what it is about. And now we're training large language models on this oeuvre, which seals the bubble of the consensus reality. I bet if I feed this exchange to any frontier LLM, it'll rebuke me harshly with the usual tut-tutting routine about how systemic racism is totally real and Scholars say so.
Regarding your actual question: I think paranoid schizophrenics, no matter how "functional" for the moment, certainly can't have guns. This isn't hard.
DaseindustriesLtd
late version of a small language model
2mo ago
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.
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.
DaseindustriesLtd
late version of a small language model
2mo ago
Except GPUs are even more interchangeable than other weapons systems.
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".
I don't think Xi is playing factorio but the CCP does obviously practice extensive industrial policy. Making it kind of ironic to go this hard against one piece of American industrial policy.
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.
It's kind of incredible to point out how China is willing to force some of their largest companies to supply demand for their domestic chip industry
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.
Absolutely. China has found its success through world trade on sea lanes policed by American military might in an environment built by American diplomats. That's what Pax Americana is.
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.
This is an ironically Trumpist take on things for as much as you rightly excoriate his team's perspectives elsewhere.
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 is also a pretty important part of this all as well
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.
multipolarity doesn't seem very stable in any case
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.
Then I return to being very confused as to why we're going to regret this.
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.
I have to note: I am undecided on what's better for me. I argue for the sake of argument. I believe the current US policy will end up making everyone poorer and American global standing lesser, as in the long term it will guarantee a separate technological civilization existing and building in and around China. So, given how undesirable your hegemony is, maybe that's overall a good thing and I should shill for export controls. Maybe this mad bet on the AGI race will work and I'm wrong, though.
The traitor, the treasonous little worm
This made me smile. Very "nationalize SpaceX" energy. You do realize that your Hail Mary attempt at preserving hegemony largely depends on him? For some reason, Loyal Americans run their hardware companies into the ground. I do think he believes that this game will continue for decades, and China is not going anywhere, it's not going to critically fall behind, and so he wants to keep a piece of that market for the US. And that can be done.
China is already exerting the maximum amount of demand and political pressure it can to try and compete on chips. The internal market demand is irrelevant. The government will guarantee every chip is sold and prop up all the companies making them. Whether or not AI labs can use NVDIA hardware has zero actual influence on the development of their ecosystem. Hardware "lock-in" on these labs is an entirely made up concept.
The internal market demand is irrelevant. The government will guarantee every chip is sold
Asinine. As it's said, "there is nothing to be learned in matters of faith". If anything, this describes Intel. No, market demand is not irrelevant, PRC corporations actually have incentives beyond 5-year plans, largely because they have slim margins. Americans really have worked themselves up into a frenzy with this doctrine that everything in China is massively subsidized and so can be unprofitable forever. It's not about subsidies, they're just more productive than you and have a more ruthless market, to the extent that the state is trying – and failing! – to arrest "involution".
Just because you hate the CCP really, really hard does not give you the license to spew bullshit. Being very confident doesn't help. It is not, in fact, possible to create a competitive ecosystem by decree, even if it's super-duper maximum pressure. This just takes too many people. I know DeepSeek has been asked to and declined to do serious training runs on Huawei due to immaturity of CANN stack. They have this choice, for a little longer. They're typical. There are maybe 2 Chinese companies doing large-scale training on Ascends, and one is iFlyTek, which has been on entity list since forever and has no choice; they haven't achieved much. Even Huawei themselves are yet to release a single compelling model, they literally can't keep top-tier people interested as they leave to companies like DeepSeek. Huawei has 200K employees, for reference.
On a smaller scale, we've seen this when Microsoft attempted to make Windows Phone a thing. Tremendous effort went into it, a formidable corporation was banging against the wall for years, subsidizing the app marketplace, and it all fizzled out. No developers, no users, no network effects, no future.
