Roastidio.us User Picks KjQKkQVg

异乡的乡音 | JustGoIdea

Wed, 03 Jun 2026 14:58:52 +0000

异乡的乡音

02 Jun, 2026

坐在街边小店里喝饮料，这几日的夜晚微风习习，甚是凉爽。旁边是一家中餐馆，不知什么时候添了点唱机。断断续续的歌声随风飘来，并不那么真切。可人在异国他乡，听见这样的声音，心里还是会轻轻一动。

也正是这样的声音，让我忽然想起十几年前在 CHCH 的夜晚。那时受邀到一位老华人家中聚餐。饭后，我们一行人去了花园里的木屋，打开点唱机选歌。让我意外的是，曲库竟然还算新，2010 年以前的许多流行歌都有。大家唱着、跳着、觥筹交错，气氛很热闹。几位上了年纪的大叔兴致起来，还拉着我们几个十几二十岁的年轻人一起对唱京剧样板戏的选段。

那些唱段，我当然听过，但毕竟离我们这一代已经有些远了，并不怎么会唱。可在那样的氛围里，也就跟着屏幕上的歌词唱了起来。唱得未必准确，也未必懂得其中所有意味，却觉得那一刻很自然。

现在想想，那些唱段对几位长辈来说，恐怕并不只是「歌」。那是他们年轻时听着长大的声音，是时代留在身体里的记忆。虽然人无法完全选择自己伴随什么声音长大，但那些旋律、腔调、唱词，夹杂着复杂的历史，也夹杂着青春、酒桌、朋友、热闹与某种已经远去的生活经验。对年轻人来说，它可能只是历史的残响；对他们来说，却是生命的一部分。

那一晚让我感到奇妙的是：在异国他乡，一群原本素不相识的人，只因为同文同种，便可以坐在一起吃饭、喝酒、唱歌、说笑。我们未必真的了解彼此，年龄不同，经历不同，观念也可能不同；可一旦饭菜端上桌，熟悉的语言聊起来，老歌唱起来，许多隔阂便暂时松动了。

在故土，「同文同种」常常不算什么。人们还会因为地域、阶层、年代、观念而彼此区分。可是到了远方，语言、口音、饭菜、歌声，突然都变成了一种暗号。不必解释太多，就能让人感到：原来你也从那里来。

这种亲近很真实，像是在漂泊之中临时形成的一个小共同体。大家短暂地把孤独放下，把异乡感放下，在一首歌、一杯酒、一句玩笑里，互相确认了一下彼此的来处。

所以乡音最动人的地方，也许并不在于它多么优美，而在于它证明：人曾经从某处来，也曾经和一些人共享过相似的声音、语言和记忆。

异乡的风一吹，那些声音就又回来了。它们不完整，断断续续，甚至未必悦耳，可正因为如此，才更像真实的故乡。故乡并不会完整地归来，它只是偶尔藏在一阵夜风里，一首老歌里，一桌饭菜里，或者一群素不相识却忽然可以欢聚的人之间。

Previous |

#LisbonDiary

Why China got rich, and India didn't - David Oks

Wed, 03 Jun 2026 01:55:37 +0000

Here is a question that I think about a lot.

In the year 1950, much as today, the two largest countries in the world by population were China and India. China was a good deal larger at the time, holding 22 percent of the world’s population to India’s 15 percent; but really the two were in a very similar position. Both of them were giant countries that had assumed their current state—India as the independent Republic of India, China as the People’s Republic of China—in the preceding three years. Both of them were among the very poorest places on earth. And both of them were about to spend decades trying, by very different means, to make themselves rich.

For China, that experience was one long nightmare. China had already been wrecked by a prolonged civil war and by a brutal Japanese invasion in the decades prior, the whole experience killing tens of millions of people. The civil war ended in 1949, with a Communist victory; but what came next was no less catastrophic. The Communists’ leader, Mao Zedong, immediately embarked on campaigns of vengeance against enemies of all stripes, murdering well over a million people in the process; he then launched on an ill-fated agricultural modernization campaign, the Great Leap Forward, that produced the largest famine in history, killing somewhere between 30 and 45 million people; and then a frenzied period of ideological radicalization, the Cultural Revolution, that suspended national life for a decade and killed another 1.6 million. By the time that Mao died in 1976, China was internationally isolated, economically stagnant, and still desperately poor.

For India, the experience was a much gentler one. India had been a colony of the British, and it was able to achieve independence without taking up arms. British institutions like the Indian Civil Service—the colonial bureaucracy, rechristened as the Indian Administrative Service—carried over into the new Indian state. There was a bout of extreme violence in the late 1940s, as the country was partitioned into Hindu-majority India and Muslim-majority Pakistan: but even that was incomparable to what China experienced. And after that episode, India enjoyed long decades of peace, stability, and democratic rule. It was led by a broad-minded secularist named Jawaharlal Nehru, who had been educated at the finest British institutions and governed in the name of science, reason, and social progress; and throughout its entire post-independence period India maintained open elections, an independent judiciary, and a free press. It never experienced anything like the Great Leap Forward or the Cultural Revolution.

I suspect that if I’d been around in the year 1950, it would have been obvious to me that India would succeed and China would not. I would have made the same bet in 1960, when China was starving tens of millions of its own people while exporting grain abroad; and I would have done it again in 1970, during the insanity of the Cultural Revolution. Nor would I have been alone. As late as 1985, prominent economists were writing articles in the New York Times suggesting that “far more than China today, India is an economic miracle waiting to happen.”

But they were wrong.

In the five decades since the death of Mao Zedong, China has grown much faster than its fellow Asian giant. China has become a manufacturing superpower and the single fastest-growing economy of the last 50 years; its per capita GDP, on par with India’s in 1976, is now about 2.5 times higher. And so Chinese people have become much better-off than their Indian counterparts. In 1987, median purchasing-power adjusted income in China was $1.88 per day, compared to $2.94 per day in India. Chinese median wages surpassed Indian ones in the early 2000s; and by 2022, China recorded a median income of $13.36, against $5.54 in India. In the 35 years between 1987 and 2022, Chinese median income rose 611 percent, while Indian median income rose only 88 percent.

So what happened? Why did China get rich, and India didn’t? What explains the Sino-Indian divergence?

Last year, I visited India and got the chance to pose these questions to a few prominent Indians, including several members of the Indian parliament. The most common answer that I heard from them was simply that India reformed later. Both India and China operated under strict state control of the economy for much of the post-1950 period; but while China began liberalizing its economy in 1978, India waited until 1991. And so China simply has a 13-year head start: no wonder it’s grown more than India has.

But that doesn’t explain why China has continued to grow faster than India. Between 2000 and 2022, long after both economies had liberalized—and with China already being considerably wealthier—Chinese growth still significantly outpaced Indian growth. So even decades after liberalization, India still underperformed China. The timing of liberalization can’t explain the divergence.

The same is true of policy more broadly. There are all sorts of ways in which Indian economic policy remains inefficient and distortionary in ways that inhibit growth; but the same is true for China and indeed for practically all countries. I don’t think that policy differences explain why India has so reliably underperformed China even at much lower levels of income.

The same is true, I should add, of explanations that cite “Chinese culture” and “Indian culture.” It’s obvious that China and India have different cultures, and that those cultures lead people to behave in different ways. But that doesn’t explain why China started to outgrow India when it did. In the early twentieth century, long before Mao or Indian independence, India was richer than China, and Indian and Chinese cotton mill workers displayed broadly similar levels of productivity. Whatever cultural advantage that China might enjoy over India, it wasn’t operative a hundred years ago.

So I want to propose my own explanation for why China got rich and India didn’t. The moment of divergence, I think, came not in 1978, or in 1991, but around 1950.

Rapid industrial development requires human capital: workers who are literate enough to be trained, healthy enough to show up, disciplined enough to come in on time, and sufficiently unencumbered by traditional life that they can sell their labor to whoever offers the most for it. Traditional agrarian societies produce almost none of these people: the peasants who made up the bulk of both India and China in 1950 tended to be illiterate, sickly, and restricted in all sorts of ways. For people to be productive workers in modern economies, all of that must be cut away. The more advanced European states spent much of the last millennium doing exactly that. And between 1950 and 1980, China succeeded—frequently through brutal means—in replicating that process: over the course of a few decades, the Chinese state modernized its society at the barrel of a gun. By 1980, as its economy opened to the world, China was a socially modern country that just happened to be extraordinarily poor. It had the human capital for rapid industrialization.

But India never underwent that transformation. Its traditional social order survived independence more or less intact; and the Indian state never managed to develop its people’s human capital as China had. When India finally opened its economy to the world in 1991, its people were simply not prepared for industrial modernity in the way that China’s were. China invested in its people; India did not.

So how did China do it? And why wasn’t India able to do the same?

The Communist road to capitalism

In 1949, after decades of war—either against the invading Japanese or against their Nationalist foes—the Communist Party of China triumphed over all its enemies and achieved complete power over mainland China. The remaining Nationalist forces fled to Taiwan; and Mao Zedong, paramount leader of the Communist Party, announced the formation of the new People’s Republic of China.

Mao had three overriding objectives for the governance of China. The first was the absolute consolidation of Communist power over the country; the second was the reconstruction of social life along Communist lines; and the third was the economic transformation of China, from an impoverished and agrarian society into a wealthy and industrial one.

On the last goal—making China prosperous and powerful—Mao failed entirely. China remained bitterly poor throughout his time in power, and all his interventions in economic policymaking proved to be disastrous. But on the first two goals, eliminating opposition to Communist rule and reshaping Chinese society according to his whims, Mao was remarkably successful: between 1949 and 1976, the Communist Party totally transformed Chinese life.

It’s hard for us today to grasp just how brutal that transformation was. From 1950 onward, every force in Chinese life that might contest the hegemony of the Communist Party was ruthlessly suppressed and destroyed. The landlords and “rich peasants” who had comprised the traditional leadership class of the villages, for example, saw their lands expropriated by the Chinese state in the 1950s; they were forced to sit through public sessions in which the peasants would “speak bitterness” to them, typically culminating in their being beaten to death. Several hundreds of thousands of people were killed in this way. During the same time period, another several hundreds of thousands were killed as “counterrevolutionaries” opposed to Communist power.

But it wasn’t just landlords, rich peasants, and “counterrevolutionaries.” Practically every representative of traditional power in Chinese life was crushed. The hundreds of secret societies and sects that had dotted Chinese life, counting about 13 million members in 1949, were destroyed in the “Withdraw from the Sects” campaign; Confucianism and other pillars of the old order were attacked and suppressed; the hundreds of thousands of small shrines that had structured Chinese folk religion were declared “superstitious” and obliterated; major faiths were brought under state supervision, and at the peak of Communist enthusiasm in the 1960s religion was banned entirely and countless ancient temples destroyed. The famous Jing’an Temple in Shanghai, built in the third century, was turned into a plastics factory.

Even family patriarchs lost much of their authority. Decisions that had once been vested in families and elders—about, say, marriage or the allocation of land—were stripped and transferred either to individuals (in the case of marriage) or the state (in the case of land). In 1950, the Chinese government passed the New Marriage Law, which banned arranged marriage, concubinage, and child betrothal, gave women the right to own property and divorce freely, and allowed women to keep their own names upon marriage. This was a radical departure from the patriarchal order that had governed Chinese marriage for all known history: and while it led to a tremendous amount of conflict, the Chinese state simply crushed opposition—branding elders as “landlords” and encouraging women to “speak bitterness” about their in-laws and grandparents. The mass mobilization of women into the workforce, under the slogan that “women hold up half the sky,” was similarly transformative: tens of millions of women were pulled out of domestic seclusion and into economic life, and thus freed from the control of their families. And so the traditional Chinese kinship unit—not merely a “family” but an autonomous institution governing the lives of its members—was destroyed.

All this meant that between 1949 and 1976, the Chinese state effectively destroyed traditional Chinese society: the social landscape of the old China, with all its complexity and custom, was simplified and smoothed away.

And in its place, the Chinese state forged a new nation along its preferred ideological lines. Economic development always eluded Mao; but human development—mass education and mass health—proved more attainable. Literacy campaigns and mass education helped raise the literacy rate from roughly 20 percent in 1949 to almost 70 percent by 1982. These gains were concentrated among women: Chinese women went from “virtual complete illiteracy” to a literacy rate of about 50 percent during the same period. The progress in health was similarly rapid: child mortality fell by 80 percent between the early 1950s and the late 1970s. Even with all the horrors of Maoist rule—including, it should be remembered, the largest famine in history—between 1949 and 1976 China recorded one of the largest sustained increases in life expectancy of any country in history, rising from about 41 years in 1949 to 61 by 1976. And for the first time in history, Chinese women were meaningfully included in public life: by the late 1970s, China had a female labor force participation rate exceeding that of many rich countries.

And so by the time that Mao died in 1976, Chinese society had been utterly transformed. It was still a deeply poor and largely agrarian country; but it had education and health outcomes far exceeding what you’d expect from a country at its level of income. And it had crushed the traditional social structures that had previously governed every aspect of Chinese life. It was a socially modern country that just happened to be extremely poor: in 1980, China had the same life expectancy as Mexico despite having a per capita GDP 80 percent lower than Mexico’s.

And this meant that by the late 1970s, even before the “reform and opening up” process started, China was perfectly prepared for industrial capitalism. The old constraints—kinship, tenancy, female seclusion—had been swept away; the Chinese workforce was mobile, trainable, and cheap. That mismatch, between China’s level of human development and its level of wealth, was bound to be resolved by rapid economic growth.

When analysts from the World Bank visited China in the early 1980s, they reported that its low-income groups were “far better off in terms of basic needs than their counterparts in most other poor countries”; if China’s “immense wealth of human talent, effort and discipline can be combined with policies that increase the efficiency of resource use,” their report said, “China will be able, within a generation or so, to achieve a substantial increase in the living standards of its people.” And that’s exactly what happened.

India’s failed social modernization

On the surface, India’s trajectory was a much happier one. India didn’t need to fight a war of independence to gain independence: power passed from British hands to Indian ones more or less by negotiation. The partition of India was horrific, killing between half a million and two million people; but it still paled in comparison to the scale of the Chinese Civil War, the Sino-Japanese War, or the Great Leap Forward. And in the decades after independence, India enjoyed stability and democratic governance. It never saw the barbarities that China experienced under Mao.

