In February 1895, the Beiyang Fleet was destroyed at Weihaiwei. It had been, on paper, the most modern naval force in Asia: a flotilla of armored cruisers and torpedo boats commissioned from German and British yards, manned by officers trained at European academies, financed by a generation of senior Qing officials who had argued for thirty years that the dynasty could become strong again if it adopted Western methods within a Confucian framework. The fleet was the visible proof of their argument. It took six hours of combat to remove it from the world.
Three years later, on the morning of September 21, 1898, the Empress Dowager Cixi staged a coup against the young Guangxu Emperor. Six of the leading reform advocates were executed in the street outside the Caishikou vegetable market. Two more, Kang Youwei and Liang Qichao, escaped into exile in Japan, from which they would spend the next decade arguing that legitimacy must transfer from the dynasty to the Chinese nation, because the dynasty had proven itself incapable of reform.
Two years after that, the Boxer Rebellion delivered Beijing to an eight-nation expeditionary force. The resulting indemnity pledged the empire's tariff revenue to foreign banks for forty years. The dynasty that had survived the Taiping Rebellion (twenty to thirty million dead, the Yangtze delta in ruins, half the country in the hands of millenarians) fell to a regional uprising in October 1911.
The interesting question is not why it fell. The interesting question is why it fell so much faster than the Tang.
The Late Tang had taken its own catastrophic shock (the An Lushan rebellion of 755) and survived for another century and a half in simplified form. Same kind of system. Same structural exhaustion. Same fiscal collapse, same provincial military fragmentation, same elite saturation. The Tang lasted one hundred and fifty-two years from onset to terminal collapse. The Qing lasted seventy.
The variable that distinguishes them is not material. It is narrative. And the mechanism that compressed the Qing's failure window is one we should be alert to in 2026, because the contemporary system has made the same kind of commitment, and the visible artifacts of that commitment are sitting on cloud-services balance sheets being read out at investor calls every quarter.
This is a dispatch about the difference between systems that quietly decline and systems that publicly promise to rise.
The Plan and Its Artifacts
The Self-Strengthening Movement is one of those historical episodes that is easy to underestimate from a distance. Read at speed, it looks like a familiar story of a declining empire trying and failing to modernize (the Ottoman Tanzimat, the Russian Great Reforms, the Mughal modernizers, all the late-imperial attempts at structural adaptation that were running concurrently in the same century). Read more carefully, it is the cleanest case in the historical record of a specific failure mode that has not yet been given a clean name.
The movement began in the 1860s as a response to the conjunction of two shocks: the Second Opium War, which had ended in 1860 with the burning of the Yuanmingyuan summer palace and the Convention of Peking; and the Taiping Rebellion, which the dynasty was only able to suppress by delegating military authority to a generation of provincial Han officials whose loyalty was conditional and whose armies had no pre-existing place in the imperial structure. By 1861 the dynasty had two existential problems and one new institutional reality: it had survived, but only by becoming something different from what it had been a decade earlier.
The Self-Strengthening response was articulated by a small number of senior officials who became, over the next thirty years, the public face of a particular argument. Wei Yuan, the foundational figure, wrote in 1843: shi yi chang ji yi zhi yi — learn the barbarians' superior techniques to control the barbarians. Feng Guifen formalized the framework in 1861: Western mathematics, Western science, Western weapons could be adopted without abandoning the Confucian foundation. Zhang Zhidong's later slogan was the most quoted: Chinese learning for essence, Western learning for practical application.
Li Hongzhang was the most powerful of them. In an 1874 memorial to the throne, he wrote that China faced san qian nian wei you zhi da bianju — a great transformation unprecedented in three thousand years. The line became the most cited phrase of the late nineteenth century, and it captured something essential about the movement: the claim was not that change was possible. The claim was that the historical situation demanded it on a scale without precedent, and that the dynasty was capable of meeting the demand.
The artifacts followed. The Jiangnan Arsenal opened in Shanghai in 1865, with a translation bureau attached that produced Chinese editions of Western technical texts. The Fuzhou Navy Yard launched its first steamship in 1869. The China Merchants' Steam Navigation Company began operating in 1872 as the first major Chinese-owned modern shipping enterprise. Telegraph lines reached Tianjin in 1879, then Shanghai, then Canton. The Kaiping coal mines opened in 1881. Cotton mills, match factories, paper mills, mining concessions. By the late 1880s the Beiyang Fleet was a deployable naval force of sixty-five ships, fifty thousand tons, with the armored cruisers Dingyuan and Zhenyuan as its centerpieces, the heaviest warships in East Asia.