We know what PAX Americana looks like and it looks pretty good actually. Billions rising up out of poverty
One of those "billions" is in China, can you really take credit for it? I call bullshit, mostly it's just post-WWII economic growth the nexus of which was the US for reasons of not being bombed out, not some profoundly benign and productive doctrine or culture or people. India is illustrative: they wanted to latch onto Pax Americana and get something out of it; what have they got so far for India proper? I am in your "sphere of influence", so to speak, and it really doesn't look like you're spreading prosperity around. In fact it looks like you have nothing to spread, you don't invest, your own riches are a speculative bubble and you mainly "supply demand". You're demolishing your nuclear infrastructure, you don't build anything except datacenters, certainly you can't boast of turning Pakistan into a solar-powered economy or something. Outside a few premium items like these very GPUs, your wares are non-competitive trash that people abroad have to be compelled to buy, you're even pathetically forcing third parties to share your tariff regime to cling to some markets (very funny in this context of "market share is useless"). Yes, in theory you could cheat with AGI, but ask yourself, if a cheat on the scale of AGI is needed to redeem your claim to hegemony, what do you, as a people, stand to contribute? Having created the solutions where you've got AGI earlier than others?
But CCP dominance hasn't even been particularly good for them. China is host to the poorest and least prosperous Chinese people in the world.
This is a very tiresome talking point. They didn't have the benefit of a sane administration until 1978, after which they've consistently had the highest growth rate of all major economies. In any meaningful sense, including consumption spending, general QoL. GDP per capita comparisons are misleading. I've been reading on Taiwan recently and it seems that they're straight up having poorer lives than coastal Mainland Chinese in comparable population centers; like, they have higher costs of living and don't have meaningfully higher salaries. This, too, is Pax Americana; not even the smallest and most important clients can be sure to prosper. What else do we compare to? Singapore, Macao, Hong Kong? Please.
Now, history doesn't start in 1978. But nations change, even under the same regime and slogans. The US of today is not the US of 1960s either.
surely you understand the "equals across the sea" isn't an option on the table. That isn't what is in store if we give up all our advantages in this sector.
No. I don't understand. Why? What happens to the US that did not "win"? Unlike the USSR, China doesn't even have a messianic revolutionary project.
I think this is just wounded ego. You're used to hegemony, it's part of your personal identity, and it slipping away, likely forever, is perceived as existential horror, with appropriate rationalizations. This sounds about as compelling as Russian noises about NATO threat and absolute rationality of going all in to "denazify" Ukraine. In reality Russia could well survive Ukrainian integration with the West, it was merely humiliating (and deserved, certainly so after 2014 when we've demonstrated our mettle in managing "people's republics") but not affecting the survivability of the Russian state, and the costs of war have already far exceeded any sane estimate for costs of doing nothing.
China will take the chips, use them to accelerate their position, including in advancing their own semiconductor industry
Like what, using AI to design floor maps? They're doing it already, it doesn't take a lot of compute. A rather contrived concern.
As soon as China has even slightly competitive chips they will crumple up NVDIA and toss it out like so much garbage.
The thing is, chips are very, very hard and ensuring the supply chain is all outside China has been one of the few truly great American political successes (not that it was hard, this chain was mostly complete when China was around $2000 GDP per capita) . The trifecta of ASML-TSMC-NVDIA (nevermind their multiple one-of-a-kind suppliers like ZEISS, and EDA software) will genuinely take China a decade or more to even approach. They will not have competitive chips. They will have (already have announced for Q1 2026) competitive systems, but those only exist because NVDIA is prevented from exporting the good stuff.
Again, I don't know what I should "rationally" shill for here. And anyway this might be too late. The US has clearly stated its hostility, burned the bridges, and will have to "lose", in a war of its own creation.
DaseindustriesLtd
late version of a small language model
2mo ago
but to do so below market rate?
This is pretty asinine. You're defending export controls with the claim that their absence would… distort markets? Do you think that's what Nvidia is trying to do, sell GPUs below market rate, despite having an unsaturated domestic market that would generate higher margins? Why do you imagine they would hurt themselves like that? Might it be time to install some loyal apparatchiks on board, or do a little witch hunt for Communist agents?
China is not going to forego building their own echo-system and hasn't for any other sector they've found strategically important.
As I've said before, "China" is not omnipotent and cannot create an ecosystem solely through political will and subsidies, they've been trying for decades and it hasn't been working so long as Nvidia was the obvious superior choice. Even now, nobody wants to use CANN if afforded the chance. I think this is how Jensen views this: he's straightforwardly fighting as the CEO of American company Nvidia, not just for line going up in quarterly reports but for enduring global dominance of his stack.