But India also never underwent the social transformation that China did.

The great power in India in the decades after independence was the Indian National Congress, which had been the central vehicle for winning independence. Between 1947 and 1989, Congress found itself out of power for only three years; its hold on power wasn’t absolute, but it was certainly dominant. But Congress wasn’t really an ideological movement. It had started in the late nineteenth century as a forum for educated Indians seeking moderate reform, and then transformed into a mass movement for independence. It was a big-tent party whose membership amounted to the entire cross-section of Indian life: left-leaning secularists, Hindu traditionalists, upper-caste chauvinists, lower-caste activists, landlords, socialists, and many members who were simply non-ideological and attracted to the charisma of the party’s leaders or the power that membership offered.

And so even though Congress dominated Indian politics for decades, it never offered a coherent program for remaking Indian society. If the Chinese Communists sought out endless antagonisms, Congress avoided them; if the Communists imposed radical change from above, crushing all who stood in their way, Congress was happy to defer to existing interests and hope for social cohesion and gradual progress.

This doesn’t mean that the leaders of the Congress Party didn’t have their own ambitions for transforming India: Nehru, who served as prime minister from independence until his death in 1964, had a strong dislike for the “superstition and deadening custom and tradition” of traditional Indian life, and wanted to solve the “insanitation and illiteracy” and “hunger and poverty” that marked the country. But Congress wasn’t united behind him: Nehru simply didn’t have the power to achieve this in a real way.

In 1950, for example, Nehru and his law minister—the famed lower-caste activist B. R. Ambedkar—introduced the Hindu Code Bill, a sweeping reform of Hindu personal law. (Under the Indian constitution, different religious communities were governed by different systems of personal law.) The bill would have outlawed polygamy, granted women the right to divorce and inherit property, and permitted inter-caste marriage. It was similar in structure to the New Marriage Law that the Chinese government passed the same year, though it stopped short of banning arranged marriages like the New Marriage Law had. But the Chinese government had imposed the New Marriage Law by fiat and steamrolled those who stood in its way. Nehru and Ambedkar, by contrast, found themselves frustrated by a wave of opposition from Hindu traditionalists: even India’s president attacked the bill, suggesting that introducing concepts “foreign to Hindu law” would “cause disruption in every family.”

And so the Hindu Code Bill died in parliament. Ambedkar resigned in disgust; and while Nehru ultimately succeeded in reforming Hindu law, he was forced to agree to enormous concessions. The law that regulated Hindu inheritance, for example, exempted agricultural land from its purview entirely, and so didn’t touch the vast majority of useful property; the law that regulated marriage included a right to divorce, but also a provision for the “restitution of conjugal rights” that gave husbands a court-enforceable right to compel their wives to return home.

But it didn’t really matter what the laws said: enforcement was weak to nonexistent. Divorce and intercaste marriage, whatever their legal status, remained vanishingly rare, because the village and the family enforced the old rules regardless of what the law said; customs that compelled women to renounce their inheritance claims, like the Rajasthani custom of haq tyag (“sacrifice of right”) or the Haryanvi custom of karewa (forced remarriage of widows to control their land rights), remained common. Even the officials charged with enforcing the laws subverted them: the administrators who registered inheritance claims, for example, would routinely pressure daughters to sign away their rights in favor of their brothers.

And that was the general pattern of attempts at social modernization in India in the twentieth century: highly publicized reforms, followed by little change on the ground. In 1961, the Indian government made dowries—payments made by a bride’s family to the groom’s family at marriage—illegal, since the practice entrenched the subordination of women and encouraged domestic abuse. But the law went entirely unenforced; dowries remained as popular as ever, and all the abuses linked to dowries still flourished. (Between 1999 and 2016, dowry-related murders accounted for 40 to 50 percent of all female homicides recorded in India.) The same was true of attempts at land reform. Several Indian states attempted land reform in the 1940s, ‘50s, and ‘60s; but enforcement was lax. Landlords managed to simply evade the rules through legal means, like transferring holdings to relatives, or registering land under fictitious names; or they simply bribed or intimidated government officials. And so very little really happened.

All of this meant that the Indian state was never able to achieve the social modernization that the Chinese state accomplished. The dense web of kinship obligations and customary authority that governed social life remained intact. Caste panchayats still adjudicated disputes; joint families still pooled and redistributed income; and women remained bound by all the strictures of traditional life.

Nor was the Indian state able to accomplish the dramatic improvements in human capital that occurred in China. Just as it was unable to reform social life, it was unable to provide effective services; health outcomes remained dismal. Within a single generation, India’s health outcomes went from comparable to Chinese ones to dramatically worse. The gap between Indian and Chinese life expectancy widened from three years in 1950 to 11 years in 1980. Child mortality told the same story. In 1950, 27 percent of Indian children died before the age of five, compared to 32 percent of Chinese children; by 1980 it was 17 percent in India, against 6.3 percent in China.

And the same with education. Nehru and his successors were keenly interested in technology and the peaks of scientific achievement; but they could never muster similar enthusiasm for mass education. So India established a network of world-class technical institutions, like the Indian Institutes of Technology and the Indian Institutes of Management, while neglecting everything else: even today, India’s elite technical universities receive the majority of government funding for higher education, while educating only 2.6 percent of the university population.

Indian mass education, meanwhile, remained abysmal. In 1990, only 55 percent of Indian children had completed primary school three to five years after the expected completion age, against 87 percent of Chinese children; and even those who did go to school often got little out of it. In 2009, when India participated in PISA—the Programme for International Student Assessment, which ranks students across countries based on test scores in math, science, and reading—it ranked 72nd out of 73 countries. (China ranked at the top.) The Indian government responded to this embarrassing result by never participating in PISA again. This lack of investment in education is visible in Indian literacy rates, which only exceeded China’s 1990 level in the early 2020s.

This lack of progress was particularly brutal for Indian women. There was no great liberation of Indian women as there had been of Chinese women: all the abuses of traditional life, from dowry murders to forced marriages, remained common. Literacy rates for women remained extremely low; in 1981, only 26 percent of Indian women knew how to read. And with women’s lives still determined by the whims of their families, the vast majority of women remained in the home: by the late 2010s, India recorded a female labor force participation rate of about 27 percent, one of the lowest rates in the world—closer to Afghanistan, at 18 percent, than to China, at 61 percent.

(India’s low rate of female labor force participation, in fact, means that India’s labor force remains significantly smaller than China’s, despite India having a larger population: in 2019, in fact, China’s labor force was 45 percent larger than India’s.)

All of this meant that by the time that India liberalized its economy in the early 1990s, it simply didn’t have the pool of high-quality, low-wage labor that China could command. Thanks to its elite technical universities, India did have a relatively small number of highly-educated engineers, who became the backbone for India’s IT services economy; but it didn’t have the workforce for manufacturing-led growth. Its workers were less literate, less healthy, and less productive than what China could offer; they were bound by caste and kinship obligations that made them reluctant to migrate for work or sell their labor freely; and because so few women participated in the labor force, India had a higher dependency ratio than China, with each working Indian supporting far more people who weren’t working.

India did grow after liberalization, of course; and by historical standards its growth was generally quite fast. But it never saw the manufacturing boom and explosive growth that China exhibited. It hadn’t accomplished the prerequisites.

Human capital is what really matters

I think that people make economic development more complicated than it needs to be. It’s true, of course, that certain policies are better than other policies, and that all sorts of things go into successful economic management: disastrous decisions can ruin everything, though (as Mao’s many disastrous decisions might suggest) not permanently.

But when you come down to it, countries are large groups of people. And the most important thing for the success of those groups is simply who’s in them: this is as true for countries as it is for companies, music bands, and sports teams. Human capital is what really matters. Whether the people can read; whether they’re stunted due to undernourishment; whether their families let them work outside the home. Human capital isn’t the only thing that matters, and of course you also need institutions that can harness the country’s human capital. But you need the human capital to be there in the first place.

One of the nice things about countries, though, is that you can change who the people are. You can teach them to read and make sure they have enough to eat; you can make sure they have the freedom to make their own decisions. This isn’t easy, and it takes a long time for it to have an effect—not least because childhood undernutrition and poor schooling have consequences that can’t really be reversed. But you really can change who the people are.

What made China a “miracle waiting to happen” in 1980 was that it had spent decades doing exactly that. By the time it opened its economy to the world, China had hundreds of millions of capable, disciplined, healthy, and literate workers; it had freed them from the constraints of traditional culture, such that market logic could triumph unimpeded by the old order; and because it had failed almost totally in economic development up to that point, it could offer those workers at unbelievably low wages. It’s not hard to see why it grew so rapidly once it opened its economy to the world.

The Indian government never made the catastrophic decisions that the Chinese government did in the 1950s and ‘60s. But it also never made the basic investments that the Chinese government made, and it never managed to challenge the traditional social order with a fraction of the ferocity that China did.

And so in all sorts of metrics—life expectancy, child mortality, literacy, female labor force participation, childhood wasting, childhood stunting, anemic pregnancies, maternal mortality—an enormous gap opened up between China and India well before they liberalized their economies. I think the true moment of divergence was not in 1978, when China began to reform its economic system, but in 1950—when China passed the New Marriage Law, while India failed to pass the Hindu Code Bill. It was then that the direction of future things was written.

India, in other words, never really did the basics. Health and education outcomes in India have improved significantly over the last few decades; and while India hasn’t exhibited the world-historical growth of the Chinese juggernaut, it has still brought an extraordinary number of people out of poverty since the start of the 21st century. At current levels of growth, India will be about as wealthy as China on a per capita basis sometime in the 2040s. It’s impressive, in fact, that India has managed to grow so much without having accomplished the social transformation that China did. Given how brutal that social transformation was, perhaps that’s a good thing.

But it was also a tragedy for the people of India. They remain significantly poorer and worse-off than their Chinese counterparts. The situation for Indian women in particular remains horrific.

So I hope that this history of the Sino-Indian divergence conveys a simple lesson: if you want your country to go from poor to rich, the most important thing is investing in your people.

When I was in India last year, one of the main things I noticed about Indian policymakers was their firm belief that with a few adjustments—industrial policy here, market liberalization there—India could start to match China’s growth record. And I don’t condemn them for thinking along those lines: good policies certainly do help a country grow. But China’s explosive growth wasn’t simply a matter of “freeing the markets,” reducing the role of the state, and announcing that it was now glorious to get rich; nor was it simply a matter of government intervention to support the manufacturing sector and subsidies for favored companies. China succeeded because it spent decades on the basics of human development and social modernization. India did not. The rest is just commentary.

This Substack is supported by readers like you.

It's Not Just X. It's Y.

Mon, 01 Jun 2026 02:29:39 +0000

Against the Quantification of Integrity

When the measure of language becomes its target, it ceases to be good language.

💡

Nerd Rating: 1/5. I discuss the origins of certain linguistic tics in LLMs and what it means for writing, student assessment, and thinking.

"It's not x, it's y."

Large Language Models gravitate toward this type of construction, called negative parallelism. It has its uses: it sets up a contrast. It's useful, especially, for reframing assumptions: "You think it's like that, but it's really like this."

It's all over social media, especially on LinkedIn, and the construction has sparked a backlash amid an ongoing war against automated language production. If you use em-dashes – you might be a bot. If you describe things that delve, quietly, or genuinely (or create lists of three, like that one), you might be a bot.

Recent overuse by language models has led many to declare it bad writing. I'm not so sure. Nobody called JFK a lazy writer when he said, "ask not what your country can do for you – ask what you can do for your country." Negative parallelism is a rhetorical device, and any rhetorical device is only as lazy or inspired as what it contains.

Automated Language Production

Now, we have AI detectors that claim to protect you from the witch hunt by looking for these patterns. You take your own writing and you run it through Grammarly, which will analyze word patterns that AI detectors might flag. Then it offers ideas for how to change them, which a) gives Grammarly the power to write for you and b) makes your writing lose any sense of rhythm or intent.

Grammarly's review of this section has flagged 27 examples of text I should change to avoid the accusation that I am a machine. For example, Grammarly identified the above phrase – "automated language production" – as 11 times more likely to be AI. It suggests that a human would be "against mechanized language synthesis" instead. The simple two-word combo, "align with" was flagged as 43x more likely to be AI-generated. Real humans say "corresponds." These are small suggestions that add up until the result resembles nothing I chose. The human voice replaced by a machine trying to sound human.

As a result, I just paid Pangram – another AI-detection company – $20 to verify that a recently submitted journal article wasn't AI-generated before submission. It wasn't, and I knew it wasn't. It agreed. That's what I paid for: not to learn whether I wrote it, but to be told it wouldn't flag me. Because if Pangram's AI system found me guilty, that's the end of my career. That's literally extortion.

And if it had flagged it, then what? It would give me a score (four valuations: high, very likely, somewhat likely, human) to assign my integrity a category. In the ecosystem we're all building, I'd have to use Grammarly to rephrase everything: using a machine to write for me to prove that I didn't use a different machine to write for me.

A Culture Hostile to Reason

Our instinct in making sense of these machines is to examine the training data. That training data is no longer "just the Web." The web is the raw meat, but this sausage is heavily pre- and post-processed. Post-training optimizes the model for whatever it's designed to do. This includes techniques such as RLHF (reinforcement learning with human feedback) and RLVR (reinforcement learning through verified rewards). RLHF has humans rank replies, then the system emphasizes those kinds of replies.

RLVR is weirder, and I suspect it's why we see "It's not X, it's Y" so often. Dismissing negative parallelism as lazy gets in the way of understanding why it's showing up everywhere. This type of language is such a powerful framework for thinking that we mistake it for a model's capacity for thought. We credit computation for the work that's done by language.

Weird Dogs

RLVR isn't a structure that watches for words and triggers some sub-process. Instead, you train a model, like you would any model. When that model is done, it predicts tokens. Lots of people are still in denial about this. Token prediction involves producing a list of candidates based on their mathematical distribution in the training data, ranking them by their likelihood given the previous words in the prompt or sequence.