1865 — 1895
most modern in Asia, 1888
February 1895
None of this was performative. The arsenals produced rifles. The fleet had real armor and real guns. The translation bureau really did translate Adam Smith and Henry Wheaton and Alexis de Tocqueville, and the texts really were read by examination candidates whose career trajectories required them to demonstrate familiarity with foreign learning. The Self-Strengthening figures were not naive. They were senior officials with detailed empirical knowledge of how badly the existing system was failing, and their argument was that the system could be augmented from within.
What they could not do was reform the fiscal apparatus that the augmented system was sitting on top of. The land tax had been frozen at Kangxi-era levels since 1713. The salt monopoly was corrupted. The grain tribute system had been crippled by the Yellow River course change of 1851–1855 and the deterioration of the Grand Canal. The maritime customs revenue, which was the single most modern and reliable source of state income, was administered by a foreigner (Robert Hart, the Inspector-General of the Imperial Maritime Customs) whom the dynasty did not fully control. The arsenals, the fleet, the railways, the mines were funded out of the personal patronage networks of provincial governors, primarily Li Hongzhang. When Li was politically weakened, the enterprises he had built lost their protector. When the dynasty needed to fund a real war, it discovered that the Self-Strengthening apparatus was a thin industrial layer over an unreformed extractive base.
Then Japan happened.
The Sino-Japanese War of 1894–95 was a conflict the Qing was supposed to win. Western military observers had reasoned through the comparative balance of forces and concluded that the Beiyang Fleet would defeat the Japanese navy. The Japanese army was small. Japan was a recent modernizer, its Meiji Restoration only twenty-six years old. China was the older, larger, deeper civilization. The numbers favored China.
The numbers were wrong. At the Battle of the Yalu River in September 1894, the Beiyang Fleet engaged the Imperial Japanese Navy and lost five major warships against zero Japanese sunk. Five months later, at Weihaiwei, the remnant of the fleet was either captured or scuttled. Admiral Ding Ruchang took poison rather than surrender. The Treaty of Shimonoseki, signed in April 1895, ceded Taiwan to Japan, recognized Korean independence as a precursor to Japanese annexation, and required an indemnity of two hundred million taels (roughly three years of Qing fiscal revenue).
The defeat was not just a military defeat. It was a public, comprehensively documented, publicly explainable falsification of the Self-Strengthening argument. The fleet that had been the visible proof of the claim no longer existed, and the country that had destroyed it was a country the Qing had treated for two and a half centuries as a tributary periphery.
The Hundred Days' Reform of 1898 was the next attempt to rescue the rising narrative: a more radical version, led by Kang Youwei and Liang Qichao, that argued the Self-Strengthening figures had not gone far enough. The Empress Dowager's coup ended it in three months. The Boxer Rebellion of 1900 was the conservative counter-attempt, which delivered Beijing to a foreign army and pledged tariff revenue for forty years. The constitutional reforms of 1908 were the third try, which produced provincial assemblies whose first elected majorities turned out to be hostile to the Manchu court. The dynasty fell in October 1911.
Seventy years from the Opium War to the abdication. Twenty-three from Sino-Japanese defeat to terminal collapse. The structural decay had been visible for at least a century. The legitimacy collapse came faster than the structure decayed.
The Faster Fall
The comparison case is the Late Tang.
The An Lushan rebellion of 755 had broken the Tang's military-fiscal equilibrium in much the same way the Opium War broke the Qing's. The frontier-command structure that had produced An Lushan was never reformed; the provincial military governors who emerged from the war became permanent. The fubing militia system was replaced by professional armies whose loyalty was conditional and whose paymasters were regional. The fiscal apparatus collapsed in the north and rebuilt itself, slowly, around the Huai-Yangtze grain corridor and a salt monopoly that had not existed before the rebellion. The Tang did not recover. It also did not fall. It persisted, smaller and thinner, until 907.
One hundred and fifty-two years from onset to terminal collapse. Twice the Qing's window. Same system type, same structural exhaustion, same external and internal pressures rearranged into a different chronology.
The variable that distinguishes them is what the dynasty said about itself.
