Evan Osnos on China’s Technological Ascent, Chinamaxxing, and Humanism versus Scientism

Evan Osnos returns to Kaiser Kuo‘s show to discuss The Future, Made in China, his New Yorker essay on a tech-focused return to Beijing, where he lived from 2005 to 2013. His argument is a reframing: the real contest is not democracy against authoritarianism but humanism against scientism — a fault line running through both countries rather than between them — and the two systems are converging on concentrated technological power by opposite routes.

Key ideas

1. The contest is humanism versus scientism, not democracy versus authoritarianism. Osnos arrived at this in the reporting rather than bringing it to the material: the language he heard in Zhongguancun resembled what he hears in Silicon Valley. The question both societies face is ‘how much do we put faith in our machines and how much do we put faith in our minds’. See Scientism.
2. Awe and Potemkin are both real, and insisting on one is the mistake. The Xiaomi Hyperfactory is an ‘industrial Epcot centre’ staged for visitors and it produces a car every 76 seconds. Osnos notes that Rush Doshi reads China’s message as ‘we are inevitable, don’t worry’ — and observes that this is also Silicon Valley’s message. Ford’s Jim Farley called the visit the most humbling thing he had seen; the discipline is holding the performance and the progress together.
3. America’s abdication is part of China’s position, and the counterfactual is genuinely hard. The China Initiative scattered scientists, research funding was gutted, and global approval fell some forty points. Osnos nevertheless declines to condemn the Sullivan-doctrine export controls: we cannot know where the AI race would sit without them. Kuo presses him on this and they part company — Kuo’s view is that constraint acted as catalyst.
4. Convergence: two routes to the same dystopia. China’s is legible — surveillance, a facial-positivity score to two decimal places. America’s is insurers using AI to deny treatment, surveillance pricing, and defence contractors with Tolkien names fusing security, policing and immigration. Osnos refuses to flatten the difference — a one-party state is not a messy democracy — while arguing that American power has concentrated ‘beyond any recognisable precedent’.
5. Knowability is now the binding constraint on China analysis. Numbers are less reliable and contacts vanish into investigations, which forces analysts to acknowledge what they cannot see. Hence Osnos’s refusal of declarative verdicts: ‘Schrödinger’s China’, held as both imminent collapse and unstoppable juggernaut, is a symptom of watchers reading themselves rather than the country.

Content

Walking Beijing again

Osnos returns every couple of years. What struck him this time was technological saturation, and a city that is quieter on more than one level — literally, because the vehicles are electric. He reaches for the peasant da Vincis of Age of Ambition, the villagers who built wooden submarines and flying machines, as the emblem of an instinct that ‘comes and goes and can get choked off’. What felt different now is that the same inventive drive is operating at a systemic level, which is both astonishing and unnerving: ‘a piece of technology reflects the values of the person at the controls’.

A friend’s line — that Beijing now out-Singapores Singapore — carries his ambivalence. Cleaner, more organised, more systematised, and slightly sterile with it. He is careful not to romanticise: both men note the reasons one should not nostalgise poverty or grime. But he holds that a certain messiness is needed ‘to create human combinations and new ideas and the unpredictable quality that we call joy’, and describes the loss of the hutong cacophony as phantom limb pain.

Hype and substance at the Xiaomi Hyperfactory

Osnos is candid that his factory visit was partly staged — a tech-tourist attraction with a golden-ticket lottery. His method against that is refusal of the single verdict. Declaring China the technological superpower of the future and dismissing it as a Potemkin village are the same error in opposite directions, and having ten thousand words is what lets him keep the twists.

His historical check is Elon Musk laughing at BYD around 2011 — ‘have you seen their car?’ — against a boxy sedan that has since become the industry’s counter-example. The lesson he draws is about the slow, steady boring of holes, and against reading any moment as the country’s fixed state: ‘the one thing that never changes is that it changes a lot and very, very fast’.

The birdcage and its limits

Pressed for the takeaway, Osnos points to the birdcage economy of Deng Xiaoping and his peers — enough ventilation for the bird to thrive without losing control of it — and calls the balance fragile. His live example is Manus, the AI company: that it existed at all showed China’s capability, and it then hit the limit of what the government would accept and was prevented from leaving. His interest is in the signal that sends to young researchers deciding whether to leave the country before having the idea.

Schrödinger’s China and the limits of knowing

Kuo puts the pendulum to him — collapse-imminent and ten-feet-tall, sometimes inside one institution. Osnos’s answer is that the reading is always refracted through how the observer sees themselves, and that this runs both ways: Beijing’s sense of what is possible depends on how it imagines the West will receive it.

He then makes the analytical point that governs the rest. China has become harder to access, its numbers less reliable, and contacts can vanish into an investigation with no public account — so analysts must acknowledge the limits of what they can see, and may be shocked at any time. Against that he sets the asymmetry with the United States, whose political inferno is on flamboyant display. Where evidence does exist, he uses it: Scott Rozelle’s survey work on the shift from optimism to pessimism, read alongside lying flat and falling marriage rates.

Chinamaxxing and the two American conversations

Kuo names the split: the threat narrative of munitions stockpiles and lithography on one side, and Gen Z ‘soaking in infrastructure porn’ — maglevs, drone shows, drone-delivered lattes — on the other, with Pew showing a marked positive shift among the young. Osnos attributes much of China’s gain in these measures to the decline in American standing rather than to persuasion.