RLVR intervenes by having the model solve math problems by writing their way to a solution, reproducing the language we would use when thinking out loud about how to solve it. When the model arrives at the correct answer, the language it used most often to get there is then emphasized in the finished model. This is (partly) what the industry calls reasoning.

What day was it that we saw that weird dog?

So, think of it like this: You are sitting with a friend. Your phones are dead. Your friend asks: what day was it that we saw that weird dog? You start by saying, "It was Thursday." Your friend says: "No, it wasn't Thursday, because Thursday I was out of town." So you say that's right, so it must have been Wednesday, because Wednesday was your mutual friend's birthday, and you both went to the party, and you saw the dog on the way to the party. Your friend says: "That's right, except, Wednesday was our friend's birthday but the party was on Friday. So we must have seen the dog on Friday."

The two of you have articulated your way to the answer, a verifiable one: you could pop on your phones and check your photos and see that yes, the weird dog picture was taken on Friday. In dehumanizing terms, your gut instinct ("it's Thursday") is what a model might spit out at first guess, and that's where models used to stop.

But you didn't. Your friend countered: "It wasn't [Thursday], it was [Wednesday]." There are more words, which narrow the window of possible answers, and then you arrive, through "its-not-x-its-y-ing," at the correct date. The two of you had actual memories and visceral experiences to work with. Language was the vessel through which these experiences were communicated and conflicts were resolved. The model, by contrast, extends language in longer and longer bursts, replicating the pattern of reasoning you two just engaged in. These longer runs re-enact that deliberation within language rather than through it.

Other high-entropy states get filled by words like "suppose..." which triggers longer speculative passages. "Because," "consider," "alternatively," even "wait" can occupy these positions. These are words that lead to language that brings contrast, exceptions, and abstraction along for the ride. If they get to a correct answer on a math problem, they get pushed to occur more often.

The Reason We Reason

When we talk about a weird dog or have conversations like it, the point of the question was not to identify the date on the calendar when the dog was encountered. It was an opening for a reminiscence. It was posed to reconstruct the memory, to revel in its surrounding context, and to deepen a connection between friends through a shared experience.

Defining reasoning this way assumes that the point of asking a question is to get an answer, that answers can be verified, and that nothing is lost in immediate closure.

Defining reasoning the way it has been used in LLMs assumes that the point of asking a question is to get an answer, that answers can be verified, and that nothing is lost in immediate closure. This has real effects on writing, and the openness to doubt is something we lose in the rapid prototyping of thought that occurs with a language model. Ambiguity, doubt, and uncertainty matter more to some ways of thinking than any immediate answer. The inner life grows in the spaces between the industrial complexes that harness every remnant of our externalized thought.

Nonetheless, the language we use in these states is the same. When AI detectors flag text as AI-generated, is it because it follows a certain structural pattern of that reasoning? Pangram and reasoning models both detect structural patterns based on how humans reason when writing. Pangram's model is trained on pre-2021 data; it then inserts AI-generated versions of the same text into its training.

So, if we publicly shame people whose text looks like it might have been written by a machine – because it mimics the language used for human reasoning – and people stop writing in ways that they internalize as "AI writing" out of fear of false detection, it sends a signal that your language for reasoning must be policed, or you too could be held up to public scrutiny.

In the end, shaming people for writing that gets flagged as AI can lead people to sidestep structures the model has learned from us: structures that are effective tools for argumentation. We take the tools of critical thinking out of the kit at the time we most need them.

For Good Measure

There's another angle to this. An AI-based essay assessment tool was tested in the UK against human graders. The system rewarded writing structures that I can't help notice look a lot like RLVR-based reasoning: "giving out higher marks based on essay length, vocabulary range and sentence complexity, which are often unrelated to academic standards," all of which are hallmarks of AI reasoning.

In other words, the LLM grades humans based on the criteria engineers use to assess the LLM.

The LLM grades humans based on the criteria engineers use to assess the LLM.

There's this old adage from economics called Goodhart's law. The econo-nese version of it is that "any observed statistical regularity will tend to collapse once pressure is placed upon it for control purposes." Or: when a measure becomes a target, it ceases to be a good measure. It could be tweaked to apply to large language models: "when the measure of language becomes its target, it ceases to be good language."

There is danger in evaluating for language patterns over its content, and both generation and detection incentivize this. Automated grading is somewhere between the two: rewarding students for employing the form of reason over the act of reasoning will only make them more tempting and more common. And yet, punishing the form risks punishing reason. Ultimately, we have to think critically in all cases, instead of deferring to the judgments of machines.

Against Automatic Thinking

I'm not convinced by the old "if you haven't done anything wrong, you don't have anything to worry about" line. I've seen 99.8% cited as a measure of accuracy in automated surveillance systems since 2018. As Arvind Narayanan has noted, that is on a per-paper basis, which compounds every time we use it. So up to 10% of college students could be falsely accused. If we collectively run every bit of text through an AI model to check whether it is AI-generated, we will generate false positives on an even larger scale.

These models concentrate real authority; companies promise they will reason on our behalf. We normalize something dangerous when we run every two-line phrase through an AI interpreter, post the result online, and say "see? They're plagiarists!"

We create a culture of self-censorship and AI-detector-pressured rewriting and paraphrasing as people strive to avoid these witch hunts. That is the opposite of protecting human expression. We should resist normalizing a trust in any machine's ability to determine matters of guilt. If using AI to write is, at its worst, an industrialization of the mind, then AI detection, at its worst, becomes a surveillance system for thought.

Monthly, for the Second Week in a Row.

Thanks for reading! As mentioned last week, I am only a sporadic poster these days, aiming for once a month. If you're paying for the newsletter and would like to calibrate your donations (or would like to start supporting it!) you are very welcome to set up or change your subscription here.

Simple Sabotage Field Manual

Wed, 27 May 2026 03:22:34 +0000

This re-publication comes from a deeply uncomfortable recognition. Open to Section 11 and you'll find instructions that could have been lifted from yesterday's management consultant: "Refer all matters to committees." "Haggle over precise wordings of communications." "Hold conferences when there is more critical work to be done." The saboteurs' playbook has become our best practices.

We puzzle over our economic stagnation, wondering why the technological revolution hasn't made our organizations faster. Since the 1970s, productivity growth has limped along at roughly half its post-war pace, except for a brief internet-fueled surge in the 1990s. The answer stares at us from these pages.

About This Publication

The original Simple Sabotage Field Manual was declassified by the CIA and is in the public domain. What you're reading here is our foreword to the republished manual, exploring its unexpected relevance to modern organizational challenges. The historical document serves as a lens through which to examine how bureaucratic dysfunction has become embedded in contemporary business practices.

Download the Complete Manual (PDF)

Leo Tolstoy opened Anna Kareninawith this observation: "All happy families are alike; each unhappy family is unhappy in its own way." But what may have been true for the Kareninas and the Vronskys is reversed for companies. All unhappy companies are alike: they run on committees, they worship process, they strangle themselves with approvals. They're unhappy in exactly the same way, following exactly the same script. The saboteurs of 1944 are mostly gone. But their manual lives on in every corporate handbook, every best practice guide, every management consulting deck. We are their greatest success.

Origins and Intent

The Simple Sabotage Field Manual was initially published in 1944 by the Office of Strategic Services. Although details remain sparse about the document's origins, we know that the Simple Sabotage Field Manual was part of a broader collection of OSS field manuals designed to codify the emerging doctrine of unconventional warfare, encompassing commando operations, intelligence gathering, psychological warfare, and guerrilla tactics.

These manuals formalized an emerging view that modern warfare reached into factories, offices, and rail yards—not just the front lines. The officers who wrote these manuals went on to found Army Special Forces and shape CIA covert operations. That wartime effort to codify irregular warfare underpins today's hybrid warfare, where states blur the line between war and peace through cyberattacks, disinformation campaigns, economic pressure, and proxy forces, thereby achieving strategic goals without conventional military confrontation.

The Simple Sabotage Field Manual occupied a unique position in this arsenal. While other manuals required trained operatives and specialized equipment, this one democratized resistance, providing ordinary citizens with techniques to create systemic dysfunction through seemingly innocent acts of incompetence and delay.

What made this approach revolutionary wasn't just the tactics, but the strategic insight: thousands of small disruptions, when coordinated with propaganda campaigns and commando operations, could multiply the impact of conventional military strikes—grinding down the enemy's war machine from within while allied forces attacked from without.

As the manual itself states: "Simple sabotage does not require specially prepared tools or equipment; it is executed by an ordinary citizen who may or may not act individually and without the necessity for active connection with an organized group."

Alephic is an AI-first strategy and software partner that helps marketing organizations solve complex challenges through custom AI systems.

We believe AI enables brands to apply software development principles to qualitative problems—the judgment-based challenges that have always defined marketing but couldn't be solved with traditional code.

We're builders, not consultants. We solve problems by shipping code, not PowerPoints. Our mission: be every CMO's first call as they navigate AI.

About Alephic

Weaponizing Bureaucracy

While the manual covers many forms of sabotage—from sugar in fuel tanks to misrouted shipments—its true genius lies in recognizing bureaucracy itself as a weapon. Physical sabotage could be discovered and repaired. But bureaucratic sabotage? It looked exactly like business as usual.

The anthropologist David Graeber explained why bureaucracy destroys so effectively: bureaucracies are "utopian"—they create an abstract ideal of perfect process and procedure that real human beings can never live up to. When people inevitably fail to meet these impossible standards, the system blames them for not following the rules. This gap between bureaucratic fantasy and human reality doesn't create efficiency—it creates exactly the kind of systematic dysfunction that wartime saboteurs learned to exploit.

During the Nazi occupation, Pierre-Jules Boulanger, the vice president of French automaker Citroën, understood this perfectly. He instructed his foremen that "production must appear respectable to the eye, but never to the heart." He therefore ordered that the small notch showing 'full' on the engine oil dipstick be filed 8 mm lower than specification. German mechanics, trusting their tools, dutifully filled each truck to the false mark. Within a few hundred kilometers, the engines seized, and Citroën's line workers quietly celebrated another truck sabotaged beyond repair.

This is the genius of simple sabotage: it weaponized the enemy's own systems against them. A worker could damage machinery by letting cutting tools grow dull or "accidentally" dropping sand into lubricating systems. Office clerks could misfile critical documents, make errors in enemy orders, or create endless bureaucratic delays. Citizens could spread false rumors, give wrong directions to enemy convoys, or simply work slowly and inefficiently. Each act seemed like mere incompetence or bad luck, but multiplied across thousands of people, these small frictions created massive operational drag.

The Uncomfortable Mirror

When the CIA made this manual widely available in 2008, it didn't immediately go viral. Instead, it percolated through the internet, gaining momentum as readers saw something unsettling in its pages.

The reason for this sustained interest was simple: readers saw their own workplaces in the sabotage techniques. The manual's instructions for office workers to "insist on doing everything through channels," "make speeches at great length," and "multiply the procedures and clearances involved in issuing instructions" weren't just effective at disrupting enemy operations—they had become standard operating procedure in offices worldwide.

By 2015, business consultants were mining it for insights. The CIA itself published articles about its continued relevance.

The saboteurs didn't invent these techniques. They simply recognized that the most effective way to destroy an organization was to make it more bureaucratic. Every pointless meeting, every form that requires seventeen signatures, every process that takes six weeks when it could take six minutes—these aren't failures of the system. They're the system working exactly as designed.

The genius of the Simple Sabotage manual is recognizing that bureaucracies naturally tend toward their own dysfunction. They create what amounts to organizational scar tissue—layer upon layer of rules and procedures that slowly strangle the actual work. The manual just accelerated this natural process. In this light, perhaps the most insidious question isn't whether we've sabotaged ourselves, but whether bureaucracy is simply the inevitable outcome of organizational success.

Sabotaging the Sabotage

The solution isn't to fix bureaucracy. Decades of consultants have tried that approach, each wave promising to streamline processes while adding new layers of complexity. The solution is to make bureaucracy irrelevant.

What makes bureaucracy so insidious is that any attempt to break it by adding more rules or processes ultimately strengthens it. You can't fix bureaucracy by creating more bureaucracy. The cure always deepens the disease. Only removal works.

However, most organizations take the opposite approach. In the United States, the percentage of managers in the labor force increased from 9.2 percent in 1983 to 13.2 percent in 2002 and continued to rise to 36.2 percent by 2020. This army of middle management acts as a buffer, insulating decision-makers from reality while creating ever more complex approval chains. Each layer exists to manage the layer below it, creating what one might call a "bureaucratic pyramid scheme."

A Personal Reflection

As an entrepreneur, I've watched this process unfold firsthand. When you start a company, it's amazing how much a small group of people can accomplish. Five people in a room can build products, talk to customers, make decisions, and move mountains. There's no process because everyone knows everything. There's no bureaucracy because there's no time for it.

But then you grow. Ten people, then fifty, then a hundred. Fred Brooks, a computer scientist and former IBM manager, warned about this in The Mythical Man-Month. Each new person doesn't just add one more relationship to manage—they add as many new relationships as there are existing team members. By the time you reach corporate scale, you're drowning in coordination. Meetings about meetings. Approvals to get approvals. Everyone spending more time syncing up than doing actual work. As Brooks memorably put it: "The bearing of a child takes nine months, no matter how many women are assigned."

You need systems to onboard people, processes to coordinate work, and managers to manage the managers. Before you know it, you're spending more time in meetings about meetings than actually building things.

I used to think this was a failure of leadership or organizational design. Now I wonder if it's simply physics—the inevitable entropy of human systems. Success breeds scale, scale breeds complexity, complexity breeds bureaucracy. The very mechanisms that protect large organizations from chaos also protect them from change.

The genius of the Simple Sabotage Field Manual wasn't in teaching people how to destroy organizations. It was in recognizing that organizations naturally destroy themselves. The saboteurs just gave entropy a little push.

Which brings us to an uncomfortable question: If bureaucracy is the price of scale, and scale is the reward for success, then isn't every successful organization destined to become exactly what the saboteurs envisioned? And if that's true, what does it mean for those of us still naive enough to believe we can build something different?

Maybe that's why this 80-year-old manual still resonates. Not because it teaches us how to sabotage, but because it holds up a mirror to what we've become.