The Late Tang produced no rising narrative. Du Fu's "Spring Prospect," written in 757 in the immediate aftermath of An Lushan's sack of Chang'an, opens: guo po shan he zai — the state is shattered, the mountains and rivers remain. The line became the canonical articulation of an entire literary register. Bai Juyi's "Song of Everlasting Sorrow" turned imperial collapse into elegy. The poets of the late ninth century (Li Shangyin, Du Mu, Wen Tingyun) wrote about ruined gardens and disappeared courts and the impossibility of recovering what had been lost. Nobody promised that the dynasty would rise again. It was simply allowed to be smaller.
The Late Qing produced thirty years of public rising-narrative construction, followed by ten years of attempted radical recovery, followed by the rapid transfer of legitimacy from the dynasty to "the Chinese nation" in the writings of figures like Liang Qichao, Yan Fu, and the revolutionary press of the 1900s. By 1905, when the dynasty abolished the millennium-old examination system in a desperate attempt at reform, the institutional link between the educated elite and the throne had been severed. The dynasty fell six years later.
The mechanism is not subtle, once it is named. A public, dated, falsifiable commitment to rise is a deposit of legitimacy against future performance. While the system is rising, the deposit pays interest. When the system fails to deliver, the deposit is forfeit at a rate that has no equivalent in static decline. The forfeit propagates through the named principals first (the figures who have publicly carried the commitment) and from there into the legitimacy of the institution they served.
The Qing court, after Weihaiwei, was running a system that had publicly bet on becoming strong again. It had lost the bet. The legitimacy that had held the system together for two and a half centuries was not gone, but it was now contingent on a specific kind of recovery that the dynasty had no remaining capacity to deliver. The Hundred Days, the Boxers, the constitutional reforms: each of these was a renewed bet, and each of them lost faster than the last. By 1911, the dynasty had run out of plausible bets, and the legitimacy ran out with them.
The Tang, by contrast, had never bet anything. It accepted simplification, allowed the provincial military governors to consolidate, allowed the salt monopoly to fund a smaller state, allowed the literary corpus to mourn what had been lost without committing the dynasty to its recovery. The result was that the dynasty's residual legitimacy was never tested against a falsifiable claim. It could not be forfeited because nothing had been deposited.
This is the load-bearing variable. It is not the rate of structural decay that determines the speed of legitimacy collapse. It is the relationship between the system's public commitments and its measurable performance.
A system that quietly declines can persist almost indefinitely. The Mughal Empire entered its long collapse in 1707 (Aurangzeb's death, after which no emperor would again project centralized authority across the subcontinent) and was not formally dissolved until 1858 (when the British Crown abolished the title in the aftermath of the Sepoy Mutiny); a hundred and fifty years of progressive simplification under successive emperors who, by and large, did not go on the record claiming the empire would be made strong again. The Holy Roman Empire of the late seventeenth and eighteenth centuries was already what Voltaire said it was (neither holy, nor Roman, nor an empire) but its constituent princes did not commit to a rising narrative, and the structure persisted until Napoleon dissolved it from outside in 1806. The Ottoman case is more complicated and instructive: it had its own Self-Strengthening period under the Tanzimat reforms, then under Abdülhamid II, then under the Young Turks, and each cycle of public commitment-and-failure compressed the timeline until the empire dissolved in eight years between the Italian war of 1911 and the Treaty of Sèvres in 1920.
The pattern is consistent. The systems that survived their structural decay longest were the ones that did not promise. The systems that fell fastest were the ones that promised most.
The Ottoman case deserves a closer look because it was the Qing's most direct contemporary, running its own Self-Strengthening cycles in the same century. The Tanzimat reforms of 1839 to 1876 were the first cycle, with Mustafa Reşid Pasha and his successors articulating a public claim that the empire could be modernized without abandoning its Islamic foundations: the same essence-application division Zhang Zhidong would formalize forty years later. The reforms produced visible artifacts: a new legal code, secular schools, modernized army units, the first Ottoman constitution in 1876. Then Abdülhamid II suspended the constitution in 1878 and ran the second cycle, of pan-Islamic legitimation and selective modernization, for thirty years. The Young Turk revolution of 1908 was the third cycle, restoring the constitution and committing the empire to constitutional renewal. Each cycle made the next failure more catastrophic. By 1911 the empire had been losing structurally for two centuries, but it was the rapid succession of failed public commitments that compressed the final collapse into eight years between the Italian invasion of Libya and the Treaty of Sèvres. The empire that had been the Sick Man of Europe for a century fell in less than a decade once the named bearers of the rising narrative had run out of credibility.