It's banking on the certainty that surrendering our major advantage in the AI race to china for no reason or gain will turn out badly for us, obviously. I can't even fathom how a thinking person could convince themselves otherwise. You've already highlighted their advantages, is your position that the race is already over despite us currently being ahead?
Are you avoiding the question, or does it not parse for you?
I think that to discuss whether "the race" is over, it's important to establish whether a race is happening and what it is that you are racing towards. The US is ahead in AI. Again, without American chips, China will be developing AI slower for the next few years. Is that a "race"? What happens when you reach the finish line? Don't huff and puff, say concretely. Do you build an AGI superweapon that disables their nukes with nanobots? Or what? What's the end goal, in the face of which every thinking person would deem hundreds of billions or trillions of dollars of profits a mere short-sighted distraction? Can you spell it out?
Our one chance at dominance in this sector is remaining ahead in AI and reaping compound interest on that lead whether it's AGI or simply accelerated AI and chip development. If it's not enough
Not enough for what? Like, what's the theory of victory here? Repeating the Great Divergence, now with automation, relatively growing so quickly that China is forever left in the dust? Lights-out factories spawning across the US, producing ungodly goods optimized by AGI, incomprehensibly advanced weapons systems, Pax Americana becoming permanent?
How likely do you think that is? And what happens if this doesn't work out?
I think the answers are basically "yes/likely/better not to think of this", and personally, I believe this is all deluded and very much in the spirit of last days of Nazi Germany. Both sides will have adequate AI to increase productivity, both will have "AGI" at around the same time, you're not going to have some dramatic inflection point, you will not leave them in the dust as a military or economic power, you'll just slow down global economic growth somewhat, and in the long run end up poorer and have a smaller slice of the global market. That's all.
DaseindustriesLtd
late version of a small language model
2mo ago
We benefited our rival for... What? A few quarters of sales for a couple of firms?
Do you realize that the entire windfall from Trump's tariff nonsense would be an order of magnitude less than those quarters, even as it destroys similar value (hundreds of billions)?
It seems Americans aren't happy with this whole concept of trade anymore. If they buy foreign stuff, that's bad because they're losing dollars, gotta reindustrialize and implement tariffs. If they're selling stuff to foreigners who aren't completely inept and subjugated, that's also bad, because then those foreigners may develop and get richer, and for an American, the world is zero-sum, so the only Deals Americans are now willing to make are that which make the other party poorer, like the humiliation rituals you subject "NATO allies" to. Trump's rhetoric around coercing South Koreans and others to "invest" (he apparently understands FDI in very childish terms, "they give us moneys because they're our bitches") completes the picture.
Yes, I admit this makes me even more sympathetic to China.
DaseindustriesLtd
late version of a small language model
2mo ago
5G/6G is not very relevant to this issue, but they have extremely advanced datacenter network architecture and their new systems are based on it. This will allow them to cope with lower performance of individual chips.
If China has the ability to leapfrog Nvidia and other western AI tech, they're gonna do it irregardless of any sanctions on chips
This is not true. People act like "China" is a perfectly coordinated single entity, a game of Factorio Xi plays, but it's still a country with different economic actors. If Huawei can't sell their crap because everyone in China who is actually good at AI uses CUDA and Nvidia hardware (like, again, DeepSeek), Huawei will not improve as rapidly. Subsidies in isolation cannot replace organic ecosystem support, they just prolong the agony, and at the current level not even China can subsidize the development of the entire supply chain, it's to the tune of hundreds of billions of dollars.
I will answer thusly.