A Possible Escape

And yet, for the first time in my career, I wonder if we might have found a way out. Not by fixing bureaucracy—that's the trap every generation falls into. But by making it irrelevant.

AI is unlike any technology we've built before. Every previous tool required us to conform to its specifications, to translate our messy human processes into rigid machine logic. AI does the opposite. It adapts to us. It becomes what I call a "fuzzy interface"—capable of understanding intent rather than requiring perfect syntax, of bridging incompatible systems without forcing standardization.

Think about what this means. All those bureaucratic layers, those translation tasks, those forms and processes, and approval chains—they exist because humans needed interfaces between other humans and systems. What if we didn't? What if AI could fill all those gaps, handling mechanical compliance while we focus on the human work?

This isn't naive optimism about AI eliminating bureaucracy. But it might—just might—let us build secret passages around it. To create organizations where the machinery still runs, but we don't have to be cogs in it, and where success doesn't inevitably lead to sclerosis.

The saboteurs of 1944 understood that bureaucracy was organizational entropy made visible. Maybe, 80 years later, we've finally found a force that can push back against that entropy. Not through more rules or better processes, but through intelligence that adapts rather than constrains.

Time will tell if this optimism is justified. But for the first time since I started building companies, I believe we might be able to have our scale and eat it too.

— Noah Brier, Co-Founder, Alephic

Nobody Cracks Open a Programming Book Anymore · unix.foo

Tue, 26 May 2026 00:10:28 +0000

There was, for a long time, a wall.

If you walked into a book store, past the magazines and the cookbooks, you’d arrive at the computer section, and along one wall there was a stretch of books with cartoon animals on their covers. A rhino for JavaScript. A camel for Perl. A python (obviously) for Python. And whatever this was:

They were thick, they cost about $50, and they had titles like “Learning React” and “HTTP: The Definitive Guide”. If you wanted to learn how to do a thing on a computer, you bought one of these, took it home, and opened it up next to your computer and typed what it said until the thing worked.

That wall is smaller now. If it’s even still there. In some stores the wall is gone and relegated to a small rack that has six books on it, three of which are about ChatGPT.

Through the first nine months of 2023, sales in the “computer book” category at Circana BookScan (the industry’s standard tracker, which costs roughly the price of a small used car to subscribe to) were down 16.9% year over year. Publishers Weekly, which had been dutifully reporting these figures in its quarterly narrative summaries, kept doing so right up through that 16.9% figure, and then in 2024 and 2025 simply stopped mentioning the category by name.

To be clear, books in general are doing fine. Total U.S. print sales reached 762.4 million units in 2025, up 0.3% over 2024, which was itself up 0.5% over 2023. The category that is in trouble is the part of it that teaches you how to make software. The American Association of Publishers’ “professional books” segment, which is the rough corporate proxy for “books your employer might buy you,” fell 22.3% in August 2025.

The book industry is fine but the technical end is bleeding out.

Quickly and quietly. There was no Napster moment for the programming book. Nobody filed a lawsuit. The publishers did not, as far as I can tell, even hold a press conference. We simply found one day that they stopped reporting the category itself. The category doesn’t die, it just stops being talked about.

You already know why, more or less. ChatGPT has over 900 million monthly active users. GitHub Copilot has 4.7 million paying subscribers as of January 2026, up roughly 75% in a year. You can’t imagine writing software without Claude Code anymore.

Stack Overflow is receiving about 3,800 questions a month, which is what it was getting in 2008, before it had finished being launched. The chatbots have eaten the demand for the kinds of answers that programming books used to provide.

The programming book was, when you look at it squarely, always a slightly absurd object. Printed text on bound paper, describing software that lived on screens, which the reader had to retype, by hand, into a screen of their own. I loved doing this and they remain some of my very fondest childhood memories. But the medium was wrong for the content. People put up with it because there was no better way to get a careful sustained explanation of a technical thing into one person’s head from another person’s.

What the book was good at, despite being the wrong format, was forcing both the writer and the reader to be slow. You cannot fake your way through 400 pages. It took a certain discipline to get through.

The chatbot does not have this discipline. The chatbot has read every book and forgotten the point of every one of them. It will explain idempotency in the precise number of words you require, and you will close the tab, and you will not remember what it told you, because you did not type it.

That last sentence is the whole thing. Knowledge, for working programmers, was always the residue of typing. Of doing. The typing was the practice! What is going away is the typing.

Which, on balance, may be fine. I don’t know. People used to lose weekends to installing Linux from a stack of floppies and struggling with WinModems, and nobody pretends that was character-building (though I now consider them fond memories too). Tools get easier. Skills shift. The kid who is right now learning to code by chatting with an agent is not a worse programmer than I was at 12, hunched over Learning Perl, retyping examples that would not run because I missed a semicolon.

That kid is a different programmer. They are, in some ways I don’t fully understand, working at a higher level of abstraction than I ever did at that age, and the things they will build with that abstraction will surprise me.

But somewhere in a used bookstore in San Francisco or Seattle or wherever used bookstores still exist, there is a 1997 edition of Learning Perl. It smells faintly of basement. Someone wrote their name in the front of it in pencil. There is a furiously underlined sentence in chapter 7 about regular expressions that was made in anger. On page 112 there are coffee stains where the caffeine blots are somehow still a valid Perl program.

The book costs three dollars. Nobody is going to buy it.

Tall poppies don’t get the calls – Economist Writing Every Day

Mon, 25 May 2026 07:49:50 +0000

Ask anyone who grew up playing basketball as the tallest player on the court and they will, each and every one of them, tell you that players were allowed to foul them harder and more often. If you were tall you didn’t get the calls, full stop. Why? We could sort through a host of mechanisms, but they all boil down to “Being tall is an unfair advantage. It’s only fair that I, the shorter opposing player, am allow to slap you, chop you, kick you, trip you, grab you.” To be honest, I don’t think this is a particularly shocking phenomenon. “Tall poppies get cut down” is a cultural cliche for a reason. What is interesting is that it persists even amidst billions of dollars in market incentives pushing in the other direction.

The latest version is happening right now as the Oklahoma City Thunder are currently doing their best to end Victor Wembanyama’s nascent career each and every night, and the referees seem particularl disinterested in religning the incentives otherwise. At the moment the Spurs are currently up 65-43 in game 4 of the series. If the series goes 7, there’s at least a 20% change Wembanyama doesn’t make it to the end. Will they break his foot smashing down in it, break his leg tripping him, or dislocate his shoulder yanking down on it from a leveraged position? Don’t know, but they’re doing their best to make it happen.

Caitlin Clark came into the WNBA as the single greatest talent prospect in the history of women’s basketball. The abuse she suffers is well documented. Wayne Gretzky was the greatest hockey player of all time, but he was arguably only allowed to reach his potential because Bobby Orr’s careers was cut in half by a league that allowed teams to abuse him with little to know punishment. Bobby Orr’s sin was that he was such a better skater than everyone else that, if allowed to play without constant grabbing, hooking, and abuse tantamount to aggravated assault, he would have walked away with too many goals, wins, and Stanley Cups. It wasn’t fair that he was so much better, so they let the players even the odds. Having watched him limp away after only 7.5 seasons, the NHL took the unofficial position that Gretzky’s teammates (specifically, Dave Semenko and Marty McSorley) were given carte blanche to assault anyone who touched Gretzkey. While perhaps not a culture-shifting solution, but Gretzky did have a 20 year career that brought hockey to new heights of popularity, so it was ostensibly effective.

But none of that gets at the underlying economics. Elite players bring big audiences to sporting events, which in turn, brings in big money for everyone. The owners, players, and everyone in between gets richer when elite players shine under the biggest lights. So why chop them down? Well, first we have a collective action problem to solve, because, yes, the entire market benefits from superstars, but their opposition during the course of play in most any game have the individual incentives to do whatever they can get away with to win. That’s why we have referees, commissioners, and a players union: to solve those collective action problems. All of those rules and institutions are in place specifically to solve problems like these. To align incentives and bargain for outcomes that maximize welfare. So why aren’t they working?

When you find cliches at the front of your mind, decent chance you’re running up against psychology and behavioral economics. And as Victor Wembanyama is learning each and every night of the playoffs, “tall poppies get cut down”. It’s not fair that he’s the first 7’5″ player with elite NBA level skills to ever play the game. You know, I was never a fan of watching Shaq play basketball per se, but I always knew he should have scored at least 40 points every night. Yes, he committed 7 offensive fouls every game, but he also received 25 fouls that went uncalled. Players were allowed to maul him because it was unfair he was so much bigger, stronger, and more athletic. His career was only as long as it was because his body could endure the abuse. There has never been another player in NBA history who could have survived even 3 seasons receiving the abuse he did.

Putting aside simple behavioral explanations, we also should consider the possibility that NBA team owners and players are so far down the diminishing marginal returns to wealth, that the median participant would actually prefer to earn less money in order to maximize their own chance at winning a championship. They want parity, of a sort. Parity, but only once the playoffs arrive. The regular season is too long and everyone needs to make money, so the abuse is minimal, but once the playoffs arrive, the collective preference is for parity delivered via weaker rule enforcement. There are only so many elite players, but everyone is capable of low-level violence. This preference for postseason parity may also explain why Oklahoma City’s best player, Shai Gilgeous-Alexander, has the reputation for simulating being fouled on every player. If you’re going to get fouled no matter what, you might was well maximize the probability of getting a foul call by forcing the referee to be observed observing the incident.

And, to be clear, parity may in fact be revenue maximizing. Just look at the NFL – the entire structure is designed to maximize the number of franchises who believe at the beginning of the season that their team has a chance to win it all. The players are relatively anonymous compared to NBA superstars, but fans are mostly there to root for laundry, and in the NFL, so long as that laundry doesn’t say NY Jets on it, there’s at least a glimmer of hope. Counter point, just look at the NFL. They understood that each team, especially once the playoffs started, had strong incentives to try to end the opposing quarterbacks career on each and every play. So the NFL introduced a battery of rules to protect quarterbacks, and it seems to have worked.

So maybe I’ve come full circle. Maybe this is what the NBA wants. But I really, really it’s hope not. Wemby is special. I’d like to see the very most of what he can become.

我是杠精

Mon, 18 May 2026 22:02:09 +0000

从小到大，我在朋友们中都是一个知识面比较丰富的人的形象：不管你想聊啥，十分钟内我都可以佯装专业人士跟你聊而不露馅。我自己也是自认为是一个通才而不是一个专才，并为通才的身份而自豪。但是到了AI时代，你再渊博也渊博不过AI，明显专才在这个时代比通才更有用：专业到一定程度肯定是AI比不了的。所以过去一年多的时间里，我有很强的挫败感：我这样的人还有用吗？我是不是应该试图转变成一个专才？

过时的本事

很久以前读过一本小说是冯骥才写的《神鞭》，讲的是清末一位高手的一身本领都在脑后的一根辫子上，真的是指哪里打哪里。但是不久后到了民国，大家都剪掉了辫子，神鞭的功夫一下子一文不值。故事的最后神鞭又找到了新的本事，我也能像他一样转型成功吗？

假如我要从通才转型为专才，这其实并不像学一门功夫。先不考虑年龄和学习能力的问题，想对一个领域钻的很深，成为专才，这和我天性相反，我对枯燥的细节向来兴趣寥寥，静不下心钻进去。更何况成为专才需要不短的时间，这世界变化快，我又预测不了未来，假如我选的专业也成了辫子功，那我不是白忙活一场？

通才的通

假如我转型不了，那通才还有什么价值？通才作为多面手，在找不到专业人士的时候顶一小会儿，提个建议什么的，这是一点意义都没有了。这也是我的苦恼之源。但是“通”还有别的意思：

沟通。我懂的多些，所以我可以和很多不同职业，不同背景的人沟通。这个意义在AI时代还存在，只是缩小了很多。假如说需要沟通的人的需求是获得帮助，那我没什么用，AI做的比我更好。但假如这个人需要的不是抽象的帮助，而是同理心，安慰，鞭策，或者仅仅是一个段子呢？那我作为一个人，还是比AI强的。

变通。我这人背景比较复杂，没有很强的地域，宗教，文化限制，我可以短暂地放弃其他背景，变得和对方背景类似，为对方设身处地地思考，或者仅仅是一起吐一吐苦水，这总有点用的。AI的问题是过于渊博，恐怕装不了单纯，而且别人已经知道它是AI了，也没法去装。这个用处有多大不好说，大概是不大。

以上两点单独看都用处不大。但是组合起来，就有了个新的用处。假如你的目的是说服什么人什么事，我就是你最好的练习对象：

我有相应的背景，懂你的语言，所以你没有鸡同鸭讲的危险。
我没有很强的预设立场，所以我不容易被得罪，但需要被说服，而且是有可能被说服的。
最重要一点，我其实很难被说服，因为我还是一个杠精！

什么是杠精

一个杠精会故意跟你抬杠，提出相反的意见。但是，一个真正的杠精不会用资历来压人，因为这胜之不武，也不会扩大或转移话题来故意惹你生气，因为这是在没有人围观的时候哗众取宠，纯属浪费精力。一个真正的杠精只会做一件事，就是指出从你举的证据到你想得出的结论中的逻辑推理有漏洞，还存在别的可能性，而且还不小。

只有一个通才才能做一个好的杠精。虽然说在辩论中，我不能用对方没有，或不接受的其他资料背景来做证据，但天下没有新鲜事，其他领域的其他经验可以让我提出其他种种对方逻辑覆盖不到的可能。相反，一个专才的经验会相对比较冷僻，难以拿来移植到当前辩论话题。

一个AI模型不是一个很好的杠精。首先，AI模型被训练地倾向于服从而不是辩驳：You’re absolutely right! 是它的口头禅。其次，就算用户强令AI去辩驳，它也无法像一个人这样在聚焦和触类旁通之间找到平衡。最后，和AI吵架赢了输了都挺没劲的。

善待你在网上遇到的杠精吧。他们虽然会惹你生气，但真的有用，而且至少是人。网上人已经不多了。

Prisoner’s Dilemma is a Scam

Wed, 20 May 2026 11:47:01 +0000

The Prisoner’s Dilemma is the most celebrated thought experiment in game theory and, I dare say, one of the most intellectually dishonest constructions in the history of social science. Not because the mathematics is wrong—the mathematics is airtight, as mathematics tends to be when you get to choose the axioms. The fraud is in the premises.