The Eurozone case is more recent and more instructive precisely because the underlying institutions did not fall. Between roughly 2010 and 2015, a small set of named principals (Mario Draghi, Wolfgang Schäuble, Jean-Claude Juncker, Christine Lagarde at the IMF) committed publicly to a specific recovery narrative for the southern periphery. The bailouts would work. The structural reforms would deliver convergence. The single currency would emerge stronger. The artifacts followed: ESM, OMT, banking union, fiscal compact. Some of the structural decay was real and persistent: Greek output never recovered to its 2008 peak, Italian productivity has not reconverged to the German trajectory, peripheral youth unemployment scarred a generation. The institutions did not collapse. But the named bearers of the recovery narrative absorbed disproportionate political costs at the national level: Papandreou fell, Berlusconi fell, Zapatero fell, the German governing coalition that authorized the first bailout never recovered its electoral position. The Eurozone case is what the trap looks like when the underlying institutions are robust enough to absorb the legitimacy collapse without falling. The principals fell. The system did not. That is the better case, not the worse one.
The Same Mechanism, Now
Which brings us, finally, to the present.
The contemporary AI infrastructure buildout (the hyperscaler capex cycle, the cross-channel investments tying Microsoft to OpenAI, Alphabet and Amazon to Anthropic, NVIDIA to TSMC, and all of them to a finite set of utility-scale power projects) is, on the metrics that matter for chokepoint analysis, a structurally fragile system. This is not a controversial claim. It is the conclusion of any serious analytical pass through the current disclosures, and it is consistent with the observation that the four largest issuers in the cycle sort cleanly into four distinct fragility profiles when their filings are forced through the same comparative schema.
I have been working through this with a small set of analytical tools, and the results are worth stating plainly. Oracle's case is externally mediated: vehicle-heavy, lessor-supported, with an exposure profile dominated by signed-but-not-commenced lease commitments and customer-prepayment structures around the OpenAI and Stargate corridor. Microsoft's case is on-balance equity-platform, with the OpenAI exposure sitting inside a much larger and more diversified balance sheet. Amazon's case is the resilience benchmark of the four: self-funded platform expansion, no Oracle-style vehicle layer, with an exposure to Anthropic that is real but more diffuse. Google's case is the new one and the most analytically interesting. It is vertically integrated chokepoint fragility: an internal frontier lab, custom chips, cloud platform, model deployment, and major external-lab investment all routed through one stack. That stack also carries a $40 billion commitment to Anthropic and a reported $200 billion Anthropic-to-Google-Cloud counter-commitment, producing a cross-channel concentration that has no clean precedent in the previous cases.
The numerical commensurability with the historical anchors is striking. The index is a concentration-weighted measure of critical-throughput dependency, built from revenue-source concentration, single-counterparty exposure, and chokepoint density across a system's core fiscal or operational pipeline. On that index, Oracle and Google both score in the same band as Western Rome's late-imperial dependence on the African grain provisioning system, and as the post-An Lushan Tang's dependence on the Huai-Yangtze grain corridor and salt-monopoly chokepoint. Microsoft sits in the medium band. Amazon sits below it.
vs. Western Rome 2.80
vs. Late Tang 2.68
per Q1 2026 disclosures
That alone would not be especially alarming. Chokepoints exist. Civilizations route critical throughput through small numbers of nodes all the time, and most of them survive most of the time. Eastern Rome had Egypt-grain dependence comparable to the Western Empire's Africa-grain dependence and persisted for another thousand years. Early Ming China rebuilt the canal system and re-centralized fiscal extraction with no sign of imminent collapse. Concentration is not destiny.
What changes the picture is the rising narrative under active construction by a small number of named principals.
Larry Fink, the chief executive of BlackRock, told the Milken Institute in early May 2026 that a trillion-dollar new asset class (futures on compute, modeled on oil futures) would emerge to monetize guaranteed future AI processing capacity, and that fears of a bubble in the underlying buildout were unwarranted. The framing did four things at once. It acknowledged a structural shortage in compute, chips, memory, and electricity. It proposed a financial-engineering response. It dismissed skepticism. It reframed the shortage as opportunity. The structural cousin to this rhetorical move is Li Hongzhang's 1874 memorial: the same combination of urgent diagnosis, programmatic response, and confidence in the system's capacity to meet the moment.