He grants the phrase and then turns it: an understanding of human relations refracted entirely through the internet would be impoverished, and the same holds for China. If you can only push on the doors that are unlocked, you are not getting the full picture. He is alert to the unfashionability of saying so to people who regard scrolling as doing their own research, and Kuo offers the sharper formulation from the Red Note moment — not a Potemkin village, but only the nicer parts of the real village.

This is where Osnos mounts an unembarrassed defence of journalism: the trade is inherently adversarial, it will always grate against the story a party wants told, and closing the aperture further would not buy trust. ‘You have to make the place knowable.’

The passage turns on two borrowed images about the limits of outside description. Jiayang Fan’s, on reading Western reporting about China: like seeing a familiar part of your own body through an X-ray — anatomically precise, revealing what the naked eye misses, and not the hand you know, lacking the sinew and flesh. And Peter Hessler’s outdoor exercise machines, built as though an engineer had been shown a brochure of American equipment and then had it taken away. Osnos applies the same to Wang Huning’s America Against America: the bone structure caught, the ligaments missed.

Thank Trump, and the American self-inflicted wound

Kuo assembles the ledger — the Renmin University report titled Thank Trump, the China Initiative scattering scientists, gutted research funding — and asks how much of China’s position is its own achievement and how much is American abdication. Osnos says the distinction matters, because countries are never assessed in a vacuum: ‘don’t compare me to the almighty, compare me to the alternative’.

His worry is that Americans experience disinvestment in basic science as an ordinary political cycle rather than as a transfer of advantage, and he poses the test case of what it will mean if the cure for cancer comes from China. The mood abroad he renders through a German interlocutor on NATO and Greenland: ‘we’re not angry, we’re just moving on’ — which he calls the most withering thing anyone who has been through a breakup can hear.

On export controls he declines to join Kuo. The counterfactual is unavailable; there was a genuine belief that these technologies were too powerful for the open market; and he judges the second Trump administration’s reversal a symptom of absent policymaking rather than a deliberate correction. He concedes the durable point — China’s persistent capacity to make more out of less was not extinguished.

Mankind 2000, oligarchy, and convergence

The essay’s philosophical heart is Mankind 2000, the European survivors of fascism wary of both communist determinism and American hubris, who warned that man could be absorbed by his machine and insisted the future belong to everyone rather than to small groups. Osnos reframes the rivalry as a race between two oligarchies: technomandarins bending technology toward obedience, and an accelerationist Silicon Valley waving off constraint.

He credits Sam Altman with the sharpest statement of the problem — that if he were not in on it, he would be asking why this group of people gets to make the decisions for all of us — and argues the question transcends borders. It is what he thinks is really being expressed when graduates boo commencement speakers who mention AI, and in the resistance to data centres: agency, control, a hand in making one’s own future. Dan Wang’s phrase for the Chinese analogue is serene discontent.

Kuo tests the symmetry, noting the techno-optimism that still pervades China — his framing is that the West is in a Black Mirror phase while China remains in a Star Trek one — and grounds it in a technocratic system where the ladder of social success runs through STEM. Osnos partly agrees, and adds that you cannot boo a commencement speaker in China without consequences, so the absence of visible dissent is weak evidence.

On convergence Osnos accepts Kuo’s reading while refusing to collapse the difference. The concentration of power in very few American hands has reached the point where entities are ‘becoming nearly sovereign’ — his image is the 2025 inauguration stage, the world’s richest people in strange political communion with the leader of the most powerful military on earth. But a country where the public will can still reach the front page is not the same as one where it cannot.

Self-strengthening redux

Kuo presses his own thesis: that America is running the nineteenth-century self-strengthening playbook and botching it. Industrial policy, state-directed AI investment, export controls, a president taking government stakes in chipmakers — Chinese means adopted in a rush, without a national conversation, while insisting the American essence remains intact. He invokes Li Hongzhang on learning the superior skills of the barbarians in order to control them, and Joseph Levenson on the widening gap between what is one’s own and what is true.

Osnos offers the other reading: that the United States has shown it can mount an all-hands-on-deck national effort, and that China may underestimate what happens when the dragon wakes. He then shifts the frame to a second Sputnik moment — not America waking to a Chinese challenge, but Americans waking to overweening control by a few technology powerhouses. His evidence is Jill Lepore’s forthcoming The Rise and Fall of the Artificial State, whose own Sputnik moment was ChatGPT, and which argues that technology will run up against fundamental American beliefs about liberty and control.

Asked who is left to defend everyone’s claim on the future, Osnos answers plainly that he does not see who it is. The parties are in a primordial-soup period; Europe has concluded it must look out for itself, which he does not think redounds to China’s benefit. He quotes Robert Kagan on the succession of world orders — Mongol, Islamic, Egyptian, American — and insists that coming to terms with what follows does not mean abandoning the values the United States has aspired to and never fully achieved. The Gramsci tag arrives, the time of monsters; Osnos notes such periods do not last long. He closes on the self-aware note that predictions are a mug’s game in a piece ostensibly about the future.

Related

• Evan Osnos — guest
• Kaiser Kuo — host
• Scientism — the concept this episode extends from epistemology into geopolitics
• Kaiser Kuo, Eric Olander, and Andrew Polk on China's 15th Five-Year Plan, ASEAN Realignment, and the Collapse of the Trade Truce — the same host on China’s planning apparatus
• China's Vanishing Jobs Target, Flexible Work, and Gen Z's Move to Smaller Cities — the Manus episode and Chinese labour-market pessimism from another angle