For those mercifully unacquainted: two criminals are arrested. Each can stay silent (cooperate with the partner) or testify against them (defect). If both stay silent, each serves one year. If one defects while the other stays silent, the defector walks free and the loyal fool serves three. If both defect, both serve two. The “dilemma” is that rational self-interest dictates defection, even though mutual cooperation yields a better outcome for both. The Nash equilibrium is mutual betrayal. Two rational agents, acting rationally, produce an irrational result. Checkmate, humanity.

This story has launched two thousand scholarly articles, several dozen popular books, and at least one documentary narrated in the kind of hushed voice normally reserved for tectonic plate collisions. The problem is that the story has roughly the same relationship to actual human decision-making as a spherical cow has to dairy farming.

The Outside World, or: Snitches Get Stitches

The standard formulation contains a clause so brazen that one wonders how it survived peer review. Let me quote: “It is assumed that both prisoners have no loyalty to each other, and will have no opportunity for retribution or reward outside of the game.”

Read that again. Slowly.

“No opportunity for retribution or reward outside of the game.” This is not a simplifying assumption. This is the removal of the entire ecosystem in which the decision operates. It is the equivalent of an aerodynamics paper that begins with “assume no air.” You can certainly derive results under such a premise, but you should have the decency not to call them results about flight.

Criminals—actual criminals, the kind who get arrested in pairs and face prosecutors offering Faustian bargains—exist in social networks denser than a neutron star. The reason is not sentimental; it is economic. Criminal enterprise is, at its core, a repeated cooperation game with extremely high stakes, and the networks that survive are precisely those that have evolved mechanisms to make defection suicidal. The Sicilian Mafia did not invent omerta because Sicilians have a particular fondness for silence. They invented it because organizations whose members rat each other out at the first prosecutorial wink have the life expectancy of a mayfly. Natural selection applies to institutions as ruthlessly as it applies to organisms, and the institution of “snitch freely, suffer no consequences” went extinct before it could leave fossils.

The Japanese yakuza have yubitsume—the ritual severing of a finger joint as atonement for disloyalty. The Russian vory v zakone maintained an entire parallel legal code inside the prison system, one in which cooperation with authorities was punishable by death. These are not colorful cultural footnotes. These are equilibrium-enforcement mechanisms, evolved over centuries by groups whose survival depended on solving exactly the problem the Prisoner’s Dilemma claims is unsolvable. The game theorist’s model is a world in which these mechanisms have been surgically removed, and the conclusion—“look, cooperation collapses!”—is presented as insight rather than tautology.

Even in the absence of organized crime, the argument holds. Neighborhoods have memories. Prisons have hierarchies. The man who walks free today because he testified will walk the same streets tomorrow, and the streets will know. The cellblock has its own information network, and it is more efficient than most corporate intranets. The notion that the “game” ends when the prison doors open is a fantasy that could only be entertained by someone whose closest encounter with the criminal justice system is a parking ticket.

The iterated version of the Prisoner’s Dilemma partially addresses this objection. When repetition is introduced, cooperation emerges: Robert Axelrod’s celebrated tournament showed that tit-for-tat—cooperate first, then mirror the opponent’s previous move—beats strategies built on pure defection. Nash himself observed, watching Flood and Dresher’s original hundred-round experiment, that Alchian and Williams cooperated far more than the one-shot theory predicted. But here is the thing: there is no such thing as a one-shot game in real life. Every human interaction is embedded in a web of relationships, expectations, and consequences that extends far beyond the immediate transaction. The one-shot Prisoner’s Dilemma is not a simplification of reality. It is a different planet.

To strip away reputation, future interaction, social networks, and the threat of retribution, and then to marvel at the collapse of cooperation, is approximately as profound as removing the walls from a building and expressing surprise that the roof falls down. The walls were doing the work. The game theorist removed them and then blamed gravity.

Pride in a Vacuum

Very well. Let us, for the sake of argument, accept the vacuum. One shot. No future. No retribution. No social network. Two strangers in separate rooms, never to meet again, never to meet anyone who has met the other. A decision, crystalline in its isolation, untouched by any consideration outside the four walls and the ticking clock.

Even here—especially here—the Prisoner’s Dilemma is built on a hidden assumption so fundamental that most treatments never bother to state it: the payoff matrix is complete.

It is not.

The matrix shows prison sentences. It assigns numbers to years behind bars. What it does not show—what it cannot show, what it deliberately omits—is the weight of the decision on the person making it. The reflection in the mirror on the morning after. The thing you know about yourself that no parole board will ever ask about and no sentence reduction can erase.

For a certain type of person—the type that game theorists call “rational” and the rest of us might describe less charitably—the matrix is indeed complete. The only thing that matters is the number of years. Freedom is good, imprisonment is bad, the conscience is a rounding error. For this person, defection is the dominant strategy, the Nash equilibrium is mutual betrayal, the mathematicians are right, and the world is a grimmer place than it needs to be.

But humans are not uniform in this regard. Pretending otherwise is the second great fraud.

Some people would rather serve three years with a clean conscience than walk free knowing they sent a man to prison for keeping faith. This is not irrationality. It is not “bounded rationality” or “cognitive bias” or any of the other euphemisms that economists reach for when humans decline to behave like their models. It is a different utility function. The game theorist’s error is not in the algebra; it is in the ontology. They assumed that the payoff matrix—that neat little grid of prison sentences—describes the entire space of consequences. It does not. The matrix is missing a row. The row is labeled self-respect.

Let us call this hidden variable pride. Not pride in the vainglorious sense—not the peacock’s tail, not the Instagram selfie—but pride in the sense of self-regard: the internal cost of knowing that you have done something you consider contemptible. For some people this cost is zero. For others it exceeds any sentence the state can impose. Between these extremes lies the entire spectrum of human character, conveniently erased by the assumption that all players are “rational agents concerned only with minimizing their prison sentences.”

The quantity that interests us is the ratio of pride to rational self-interest—the pride/mind ratio. Let us call it ρ (rho). A person with ρ = 0 is the game theorist’s ideal rational agent: a spreadsheet with legs. A person with infinite ρ is a saint or a fool, depending on your theology. The rest of us live somewhere in between, and the exact location has consequences that the standard Prisoner’s Dilemma refuses to compute.

Let us compute them.

The Balanced Game

The Setup

In the standard formulation, the payoffs for player A (expressed as negative sentence years—higher is better) are:

Defection dominates: regardless of what B does, A is better off testifying. Against a silent B, testimony yields freedom (0) rather than one year (-1). Against a testifying B, testimony yields two years (-2) rather than three (-3). The logic is impeccable. The premises are what is crippled.

Now let us introduce pride. When a player defects, they incur a psychological cost proportional to ρ. But—and this is the crucial nuance—the magnitude of that cost depends on what the opponent did.

Betraying someone who trusted you and stayed silent is, for most humans, a heavier burden than mutual betrayal. If both of you defect, you can at least tell yourself he would have done the same. If you defected while your partner kept faith, there is no such refuge. We capture this asymmetry with a second parameter γ (gamma), ranging from 0 to 1:

Defecting against a cooperator costs the full ρ in shame. You betrayed trust.
Defecting against a defector costs γρ in shame. Discounted by the knowledge that trust was never on the table.

The modified payoff matrix for player A:

When ρ = 0, this collapses to the standard Prisoner’s Dilemma. When ρ is large, cooperation dominates. Somewhere between these extremes, the game balances.

The Derivation

Suppose B cooperates with probability q. Player A is indifferent between strategies when the expected payoffs are equal:

E[stay silent] = q(-1) + (1 - q)(-3)      = 2q - 3
  E[testify]     = q(-ρ) + (1 - q)(-2 - γρ) = q(2 + γρ - ρ) - 2 - γρ

Setting these equal:

2q - 3 = q(2 + γρ - ρ) - 2 - γρ
   2q - 3 = 2q + q(γ - 1)ρ - 2 - γρ
  -1 + γρ = q(γ - 1)ρ

Solving for q:

q = (γρ - 1) / ((γ - 1)ρ)

For the game to be balanced—each player cooperating with probability exactly 1/2—we set q = 1/2 and solve:

2(γρ - 1) = (γ - 1)ρ
  2γρ - 2 = γρ - ρ
  γρ + ρ = 2
  ρ(1 + γ) = 2

Which yields the boundary:

ρ = 2 / (1 + γ)

What This Means

The formula says: the critical pride/mind ratio depends on how much you discount the shame of mutual betrayal relative to the betrayal of a cooperator.

γ = 0—“Betraying a betrayor costs me nothing.” The hardened realist. Full shame for ratting out a loyal partner, none for mutual defection. Here ρ = 2. Translation: if your self-respect is worth less than two years in prison, you defect. If it is worth more, you cooperate. At exactly two years, you flip a coin.

γ = 1/2—“Mutual betrayal stings, but only half as much.” A plausible default for most humans. Here ρ = 4/3 ≈ 1.33. Your self-respect needs to be worth roughly sixteen months of freedom to make you indifferent.

γ = 1—“All betrayal is equally shameful.” The person for whom the act of defection itself, regardless of context, carries the same moral weight. Here ρ = 1. One year of self-respect against one year of prison. The exchange rate is 1:1.

The full table:

γ       ρ = 2/(1+γ)   Meaning
  ──────────────────────────────────────────────────────
  0.00    2.00          Shame only for betraying trust
  0.25    1.60          Mutual betrayal costs a quarter
  0.50    1.33          Mutual betrayal costs half
  0.75    1.14          Nearly as shameful either way
  1.00    1.00          All betrayal equally shameful

The Three Regimes

The complete picture has three regimes, governed by ρ and γ:

ρ < 1—Defection dominates. Classical Prisoner’s Dilemma territory. The shame of betrayal, even at full weight, is less than the one-year difference between cooperation and defection. Mind rules. Pride is a footnote. The game theorists are right, for this particular subspecies of human.

ρ > 1/γ—Cooperation dominates. The shame of defection, even mutual defection, outweighs the material benefit. The “dilemma” ceases to exist; staying silent is strictly preferred regardless of what B does. (Note: when γ = 0, this threshold is infinite—you cannot be shamed into cooperating against a defector if mutual defection costs you nothing.)

1 < ρ < 1/γ—Mixed strategy. Neither strategy dominates. A player cooperates against cooperators (the shame of betrayal exceeds the one-year gain) but defects against defectors (the discounted shame is worth less than the one-year saving). In this regime, the equilibrium is a probabilistic mixture, and the balance point—exactly 50/50—sits at ρ = 2/(1 + γ).

The Missing Variable

The Prisoner’s Dilemma is not wrong. It is a theorem, and theorems are not wrong; they are either correctly or incorrectly derived from their axioms. The derivation is correct. The axioms are the problem.

The first axiom—no outside world—removes the entire evolutionary apparatus that makes cooperation the default rather than the exception in human affairs. This is the axiom that everybody criticizes, and rightly so. The second axiom—the payoff matrix is complete—assumes that prison sentences are the only relevant consequence, which is tantamount to assuming that human beings have no inner life, no self-respect, and no conscience. This is the axiom that almost nobody questions, because it is not stated as an axiom at all. It is smuggled in as a two-by-two grid and accepted without examination.

The arithmetic above shows that the omission is not merely philosophical. It is quantifiable. For any given shame-asymmetry γ, there exists a precise pride/mind ratio ρ = 2/(1 + γ) at which the dilemma dissolves into perfect indifference. Below this ratio, the game plays as advertised. Above it, cooperation is not altruism—it is self-interest, properly accounted for.

The question the Prisoner’s Dilemma actually asks is not “will rational agents cooperate?” It is something far more personal and far more interesting: how much is your self-respect worth, measured in prison years?

The answer, unlike the dilemma, is not the same for everyone.

That, perhaps, is the real dilemma.

Why Japanese companies do so many different things

Mon, 18 May 2026 23:48:04 +0000

Photos from Lars Tunbjörk’s “Office” series

Consider Toto.

If you spend much time in American public bathrooms, or rather if you’re simply a particularly attentive patron of American public bathrooms, you’ll probably have noticed Toto’s toilets at some point or another: they’re distinguished by a quite memorable serif-font “TOTO” logo. Toto toilets aren’t quite dominant in American bathrooms, since they have healthy competition from our homegrown toilet champions American Standard and Kohler—though Toto is doing better and better as Americans start to fall in love with the bidet-toilet—but globally Toto is the world’s largest manufacturer of toilets and bidets. And in its home country of Japan, Toto is simply everywhere: 80 percent of Japanese homes contain a Toto bidet-toilet.

And if you’re a longtime Toto shareholder—maybe an investor with a particular interest in bathroom fixtures—this has been a wonderfully lucrative year for you. Toto’s stock is up 60 percent year to date; in just the last few weeks, it’s risen by 30 percent. Toto is doing better than ever: its net profit, in the first quarter of 2026, was up 230 percent year over year.

But Toto’s remarkable year doesn’t have much to do with toilets or bidets. Toto might have been founded in the 1910s to “provide a healthy and civilized way of life” through affordable toilets, and in the decades since might have become the global leader in the bathroom game. But Toto also does a lot of other things. Toto manufactures not just bidets and toilets but also bathroom tiles, prefabricated bathroom modules, faucets, modular kitchens, photocatalytic coatings for buildings, and assistive equipment for the elderly. And, most importantly, Toto has a very lucrative sideline in the fabrication of memory chips.

Since 1988, in a once-obscure corner of the company called the “advanced ceramics division,” Toto has been producing a very particular component called the electrostatic chuck, or the “e-chuck.” The e-chuck is a sort of high-precision ceramic plate, about the size of a steering wheel, that uses electrostatic force to hold a silicon wafer perfectly flat and thermally stable while memory chips are etched into it with bombardments of plasma. Making these components is extraordinarily difficult, since the ceramic body needs to have near-zero particle generation and be polished to submicron flatness: and this means that there are only a few companies in the world that are capable of manufacturing e-chucks reliably. Almost all of them—Shinko Electric, NGK, Toto, Kyocera, Sumitomo Osaka Cement, Niterra—are based in Japan.