David Sacks, who serves in the current administration as the AI and Crypto Czar, posted on X in early May that AI capex was already responsible for seventy-five percent of US GDP growth in the first quarter and that, in his words, "stopping progress in AI would be equivalent to halting the U.S. economy." This is not a casual claim. It is a public, dated, falsifiable assertion attached to a named principal whose role in government requires him to defend it. If the AI-as-GDP-tailwind reading does not hold for two consecutive quarters in 2026 or 2027, it will not be a quiet revision. It will be an event with a name.
Jensen Huang, at GTC 2026 in March, revised the AI compute demand projection to one trillion dollars through 2027 and is now actively framing AI compute, in conversations with national governments, as Sovereign AI: a foundational state resource on the order of energy or water. Sam Altman has been, for years, the most articulate advocate of an abundance-and-AGI rising narrative whose institutional artifact is OpenAI itself. Marc Andreessen's "techno-optimist" framing has, since 2023, performed a similar function in the venture-capital register. These are not equivalent statements. They have different scopes, different time horizons, and different falsifiability profiles. But each of them is the same structural object: a public, named, time-bounded claim about the trajectory of a structurally fragile system.
Each of them is collateral.
The Named Bearers
The Late Qing's Self-Strengthening figures were not a class. They were a small set of named principals whose roles required them to combine technical-administrative claims with rising-narrative legitimacy. Wei Yuan, Feng Guifen, Li Hongzhang, Zhang Zhidong, Kang Youwei, Yan Fu, Liang Qichao: fewer than a dozen figures who carried the weight of the public argument across the thirty-year arc. When the argument failed, they failed individually, in sequence, ahead of the system they had been arguing for. Li Hongzhang's reputation never recovered from Shimonoseki. Kang and Liang spent the rest of their lives in exile. The dynasty fell after the principals had already lost their public credibility.
The contemporary system has its equivalent. The set is small, identifiable, and observable in real time.
Sacks is the policy-facing node. His role as Czar, his X presence, his public defenses of the AI capex thesis, and his willingness to make specific GDP claims position him as the most legitimacy-exposed figure in the administration's AI-policy apparatus. If the rising narrative he is publicly committed to fails at the macroeconomic level, the political cost lands on him first.
Fink is the capital-allocator node. His role as BlackRock CEO, his platform at Milken and Davos, and the specificity of the futures-on-compute commitment make him the most legitimacy-exposed figure in the financial-architectural register. If the asset class he has publicly named does not materially exist by the end of 2028, the cost lands on him before it lands on the underlying buildout.
Huang is the technical-claim node, most exposed on the Sovereign-AI framing and the trillion-dollar demand projection. Altman is the AGI-utopia node, most exposed on OpenAI's own structural commitments and on the public claims the company has made about its trajectory. Each of them has constructed a public position that is observable and dated, and each of those positions is currently sitting as collateral against a buildout whose chokepoint structure scores in the same fragility band as systems that have historically collapsed.
The Late Qing did not fall because Li Hongzhang was wrong. It fell because the public commitment he carried was load-bearing, and when it failed visibly, the legitimacy of the entire dynasty became contingent on a recovery the dynasty could not deliver.
The argument is not that any of the contemporary figures will fail. Predictions of that form are unfalsifiable in the wrong direction; nobody is in a position to time the cycle, and most who have tried have been wrong. The argument is narrower and more useful.
It is that the conditions for a Late-Qing-style legitimacy collapse are present in the system. High concentration of critical throughput, measurable on a concentration-weighted index commensurable across two and a half millennia. Active rising-narrative construction by named principals, observable in real time on the public record. Cross-channel exposure at identifiable nodes, mappable through the disclosed financial commitments. The combination of these three is the signature.
When a future failure event arrives (and the named commitments are dense enough that some failure event will arrive somewhere), the question is whether the legitimacy collapse propagates through the named principals first, ahead of the underlying structural collapse. That is the trap. That is what made the Qing fall in seventy-two years instead of one hundred and fifty-two.
What's Different, What Isn't
It is worth being honest about the ways the analogy doesn't hold.
Modern hyperscaler-class issuers have capacities the Qing dynasty lacked. They have transparent capital markets, audited disclosures, multiple regulators, and a global investor base that prices their commitments in real time. They have the option of failing partially without failing terminally; Oracle can refinance, Microsoft can reallocate capex, Alphabet can renegotiate. The contemporary AI buildout is not the Beiyang Fleet. It is a financial-engineering exercise sitting inside a regulatory and capital-markets architecture that has no equivalent in late-imperial China, and the named principals are not running a 270-year-old dynasty whose legitimacy depends on a Mandate of Heaven that can be lost. The Qing fell because there was no exit ramp short of the dynasty's end. The contemporary system has many exit ramps.