My default moral intuitions aren't that different from modal American ones 50 or so years ago. Russia is a multiethnic society, clearly defined by one ethnos and culture (mine, to a first approximation), which does not possess the instinct or inclination for clannish diasporic behavior. We are more ethnocentric than modern Westerners but not by far. We assimilate easily in Western societies, find them an upgrade to our own, and generally agree with the way of the West, whereas the ways of The Rest are seen as unfair and backward, if demonstrably adaptive on the personal and sometimes collective level. Ethnocentrism specifically has been investigated in a toy model I like, by one Artem Kaznatcheev and friends, in Canada in 2013, and the conclusion was that it “…eventually overcomes its closest competitor, humanitarianism, by exploiting humanitarian cooperation across group boundaries as world population saturates”. Intellectually, then, it appears necessary to develop a system that can defend that superior humanitarian way, and the unpalatable conclusion is that in practice it amounts to something not unlike ethnocentrism – aggressive policing of defectors, attention to proxy metrics of defection, operational presumption of non-assimilation, and rejection of comforting lies about universality and natural attractiveness of preferred values. Western experiment with mass immigration and “race-blind meritocracy” is clearly a cheap hack motivated by Western cognitive biases, myopic financial considerations and such, not any concern for long-term preservation of the Way. Similar thoughts are common for sympathetic peoples on the periphery of Western civilization, which is why we see Slavs, assimilated Jews, Moldovans and so on disproportionately represented among the European right.
On a more philosophical level, I don't know. Ethnocentrism is a crude but effective means to preserve the intrinsic direction of a people, it increases the activation energy for changing course, and lets the direction be explored further. Moral intuitions aside, I'm not convinced that the Western direction is truly superior; it would be premature to say so.
I'm pretty open to the idea that diversity is good — real diversity, not this consumerist Western appreciation of cuisines, not the lukewarm respect for ethnic varieties of ideas the West already accepts. Why is diversity good? I consider great men of history to be scientists, and civilizations to be ongoing longitudinal natural experiments – about the limits of human nature, society, what kinds of minds should be incentivized to develop, what notions of goodness are viable and lead to more adaptive behavior, better instrumental outcomes and, in the limit, to greater collective and individual flourishing. Some experiments achieve negative results, invalidating the hypothesis for observers, but it's always a pity if this happens for contingent reasons like a natural disaster, or an opportunistic alliance, or luck of the draw in relative timing of access to some preexisting technology. I am distraught at being unable to know how failed civilizations could have developed to their “mature” stage, given a couple more millennia of literacy and a handful of extra IQ points, or passing a good reform, or contacting a powerful idea earlier; where would they have met their ceiling. When possible, one should run experiments in controlled conditions, after all.
The West is a beautiful experiment, plausibly the most successful that has ever been proposed. Its core thesis, stripped of the ever-changing scaffolding, is something like “human nature inherently has the spark of God's love and wisdom, therefore individual freedom is good and barriers to its realization are at worst prejudices, at best training wheels and must be systematically removed”. It's been working very well. But this near-genocidal desire to universalize the way smells less and less like sincere proselytism, and more like anxiety, fear of the hypothesis getting falsified. The truth is, the West has no clue as to what made or makes it work, beyond currently-commoditized pieces like capitalism (but whence capitalism? If “because freedom”, why does it port to societies that don't adopt the rest of the package?) Americans sometimes boast of “nation-building” Germany and Japan, developing “institutions”, and that reveals the hollowness and vulgarity of the doctrine. Germany and Japan, seriously? Japan got destroyed in the first place precisely because it got competitive, while remaining philosophically largely alien; Germany was actively advancing a divergent branch of the Western thesis. And today, both these nations are deeply troubled. Nation-building in less performant societies has a dismal track record. The West doesn't really know what to teach others so that it sticks. Really, what made the West into what it is, what was the generative function behind those generically adaptive innovations? Christendom (adopted Middle Eastern teaching, effectively dead now)? “HBD” in the sense of high IQ and conscientiousness – OK but why did it happen, just deep time migration patterns, cold winters, founder effects? What's the lesson here, pray to RNG? Galton-style Social Darwinism, Gregory Clark's Anglo class eugenics (grotesquely replaced with education-mediated assortative mating, also largely dead, and their practical implications made taboo in the resultant society)? “High trust” and non-kin cooperation by default (as covered here, a giant exploit for people who practice kin cooperation, and thus a mere bootstrap phase)? Science? Everyone above 95-ish median IQ can do science. The discovery was invaluable, but can “the West” come up with anything of that caliber ever again? Rather, can you even do what you used to do? We seem to be near the end of the session. Do you even know if you want to live? When you have full automation, will you put forward an argument for not just exercising your freedom to pass away replacing yourselves with machines, like you're currently doing with immigrants? Of what nature will that argument be?