For most of its history, the advanced ceramics division was a rounding error on Toto’s balance sheet: the money maker, as it had been since the 1910s, was the toilet and bidet business. But we’re in a new era. Demand for AI is exploding, meaning that demand for the high-bandwidth memory that AI data centers require is exploding, meaning that demand for memory chips is exploding, meaning that demand for e-chucks is exploding. And so Toto’s advanced ceramics division is suddenly the company’s largest business, generating the majority of its operating profit. Toto’s leadership, suddenly awash in AI-driven revenue, announced that they would double down by investing hundreds of millions in expanded electrostatic chuck production: the toilet company had become, quite unexpectedly, a supplier to the semiconductor supply chain.

The Toto story is a fun and interesting illustration of corporate diversification and how strange bets can pay off. But that type of diversification—a toilet company that also produces photocatalytic coating and high-precision components for semiconductors—isn’t really unique to Toto. Practically every company in Japan seems to do a thousand very different things.

Consider, for example, Kyocera, another one of the e-chuck makers. Kyocera was founded in 1959 as a producer of ceramic insulators for cathode-ray tubes; today it manufactures not only industrial ceramics but also printers, smartphones, ballpoint pens, kitchen knives, solar PV modules, lens components, industrial cutting tools, automotive camera modules, electronics components, semiconductor packaging, biocompatible tooth and joint replacements, UV-LED curing systems, LCD systems, medical products, and lab-grown gemstones. Or another e-chuck maker. Sumitomo Osaka Cement, as you might have been able to deduce from the name, produces cement and ready-mixed concrete; but it also produces optical components, measuring instruments, industrial ceramics, artificial marine reefs, cosmetics and nanoparticle materials.

And this degree of diversification extends to many of Japan’s most famous companies. Yamaha, for example, manufactures pianos, motorcycles, guitars, drums, boats, snowmobiles, ATVs, audio equipment, golf clubs, tennis rackets, home appliances, specialty metals, molding and bonding equipment for semiconductors, and industrial robots. Hitachi makes nuclear reactors, power grids, railway systems, elevators, semiconductor manufacturing equipment, medical imaging devices, data storage, IT consulting, and industrial machinery. Even a company as simple as Oji, Japan’s largest paper company, has been drawn into the production of disposable diapers, functional films, adhesives, cellulose nanofibers, and wood-based EUV photoresists; and it also operates a hotel, an airport catering business, a concert hall, and an insurance agency.

All of which is to say: Japanese companies do a lot of things.

There are, of course, other countries with companies that “do lots of things”: much of Indian economic life, for example, is defined by the sprawling activities of a few large business clans—the Adanis, the Ambanis, the Tatas, the Birlas. But India is a relatively poor country with a low level of economic specialization, and the sprawling conglomerates that dominate its economy focus on relatively simple things like cement, steel, ports, and telecommunications. Japan, by contrast, is a wealthy, developed society—by one measure, the most economically complex country in the world. What’s striking about Japanese companies is not that they do lots of different things but rather that they do them very well. There are all sorts of high-precision inputs—the e-chuck being just one example—that are produced virtually only by Japanese firms.

This is very different from how most wealthy countries operate. American firms, for example, tend to prioritize focus above all else: it would be bizarre for an American paper mill to also operate a concert hall and an airport catering business, or for American Standard or Kohler to somehow have something to do with semiconductors. Even a country like Germany, which matches Japan in its depth of high-precision firms, has nothing like Japan’s corporate diversification. Only a few large conglomerates, like Siemens, have anything approaching the lateral breadth of the Japanese firm. South Korea—whose economic system was not coincidentally modeled off the Japanese one—does have a few chaebol conglomerates, like Samsung and SK, that truly do as many things as Japanese companies. But these are economy-dominating, state-entangled megafirms, cultivated as national champions by Korean industrial policy. They look nothing like, say, Sumitomo Osaka Cement, which is hugely diversified despite being relatively small. (“Look what they need to mimic a fraction of our power!”)

So why are Japanese companies like this? Why do they do so many different things? And how do they manage to do so all those different things so well?

Here is the answer I want to suggest: Japanese companies excel in lots of very different domains because it’s inherent in how they’re structured. The form of the corporation that we know and love in the United States—specialized, market-oriented, governed by shareholders—is just one form that the corporation can take; but it’s not the only way to coordinate capital and labor in a successful and profitable way. The protean corporations of Japan are best understood as a different species of thing altogether: better at some things, worse at others, but still highly adapted to their particular environment. And the things that they’re very good at turn out to be extraordinarily helpful for all sorts of things in which American companies tend to struggle.

To see why, we need to learn a little bit about the economics of industrial organization.

Companies are bundles of practices

In 1990, two economists—Paul Milgrom and John Roberts, both of Stanford—published a paper called “The Economics of Modern Manufacturing.” You should forgive them for the rather bland title. It was a very interesting, and very influential, paper.

Milgrom and Roberts started out by noting that manufacturing was “undergoing a revolution.” One paradigm of production was getting swapped out for another. In the past, there had been the “Fordist” paradigm: the factories that worked in this paradigm had long assembly lines of standardized goods, large buffer inventories, narrow and repetitive jobs for their workers, and dedicated single-purpose machinery. But that approach was being superseded by a new model: “a vision of a flexible multiproduct firm that emphasizes quality and speedy response to market conditions while utilizing technologically advanced equipment and new forms of organization.” This was the “post-Fordist” vision. In practice, this meant shorter production runs, rapid changeovers between products, smaller and more frequent deliveries from suppliers, workers trained to operate multiple machines and diagnose problems on the fly, and quality control embedded at every stage of the process. It was an entirely different way of producing things.

The question that Milgrom and Roberts wanted to answer was simple: why did all of these changes come as a package?Maybe it made sense for a specific firm to adopt shorter production runs; but why did it also make sense for them to do everything else in the “post-Fordist” category? Why did the changes seem to be so tightly clustered, with firms either having none of these practices or having all of them?

The explanation that Milgrom and Roberts offered was that the practices were complementary. Adopting any one of the “post-Fordist” practices raised the returns to adopting others, such that adopting only one of the practices didn’t make nearly as much sense as adopting the entire set.

Milgrom and Roberts formalized their argument using the mathematics of supermodular functions. But you don’t really need to know anything about math to understand the idea intuitively.

Here’s an illustration. Let’s say you run a factory. You decide that you want your lines to produce fewer defective goods: maybe you want to improve your yield from 95 percent to 98 percent. So you decide to invest in better training for your workers: maybe training now lasts six weeks instead of two weeks. This works, and now your yield is higher; but that change makes other things more attractive too. For example: now that your yield is higher, it makes sense for you to reduce your inventory, since fewer defects mean you no longer need a large buffer of spare parts to replace the bad ones. So now you’ve cut your inventory: but now it makes sense for you to shorten your production runs and switch more frequently between products, since without a mountain of inventory to work through you can afford to change what the line is making. And if you’re switching frequently between products, then it makes sense for you to invest in flexible, reprogrammable machinery instead of dedicated, single-purpose equipment. So one relatively small tweak shifts the entire calculus of what you do.

In short: each practice makes the others more valuable, and each practice is valuable because it’s implemented alongside other complementary practices. Doing just one of these things—investing in flexible machinery, for example—doesn’t really make sense alone. The practice needs to work well with all the other practices that you have.

So the correct way to think about organizational practices, Milgrom and Roberts suggested, was as bundles. A complete bundle of practices was worth more than the sum of its parts; and each part was worth less in isolation than as part of a bundle. So there was a coherent “Fordist” bundle of practices, and a coherent “post-Fordist” bundle. But there wasn’t much in between.

“The Economics of Modern Manufacturing” turned out to be the cornerstone paper for an entire paradigm of thinking about firms and how they work. (Milgrom won the Nobel Prize in 2020, though mainly for his separate work on the theory of auctions; you can watch a delightful video where he’s woken up in the middle of the night by his neighbor and corecipient, Robert Wilson, because he was sleeping and didn’t answer the call from the Nobel committee.)

Most importantly, the Milgrom-Roberts framework gave a strong answer to the question of why firms are the way they are, and why it’s so hard for them to change. A firm that uses one coherent bundle can’t easily move to another: changing one practice without changing the others will typically make the firm strictly worse off.

So if we want to know why Japanese companies have one apparently unusual practice—why they’re so diversified into countless unrelated industries—we can’t really answer the question in isolation. We need to ask which bundle of practices they employ.

And luckily for us, people have looked into this question. The central figure here is the economist Masahiko Aoki, who taught at Stanford alongside Milgrom and Roberts and worked closely with both of them. Through the papers that their collaboration produced—some by Milgrom and Roberts, others by Aoki alone—we can sketch a picture of what the Japanese corporation is, and why it works the way that it does.

Japanese companies are nothing like American ones

The first thing we should note is that Japanese companies do a lot of things differently from Western companies.

The most important of these, by far, is lifetime employment. Japanese firms tend to hire only at the very bottom, plucking new recruits straight from high school or university; they have all of those new recruits start on the same day of the year (the first of April); and they generally expect to keep these employees until they retire. Mass layoffs are essentially unheard of. Even in times of acute distress, a Japanese firm will go to great lengths to find its employees positions at smaller affiliates rather than releasing them onto the labor market. And individual performance isn’t really a huge criterion in someone’s career. Promotions are based largely on seniority; pay differentials between ranks are modest; and bonuses are tied to the performance of the firm.

Because they work for the same company for their life and socialize largely within that firm—nomikai drinking parties with colleagues are part of everyday corporate life—Japanese workers are often deeply attached to their company. Some employees even wear lapel pins to indicate their corporate loyalties. (For a time employees also sang corporate anthems, though that tradition has faded.) There are unions, but they’re organized within the firm: rather than a “national autoworkers’ union” that organizes in both Toyota and Honda, there is a “Toyota union” and a “Honda union” that don’t have much to do with each other.

And this means that Japanese companies strive to avoid financial pressure from outsiders. Relationships with suppliers are longstanding and entrenched: many Japanese companies have been working with the same suppliers for 50 years or longer. Outside investors seeking to interfere in this happy picture will find few avenues for influence. A standard Japanese firm’s board of directors is composed almost exclusively of the firm’s own senior managers; a large fraction of the firm’s equity is held not by outside investors but cross-held by other Japanese firms; and most of the firm’s financing comes from a single “main bank” that provides loans and monitors performance.

And as a result, Japanese companies don’t really try too hard to return profits to shareholders. Earnings are mostly reinvested, and investor dividends are kept low. For a long time, Japanese firms would spend as much entertaining the managers of other firms as they would on dividends to shareholders.

The crucial thing, Aoki suggests, is that we understand all of these distinctive features—lifetime employment, no benefits for individual performance, hostility to outside financing—as reflecting a particular bundle: a “J-firm” bundle, as he calls it, as opposed to the “H-firm” bundle that you encounter in the United States or Europe. The core difference, Aoki says, is that while in the H-mode production is organized vertically, in the J-mode it’s organized horizontally. (H is for hierarchy; J is for Japanese.)

Consider, for example, the famous “Toyota Production System,” the philosophy that determines how Toyota makes its cars. In a Toyota factory, there’s a rope called the andon cord that runs along the assembly line, within the reach of every worker. Anyone who spots a defect—like, say, a misaligned door seal, or a bolt torqued to the wrong specification—can pull the cord and halt production at any time; once they’ve pulled the cord, the workers and team leaders closest to the problem will converge and try to solve it on the spot. In an H-firm factory, by contrast—you can think of a classic Ford plant here—defects are reported to a line manager, who will make a report and send it up the chain of command, and the higher-ups will solve the problem.

The andon method is really the J-mode in miniature. Information flows laterally, authority to act is widely distributed, and the people closest to the problems are the ones who fix it. And one result of the Toyota-style approach is that Japanese automakers have produced fewer defective cars than their American competitors for a very long time.

But Aoki points out that the horizontal coordination embodied by the andon cord doesn’t work without other practices as well. For example: horizontal coordination requires that workers know each other’s jobs, since a worker who spots a problem in one area of the line can only act on it if he understands what that area is supposed to be doing. But in order to understand each other’s jobs, workers cannot be specialized: they have to rotate across different workplace functions to the point where they’re familiar with much of the plant’s operations. In order to rotate across different workplace functions, they need broad training; and it makes no sense to train them broadly if you don’t keep them for a very long time. And if you have generalist workers who are around for a long time, you can’t reward them based on how they do in one role, because then they’d have no desire to leave that role for another role where they might do worse. Instead you have to pay them based on company performance, and promote them based on seniority. And you also have to give them an ironclad commitment not to fire them if economic conditions worsen: if they can get laid off at any moment, why would they invest years of effort in learning all the idiosyncratic things that your firm does?

So now you have a firm that has lots of lifetime employees who can’t be fired, and whose skills are tailored to what your firm needs rather than to a particular occupational category transferable to any employer. That works very well for your company’s employees; but it makes no sense to outsiders. So the system only makes sense if the company is also insulated from outside pressure, whether from organized labor or from organized capital. Thus the other features of Japanese corporations: firm-level unions, insider-dominated boards, and broad hostility to outside capital.

So something as apparently simple as horizontal coordination only makes complete sense once coupled with an entire bundle of different things. That’s why attempts to install the andon cord and other aspects of the Toyota system in American car factories have generally produced mediocre results. American automakers noticed the superiority of Japanese cars a long time ago, and tried to implement the andon cord: but it just didn’t work with how their companies were organized. In 2007, workers at a Toyota plant in Kentucky pulled the andon cord 2,000 times per week; workers at a Ford plant in Michigan pulled it just twice a week. You can’t get all the benefits of a single practice without installing the complete bundle.