This is real. The analogy is not that the AI buildout will collapse the way the Qing did. It almost certainly won't. The analogy is narrower, and narrower is what makes it useful.
The mechanism (that public commitments to rise function as legitimacy collateral, and that visible failure of those commitments delegitimizes the speakers ahead of the underlying system) is robust across cultures, eras, and technological substrates. It applies to the Tanzimat reformers in the Ottoman Empire. It applies to the Russian Whites after 1917. It applies to Gorbachev after the Soviet collapse. It applies to the technocrats of the Eurozone after 2010. It applies, in a smaller register, to corporate executives whose public guidance is missed for two consecutive quarters and whose continued tenure becomes a market question.
What is specifically structural about the contemporary moment is the density of concurrent public commitments, the named principals carrying them, and the chokepoint concentration on which the commitments rest. None of these factors is new. The combination of all three at the present scale is a thing the historical record has limited precedent for. The closest historical analog is the Self-Strengthening Movement, and the Self-Strengthening Movement ended badly.
The honest version of the forecast is that some commitment, somewhere in the named set, will fail visibly within the next two to three years. The named commitments are dated. The exposure is mapped. The failure will not be unprecedented. The question is which one fails first, and how the failure propagates.
The Tang lasted longer because it did not promise. The Qing fell faster because it did. The contemporary system has promised.
◆ End of dispatch
Fieldnotes does not normally make dated predictions. The analytical voice of this publication is structural rather than oracular, and most of what passes for forecasting in financial and political journalism is the unfalsifiable confidence that nothing important will change. I want to be specific about the watch in this case, because the argument above is empty without something time-bounded attached to it.
Four artifacts in the contemporary set are sufficiently specific to be falsifiable.
The first is Oracle's investment-grade credit access through the first half of 2026. The company is executing a forty-five-to-fifty-billion-dollar 2026 financing plan against leverage that is already elevated, escalating capex, and a single-customer corridor through OpenAI and Stargate. As of mid-May 2026 the access is intact: no CDS blowout, no ratings breach, no syndication failure. The window closes on June 30. If it closes intact, the prediction holds. If it closes with stress, the analytical pattern has activated on the first artifact.
The second is the AI-as-GDP-tailwind claim, attached publicly to Sacks. If the figure does not hold for two consecutive quarters in 2026 or 2027, the political cost will land on the named principal before it lands on the administration. This is the specific test of whether the legitimacy-collapse mechanism propagates through the named bearer first.
The third is the futures-on-compute commitment, attached publicly to Fink. If the asset class does not materially exist by the end of 2028, the BlackRock CEO bears the legitimacy cost ahead of the underlying buildout. This is a longer-dated test, but the specificity of the commitment makes it cleaner than most.
The fourth is the cross-channel concentration around Anthropic. The frontier lab is simultaneously a reported forty-billion-dollar equity beneficiary of Alphabet, a reported thirty-three-billion-dollar equity beneficiary of Amazon, the source of more than forty percent of Google Cloud's reported revenue backlog, the holder of a hundred-billion-dollar commitment to AWS, and a primary customer for Broadcom's custom-silicon roadmap. Any single contractual failure at this node travels through three hyperscaler balance sheets, two cloud businesses, and one chip supply chain at the same time. There is no equivalent concentration point in the historical record at this scale of cross-channel exposure.
These are the specific watches. They are not predictions of collapse. They are predictions about where the legitimacy stress will appear first if the structural fragility activates. The Late Qing pattern is that the named bearers are delegitimized ahead of the system. If that pattern holds, we will see it in the specific public positions of the specific named principals before we see it in the underlying balance sheets.
The Tang lasted longer because it did not promise. The Qing fell faster because it did. The contemporary system has promised, and the promises are dated, and the named principals are observable. The watch is live.
What we are watching, in the months between now and the next visible failure event, is whether the analytical pattern holds across two and a half millennia and an entirely different technological substrate. The hypothesis is falsifiable. The watch is named. The artifacts are public.
That is the warning from Self-Strengthening. It is a hundred and thirty years old. It is also the most specific thing you'll read this week.