China is another large old civilization. They've been running their own experiments. Their most enduring research program is Confucian. At the risk of butchering it, Confucianism says something like “humans aren't very good and are prone to self-interested behavior. Individual humans are not even human, they can only be elevated from monkeys via social context, and even then they default to barbarianism. But if compelled to cultivate “virtue”, starting on the mundane level of filial piety, hierarchical propriety and standardized ritual, escalating to mental discipline and scholarship, if rewarded with reproductive opportunity for utmost compliance, if the peace is maintained for many generations – they can build hierarchical societies of unbounded scale and splendor; and eventually, more of them become Superior Persons capable of and entitled to correct independent moral reasoning, and those will ennoble everyone else”.
It is debatable how seriously that has been pursued, but I'd say at least as seriously as the Christian/Western program. Both have undergone course corrections that arguably reflect growing out of their scaffolding and purify the original strategy. The West going from theism to deism to non-superstitious interest in the Universe, generalizing the validity of “love” and “freedom” beyond traditional norms. The East purging “thieves of virtue” along with ossified ritual and adopting a more common-sensical epistemology. To an extent this can be decried as trivialization and loss of function on both sides. There's been substantial convergence, but the divergent bits are what's at stake. Right now, I think the Eastern project is showing more promise, and the West is no longer in a position to lecture them on how to steer it. They're more ethnocentric? Less individualistic? They're authoritarian? Their society feels wrong? OK, I hear you. But they're solving social trust, they're solving – in their own way, less charismatic and more transactional – international relations, they're even solving creative expression, while having solved long-term large-scale coordination to a greater extent. And crucially, this isn't their first rodeo, they've had massive collapses and comebacks, they're the only major player that has a sophisticated applied discourse about civilizational recovery. Isn't it saying something that they've fallen behind, failed catastrophically, but have recovered, denied you the option of converting them, and are again pursuing their own program? Isn't it exciting that another solution can exist? Aren't you curious of where it will peak? Of course, they're doing well enough that another questions, for example “how much of the light cone will we be able to claim at this rate” are becoming salient for savvy observers. So it is necessary that they be treated as competitors, not just an interesting alternative path.
Jews, likewise, are a unique research program. They have an insular doctrine of their exceptionalism and special nature of themselves as “the people of God”, their moral obligation to biologically and culturally perpetuate themselves, a very long story of surviving and adapting, institutions built for venerating and reproductively rewarding exceptional individuals who have superior insight into God's will, they're punching so far above their weight that it's almost comical… and all of that hinges on extreme, almost naive ethnocentrism. They've mixed with Western peoples, experienced some assimilation, and now we're watching them return to a more traditional (indeed, exxageratedly traditional) form, with large Orthodox and Ultra-Orthodox subpopulations having a vast fertility advantage over secular ones which, so long as they exist, provide a bridge to Western decisionmaking, invent spins like “Judeo-Christian liberal democratic values” and “our common Islamic/Communist/Han enemy”, and generally add confusion (partially their own). It doesn't take a genius to recognize that this research program, while fascinating on its own, can derail the Western one, and on top of your own dysfunction and anxious miscalculations it can create very ugly outcomes. We've seen trial runs in the Middle East, and the worst part is that you can barely articulate that it was mostly about them, not your “strategic interest” or oil or some other bullshit. So they, too, ought to be treated as competitors. It's okay, they can take it.
That's pretty much all relevant live players; smaller or less coherent players, who have a latent opportunity to expand their niche, are also more ethnocentric. The West is uniquely non-ethnocentric and has unique moral narcissism about this fact, largely owed to successes of the last several centuries. I think the jury is out on whether this system is sustainable or has the highest ceiling, and you're not entitled to try and “enlighten” others, but you're clearly valuable enough to think of how you can preserve and improve your program in a world of ethnocentrists, and that's what you should be doing now.
Spawn a great man or something, I don't know.
More options
Context Copy link