And the complete Japanese bundle, I should say, ends up producing something with entirely different objectives and interests than the American bundle. The H-firm exists to make money, or rather to return money to shareholders; but the J-firm, run by its employees and largely indifferent to the interests of shareholders, exists simply to continue existing. That’s why Japanese companies are so protean and willing to change what they do. Nintendo was founded in 1889 as a maker of handmade playing cards; in the 1960s, it was pushed out of the playing cards game by a wave of competition; and it spent several years experimenting with new markets—taxi services and instant rice, though contrary to the rumors not love hotels—before finding its way to video games. Fujifilm, which faced a near-total collapse of photographic film in the 2000s, simply used its expertise in chemical coatings and fine optics to pivot into cosmetics, pharmaceuticals, LCD films, and semiconductor process materials.

And that basic impulse toward survival is why Japanese companies are so insistent on diversification. If you’ve made a commitment to keep people employed for life, then you need to create jobs for them if their current jobs stop making sense: indeed, you might need to keep them employed even if you can’t find anything for them to do. If you’re not very worried about profitability, and have lots of well-trained generalist employees, then it makes perfect sense to reinvest your company’s earnings by expanding into new industries: doing so not only allows your company to survive longer—your company’s portfolio of bets is now more diversified and thus lower-risk—but also ensures that you’re able to keep your surplus workers busy in one way or another.

Bundles are hard to make and hard to unmake

If bundles are self-reinforcing, then the bundle, once established, will be very hard to dislodge. The only way to get from one peak to another is to change many things at once: and that kind of wholesale transformation almost never happens under normal conditions. It only happens during moments of acute crisis which make necessary a wholesale transformation in how things are done.

In the case of the Japanese bundle, that moment of acute crisis was the Second World War.

In the 1920s, the Japanese economy looked, on a structural level, quite American. It was already an industrial society—not quite a leading industrial power, but by far the wealthiest country in Asia—and it already had shipyards, steel mills, stock exchanges, and a growing electrical machinery sector. Heavy industry was dominated by a few family-owned conglomerates, the zaibatsu; but they operated more or less as normal firms, raising capital on public equity markets and operating under shareholder discipline. Workers, for their part, moved freely between firms and organized labor unions.

But in the 1930s and ‘40s, as Japan mobilized for total war in Asia and the Pacific, that system was reworked entirely. Total war required the rapid expansion of arms production, which meant channeling virtually the entirety of economic production into heavy industry. Japan became a “national defense state.” Capital was rerouted out of the equity markets and into the banking system, where it could be allocated under state supervision; firms were instructed to prioritize employees over shareholders in order to maximize production; wages were standardized by seniority to suppress bidding wars for skilled labor and keep workers in place. The economist Yukio Noguchi calls this planned economy “the 1940 system.” The entire point was to orient every aspect of economic life toward maximum production at all costs.

Of course, Japan wasn’t alone in that regard. Every major belligerent in the Second World War adopted some version of a production-oriented planned economy characterized by some kind of financial repression. But in Japan, the 1940 system outlasted the war by decades. Japan was defeated in 1945, and occupied by the American military until 1952; but after an abortive attempt to reorganize Japanese economic life, the Americans decided that the Cold War instead demanded the strengthening and entrenchment of the Japanese system. (This was dubbed the “Reverse Course.”) And the 1940 system, in its essence, survived: and the Japanese firms that emerged in the twentieth century were the result.

And this system, as it turned out, was really good at particular things. Aoki’s key insight was that the J-mode had a comparative advantage in environments of moderate volatility: situations where conditions changed frequently enough that rigid central plans would be outdated before they were executed, but not so radically that only top-down strategic intervention could cope. In an environment of stable, predictable demand, the H-firm did fine; in an environment of extreme disruption, where the whole product line had to be rethought, centralized authority was indispensable, and the H-firm also did fine. But in between—where the challenge was to make constant small adjustments in a changing but recognizable paradigm—the J-firm excelled.

And this was exactly what Japan needed. The postwar challenge was catch-up growth: Japan had to grow fast, and to do that it had to absorb and improve upon technologies that the West had already pioneered. J-mode firms—with their collaborative cultures, deep pools of broadly trained workers, culture of incremental shop-floor refinement, and large pools of patient capital—were perfectly suited to the task. They could throw enormous amounts of patient capital at a problem, spend years refining a process without any imminent expectation of profit, and keep hundreds of broadly trained workers iterating on the shop floor until the quality of the output was world-class. And since profitability was never the primary objective, there was no pressure to abandon a difficult market for an easier one.

And the results truly were remarkable. By the 1960s, Japanese firms had begun to displace American ones in countless manufacturing sectors, from automaking to television manufacturing. Soon Japanese manufacturing was the envy of the world. Between 1946 and 1986, Japanese real per capita GDP grew tenfold, one of the highest rates of growth in recorded history.

But catch-up growth, by definition, has to end: at some point you’ve caught up, and the challenge at the frontier is not only to refine what’s already known but to invent what is not known. And paradigm invention is precisely the sharp discontinuity for which the J-mode has no particular gift. Consensus-driven, horizontally coordinated organizations are very good at refining what already exists: but they are very bad at deciding what should exist.

That basic weakness is why Japanese firms are so dominant in some domains and entirely absent in others. Japan excels in automotive manufacturing, machine tools, industrial robotics, optics, and precision materials: domains characterized by incremental refinement. But they have very little to add in software, internet platforms, artificial intelligence, or electric vehicles. The architecture of the Japanese firm is built to perfect a domain through progressive advancement; it’s quite poorly suited to sharp discontinuity.

Consider Sony, which by the 2000s manufactured the world’s best portable music players, its best small cameras, its best mobile displays, and its best lithium-ion batteries: every component of what would become the smartphone. Purely on a material basis one would expect that Sony was the best-positioned company in the world to make the smartphone. But Sony didn’t do it. It was Apple, an H-firm par excellence, that reimagined the entire product category from the top down, largely because Apple was organized to give extraordinary power to a single visionary leader.

By the time that Aoki, Milgrom, and Roberts were writing in the last few years of the twentieth century, the shine of the Japanese model had already begun to fade. Asset prices in Japan had begun to deflate in 1990, inaugurating the country’s “lost decades.” Firms that had balanced against their assets at inflated prices now had more debt than they were worth, and the closely-affiliated banks that had lent them the money were now buried in so much bad debt that marking the loans to market value would have destroyed both the companies and themselves. Bankruptcies and mass layoffs were impossible in the Japanese system: a wave of mass layoffs or corporate restructurings would have undermined the entire social settlement that governed Japanese life.

So debts weren’t called: banks and companies simply soldiered on, “zombies” suspended in a state between life and death. Japanese business was no longer the envy of anyone in particular.

The Japanese bundle was exceptionally good, indeed world-historically good, at catch-up growth; but it was very bad at figuring out what to do once it found itself in trouble. Organizational bundles are remarkably resistant to change, even as conditions themselves change.

This is exactly what those who tried to reform Japanese corporate life in the last few decades have discovered.

In the 1990s, Fujitsu and other electronics firms experimented with performance-based pay: the idea had worked well at American firms and seemed like an obvious way to make Japanese workers more productive and thus to get the economy out of its slump. But performance-based pay didn’t cohere at all with the rest of the Japanese system. Team cooperation broke down, because output was measured individually and helping a colleague now hurt you in the rankings; senior engineers stopped mentoring juniors, because mentoring was uncompensated and the mentored juniors became future rivals; and managers struggled to keep their teams from disbanding. By 2001 Fujitsu had abandoned the practice. The episode became so infamous that one former Fujitsu executive wrote a book about it, titled The Downfall of Performance-Based Pay at Fujitsu as Seen by an Insider.

This is exactly what the economics of industrial organization would predict. High-powered individual performance pay makes sense when jobs are narrow, tasks are clearly measurable, and cooperation is inessential; but it makes no sense within the bundle that defines the Japanese firm. The same goes for practically every institution: piecemeal changes to a coherent bundle of organizational practices don’t really work; they only make things work less well. A reform that moves one coordinate but leaves the others in place produces a kind of organizational chimera, an entity that has lost the coherence of its old bundle without gaining the benefits of the new one.

But the Japanese bundle, however antiquated it might seem, still does result in some of the most remarkable companies in the world. The type of deep process knowledge that has accreted within companies like Kyocera and Toto is almost impossible to replicate. The American bundle of practices, with its emphasis on profits, entrepreneurship, and financialized risk, is probably the world’s best at innovation and frontier discovery. But as we are now discovering with the global rush on memory chips and other esoteric parts of the semiconductor supply chain, our entrepreneurial American system only works completely if it’s paired with a very non-entrepreneurial system like the one that we find in Japan.

This Substack is supported by readers like you.

Nobody Pushed Back: Why Engineers Stay Silent Until It's Too Late | How to Center a Div

Mon, 18 May 2026 12:19:58 +0000

TLDR: Most architectural disasters aren't a knowledge problem. The engineers knew. Speaking up just wasn't worth it.

Someone walks through an architectural decision and nobody in that room actually agrees — they just act like they do, because saying what you really think is socially expensive. Meeting ends, decision gets made. Six months later production blows up and everyone says "we knew this would happen."

Yeah. You did.

Most companies don't collapse because someone made a bad call. They collapse because the people who saw it coming kept their mouths shut. And the reason they kept their mouths shut has nothing to do with technical ignorance — it's that speaking up cost more than staying quiet.

The Pattern

Every major architectural disaster has the same structure underneath: there's a technical, visible problem, someone — usually more than one person — can see it, but pushing back costs something, so the decision passes under the name of "alignment" and eventually the system breaks.

Pay attention to that word — alignment. In most companies, alignment is just the corporate name for silencing dissent. It doesn't mean everyone agrees. It means nobody says out loud that they don't. Those are different things.

Nokia, TSB, Boeing, Microsoft

Nokia's engineers knew Symbian was a sinking ship. Not built for touchscreen, fundamentally wrong architecture for building an app ecosystem. When the iPhone launched, the sharpest read on what it meant came from inside Nokia. INSEAD researchers went back and interviewed 76 Nokia executives and engineers after the collapse; what they found was that people knew, they just didn't say so — because at Nokia, being the person who brought bad news upward was a career risk, not a career move. The information existed. It just never traveled. Nokia sold its phone division in 2013 for $7.2 billion.

TSB Bank migrated its legacy IT infrastructure in 2018 with a "Big Bang" cutover — one shot, clean break. The independent inquiry report runs 262 pages, and one sentence stands out: technical objections were raised, but "not taken into account." The go-live schedule won. The system collapsed, 1.9 million customers couldn't access their accounts, regulators handed down a £48.6 million fine. When the objection wasn't taken into account, did anyone push back a second time? A third? Nobody wrote that part down.

In 2020, Boeing's internal messages made it to Congress. Engineers had been writing to each other: "This airplane is designed by clowns, who in turn are supervised by monkeys." Nobody's asking what they wrote to management — we already know the answer. MCAS was pulling data from a single sensor, and the Congressional report documents that the technical risk was known — and that production schedules and cost pressure buried it. The full story involves FAA certification dynamics, supplier relationships, and decades of cost-cutting; it's not reducible to a single cause. But the pattern holds: the people closest to the problem knew, and that knowledge didn't travel up. 346 people died.

Microsoft built Windows Phone on the Windows CE kernel, then changed the architecture entirely, then existing devices couldn't receive updates, then the ecosystem collapsed. The engineers on the mobile side could see where it was heading; after the Nokia acquisition, an Android-based alternative was even prototyped internally — the Nokia X project. Management called it "disloyalty to the Windows vision" and killed it. Same story: architectural dogma, suppressed pushback, platform death. $7.6 billion written off.

Four companies, four industries, same mechanism. The problem wasn't technical — it was that nobody could speak freely.

Why Nobody Speaks

"Why didn't anyone speak up?" is the comfortable question. It puts the blame on individuals and lets the system off. The harder question is: what happened to the last person who did?

In companies, pushing back comes with a label. "Not a team player." "Always has objections." "Negative energy." An engineer who experiences this once doesn't put themselves in that position again; an engineer who watches it happen never does at all. And most engineers know this perfectly well — they just call staying quiet "professionalism." That second part matters, because over time this silence stops being a necessity and becomes a habit. The system squeezes, and people bend — and then they start telling themselves they chose to bend: "not my job," "they know better," "nobody's going to listen anyway."

Then there's HiPPO — Highest Paid Person's Opinion. The most senior person opens their mouth and the room folds, regardless of whether what they said made any sense. Everyone nods. They call it alignment. It's surrender with better branding.

Then there are metrics. A/B test says "the popup gets more clicks," conversion's up, dashboard's green — discussion over. Metrics in most companies don't measure reality. They close arguments.

Being right in a company and being safe are not the same thing. If those two things have drifted that far apart, the decisions being made are no longer driven by technical reality — they're driven by the social dynamics in the room.

The Cost of Silence

When the first bad decision goes through unchallenged, the next one is easier — not because people stop caring, but because the precedent is set. You learn what this place does with objections. "That's just how we do things here" isn't a policy. It's scar tissue. The new engineer sees the same dynamic, does the same thing, doesn't question it — questioning is weird because nobody questions anything. Calling this technical debt is wrong; technical debt is a money problem. This is learned helplessness. Silence eventually becomes a technical property of the system, built in, impossible to refactor out.

The Boeing engineer wrote "designed by clowns" to a colleague, not to management, because writing to management didn't feel possible. That knowledge never left the company — it circulated internally, accumulated, and eventually leaked.

What Pushback Actually Is

It's not saying "this is wrong" — that doesn't work. Real pushback is making problems visible: putting a price on the decision, naming what can go wrong, making the tradeoffs concrete.

"What does this decision cost us in 18 months?" "How are we handling this risk in testing?" "What's the rollback plan if this goes sideways?"

None of those sound like an attack. But they force the decision to justify itself, and they give cover to everyone else in the room who was thinking the same thing.

What companies need isn't better engineers or better architecture. It's an environment where someone can say "this is going to go badly" and not get punished for it. That environment doesn't build itself — and it doesn't appear by asking individuals to be braver. It gets built through explicit decisions: separating the decision from the person who raises the objection, making postmortems blameless by default, treating "I was wrong" as a signal of good judgment rather than weakness. Amazon's "disagree and commit" gets cited a lot here — and it's not perfect, the "commit" part sometimes just overrides the "disagree" — but the underlying idea is sound: you can record your objection, have it heard, and still move forward without pretending you agreed. That's a structural thing, not a personality thing.

Somebody implemented the 10 popups you close every time you open an app. Somebody moved that button to the wrong place. Somebody pushed TSB's system live on that timeline. In every one of those rooms, there was someone who knew the decision was wrong — probably more than one person.

It's not that nobody knew. Everyone knew. Speaking up just wasn't a rational choice.

diminished sympathy for the depressed | vishalblog jr.

Sun, 17 May 2026 03:43:34 +0000

diminished sympathy for the depressed

15 May, 2026

for the last two years i have been mentalizing quite a lot, and i now can confidently conclude that this habit has not expanded my empathy. to the contrary, i've seen what's going on in people's minds, and i don't like it.

i am, in some ways, a highly sympathetic audience to those who are hurting. i like to listen, i have a good memory, i like to hear authentic speech (though depression-talk is rarely authentic -- more on this later), and i spend a lot of time imagining, in the style of father brown, the internal states of those around me, and how such things relate to actions done. if you have something delicate you need to express to someone, and you choose to express it to me, i am always grateful and touched to hear it.

i also find it hard, in the final analysis, to really censure people. i have managed to maintain friendships with compulsive liars and sex addicts, for example, because after a certain point i remind myself that some people just can't help themselves. some things are in the genes / in the blood / in the soul / in the fabric of reality itself. there is a certain pleasure in learning to take the bad with the good.

that being said, my sympathy for depressed people has lately attenuated radically. i don't mean to imply that being depressed is a failure of character or of morality; i only mean to say that the presence of depression is neutral with respect to character. polite social scripts dictate that a depressed person is to be extended a lot of gushing sympathy. i now believe this reflex is wrong and harmful. there are many, many malicious and deleterious people in the world who have learned to game this instinct.

accordingly, i have compiled a list of dangerous depressive-type archetypes worth being on the lookout for.

those who see everyone as worthless

in high school i had a friend who both hung out with me and with a popular set filled with people who we both agreed were a bit vacuous. near the end of our senior year, i sat with this friend in a park and scolded him for spending too much time with the stupid-crowd and not enough time with me. after listing out the flaws, person-by-person, of this set for ten minutes, my friend stopped me and said, "i'm not disagreeing with any of this, but maybe you've noticed i'm spending time with you, and not them, right now?" this reply produced a spike of shame so acute it reorganized my consciousness. though i was not jealous of this "set," they were all good and interesting people, and my inability to recognize this revealed a limitation in my character. i turned this limitation into my friend's problem, by doing splash damage to him.

i now see this pattern as belonging to a genus: when one's self-esteem is so low you devalue those around you. an obvious mistake, you'd think, but there is a certain type of depressed person who does this all the time. it almost as if they're thinking, "if these people are spending time with me, they must be just as worthless." of course, no one actually believes this explicitly, but the pattern of behavior is common enough. for instance, i had a friend group which would meet every evening to chat and game, and one particularly depressed member of this group would occasionally lash out at everyone else with little jabs like, "wow, we're all such losers, spending all our time gaming..." meanwhile, i'd be thinking, "maybe you currently are a loser who does nothing, but i put in a full day at work, then set a deadlift PR, then did an hour of jiujitsu, then read a play, and now i'm gaming with you because i like you."

it's nice to believe that people who are depressed and low-status wouldn't be status-competitive, but this is far from true. some people are natural competitors who believe they're surrounded by a vast and lowing herd of human cattle. a subset of these people have been, by neuro-chemical happenstance, brought low by depression, but without developing apposite humility. linton heathcliff, from wuthering heights, is a model for this type: a mewling weakling who has been relentlessly brutalized, but nevertheless is himself a bully, who spends all of his time mocking and humiliating his illiterate cousin, hareton. linton was a man-child who couldn't have known better -- what's your excuse?

those who cry-bully

a cousin to the previous category. mostly self-explanatory. a depressed person who likes to dish it out but can't take it. everyone just has to stand around and nod politely while getting insulted, unless they're willing to risk watching this person have a mental breakdown.

last time i was exposed to a person of this type, i sat around and listened to them belittle all of our mutual acquaintances, while we both pretended like it wasn't perfectly obvious to me that this person mocks me, as well, in different company. the two most prominent examples of this type i know personally are (i'm sorry to say) women who think they are exceptionally empathic. there's nothing to do with this sort except avoid them. very bad when given authority over others.

those who amplify discord

this is very subtle, and is difficult to talk about across gender lines.

everyone in my family has sub-clinical bpd. by this i mean we form strong attachments very quickly, and then feel betrayed very easily. this unfortunate whippiness hasn't caused us to materially damage our lives (hence "sub-clinical"), but it's a very noticeable part of our personality structure. when i'm crashing out, i do something i call "discord amplification," which looks like this: someone says something i don't like; then i think about why what they said is wrong, and how what they said was poorly observed/argued; then i start asking myself trap questions like, "why did they think it was okay to say that to me in the first place?", and "what did they really mean by that?", and "if they really cared about me, wouldn't they have said X instead of Y?"; then i start enumerating the flaws in their character. and with that, we're off to the races. the discord feeds off itself. i have learned (i think) to mitigate some of this as an adult through certain strategies and mental exercises.

some people, particularly certain depressive-types, do just this sort of thing, but on the social level. perhaps your boyfriend said something you don't like, and then your discord-amplifier friends sit you down over brunch and start asking you the trap questions i listed out above. as easily as that, a whole community can give itself sub-clinical bpd.

one hysterical/demoralizing example: i used to be a part of philosophy discord server, which was primarily a place to coordinate weekly reading groups over voice chat, but also saw a lot of off-topic bantering in various text channels. over the five years i rooted around that server, i saw the following happen three separate times: text-only transwomen would start fighting over some topic, then split into two tribes, and escalate and escalate and escalate their conflict until admins were forced to start banning users, which would prompt the losing faction to leave the server en masse. again, this happened three separate times. i remember the second schism erupted because a transwoman in genchat said "makeup is transphobic" and another transwoman said "saying 'makeup is transphobic' is transphobic," and they started amplifying so hard that most-all the server's transwomen took sides, and eventually the losing side left the server (i am not joking about any of this). discord on discord! because of the dynamics of amplification and identity, calm outsiders couldn't say the obvious things, like "you're all trans, obviously none of you are transphobic" or "please talk about philosophy instead."

we don't have the social infrastructure to prevent community-level discord amplification. i hear a lot of bad ideas, like "things will de-escalate if we make sure everyone gets heard." this is a doomed strategy because amplifiers love being heard! my friend had an ex who would get in literal 30-hour amplification arguments with him; to fix this, the ex proposed going to a couple's therapist; a potential therapist turned them down as clients, because he accurately identified that the ex really wanted the therapy to be a new venue to force her boyfriend to sit down and listen to her amplify some more.

my most recent strategy for preventing auto-amplification has been to tell a friend in person, "wow, i'm having an epic bpd crash-out right now," then go off to stare at a sack of potatoes or something for ten minutes, then come back and sit and listen to people talk for fifteen minutes or so. around this time, out of nowhere, i will feel cured. this has been very consistent, and has stopped me from losing days to rage. i have no idea what the community-level analogue to this looks like, however.

i am quite sad about discord amplification, because it is so hard to stop, but it is ubiquitous, since it is parasitical on useful social infrastructure (e.g. whisper networks). for instance, i had two friends, a man and a woman, who i knew would naturally like each other (how did i know that? because i told them all the same jokes), but they kicked off their acquaintance on a minor bump which they both amplified in separate, gendered ways. luckily they both came down to earth before a losing faction got exiled, but it was astonishing how unhelpful everyone around them was.

those who demand love

i have recently observed someone crashing out of a social scene due to the following parameters: they were a perfectly likeable person, but would remind everyone daily of how they don't have many friends and are looking for more friends, and how they were confused about why they didn't have more friends, and would cut conversations short to ask people "why aren't we closer friends? what is it about me that is preventing us from being closer friends?", which is just a really unpleasant set of questions to be on the receiving end of. there was nothing except time preventing those budding friendships from really sprouting, but for some reason this person demanded full-bore friendship real quickly. essentially, they were being a "friend-pest." unfortunately, there was something more than social incompetence at the root of this.

the wisdom of la rochefoucauld

all of these dynamics are described by some as "toxic." i don't describe them this way, because i think therapy-language is a huge trap. therapoids seek to create a sufficiently expressive language that will fix all these problem -- to separate that which is "toxic" from that which is "healing" -- but the language of self-love is always exploitable. think, for instance, of the infinity of perverse use-cases the concept of "healthy boundaries" can be smeared over. scott alexander observed that concept handles in self-help are usually created and spread by the least qualified:

Advice is disproportionately written by defective people. Healthy people perform naturally and effortlessly. You walk so gracefully that a million man-hours into bipedal robots fail to match your skill. But if some stroke patient or precocious one-year-old asked your secret, you would just say “I put one foot in front of the other.”

If you want good advice about how to walk, ask someone with cerebral palsy. They experience walking as a constant battle to overcome their natural constitution, and so accumulate tips and tricks throughout their lives. Or ask a physical therapist who works with these people and studies them. Just don’t ask someone you see walking especially briskly down the street.

Relationships work the same way. Go to an elderly couple who have been happily married for fifty years, and they’ll give you vapid old-person advice like “Treat every day as a gift from God.” But go to someone who’s struggled with every one of their last thirty-seven relationships, and they’ll be full of suggestions! They’ll tell you all sorts of fascinating things about boundaries and gaslighting and the four-hundred-and-ninety-four principles of nonviolent communication.

the problem with telling people to start by learning how to "love themselves" is that human beings are already stuffed full of amour propre. the older i get, the more truth i perceive in the maxims of la rochefoucauld:

1: What we term virtue is often but a mass of various actions and divers interests, which fortune, or our own industry, manage to arrange; and it is not always from valour or from chastity that men are brave, and women chaste.

4: Self love is more cunning than the most cunning man in the world.

11: Passions often produce their contraries: avarice sometimes leads to prodigality, and prodigality to avarice; we are often obstinate through weakness and daring though timidity.

31: If we had no faults we should not take so much pleasure in noting those of others.

86: Our distrust of another justifies his deceit.

90: In the intercourse of life, we please more by our faults than by our good qualities.

146: Usually we only praise to be praised.

241: Flirtation is at the bottom of woman’s nature, although all do not practise it, some being restrained by fear, others by sense.

i hated this stuff when i was younger, because i thought it was all hyperbole and cynicism. of course it is hyperbolic, but i do not think it is cynical. understanding that a person can be both a depressive and a horrible bully is, when viewed from the right angle, actually a pretty funny and endearing thing to know. i find i like people more now. i am, however, much more cautious about expressing sympathy.

Can America's Trains Handle the World Cup? - The Atlantic

Wed, 13 May 2026 21:01:08 +0000

The planet’s biggest sporting event, the World Cup final, will take place this summer in MetLife Stadium, which is presently known as New York New Jersey Stadium because FIFA has strict rules on corporate branding. The stadium—whatever you want to call it—is located in the marshlands of New Jersey, about nine miles from Midtown Manhattan. On the day of the final, as on the dates of seven other matches throughout the World Cup tournament, an estimated 80,000 fans will converge at its gates.

But how will they get there? Some will drive, even though they’ll have to pay $225 to use one of the 5,000 available parking spots at a nearby shopping mall that is connected to the stadium area by pedestrian bridges. Others will buy a seat on a shuttle bus—originally $80, cut to $20 after last-minute maneuvering by New York Governor Kathy Hochul. (Some of these will be yellow school buses.) Or they will cough up whatever amount ride-share apps are charging on those days. And the rest—up to 40,000 people for each event—will take their chances on an infrequently used branch of New Jersey Transit that has struggled with large crowds in the past.

Read: A ‘Death Train’ is haunting South Florida

In the coming months, America’s patchwork railway system will be similarly challenged—and its weaknesses exposed—across all 11 U.S. sites of World Cup matches. In Dallas, most people who are going to the stadium will either have to pay for expensive parking or take a commuter rail to a charter bus. Kansas City will rely entirely on charter buses. Where direct rail access is available, the trains aren’t likely to be convenient, and tickets may be outrageously expensive. New Jersey is a case in point: Last month, NJ Transit announced plans to charge $150 for each round-trip journey on a route that would otherwise cost less than $13.

That price was later reduced to $105, thanks to donations from various unnamed companies, then reduced again to $98 just before tickets went on sale—but the fact of any of these fares suggests a deeper problem. NJ Transit President and CEO Kris Kolluri explained the dismal math behind this pricing at a press conference in April, alluding to the agency’s enormous debt and degrading equipment. To transport all of those people to the stadium, he said, the agency would need to spend about $6 million a game, mostly for labor and security, as well as for maintenance work on 50 railcars; this would include the purchase of new wheels, axles, and air-conditioning units “to make sure that we don’t have the challenges we typically do.” Such costs could be passed on to New Jersey taxpayers, Kolluri pointed out, but “no one that I have spoken to thinks that that’s (1) fair and (2) reasonable.” So instead, the agency has done some simple arithmetic: $6 million in operating expenses divided by 40,000 riders equals $150.

From the start, the situation has had all the makings of a political brouhaha. When FIFA complained that the fare was too expensive, New Jersey Governor Mikie Sherrill argued that the association, which stands to bring in $11 billion in revenue from the tournament, should subsidize or cover the fares itself. A FIFA official shot back that the hiked-up fares would “diminish the economic benefit and lasting legacy the entire region stands to gain from hosting the World Cup.” Then the New York Post’s editorial board took issue with NJ Transit’s plan to close off its section of Manhattan’s Penn Station for long stretches on match days, arguing that the agency was “dissing” its regular riders. Separately, Pennsylvania Governor Josh Shapiro boasted that fans could get to and from the matches held in Philadelphia using the region’s SEPTA rail system for just $2.90.

Kolluri said that NJ Transit’s special challenges justify the (much, much) higher fare. The Philadelphia stadium is in the city, for example, and SEPTA trains already go there every day. MetLife Stadium, however, has no regular train service. It “is a suburban stadium,” he said, which is “very different fundamentally.” Isn’t that the problem, though? Europeans have lately been wondering on social media why this stadium was constructed where it is in the first place—stranded miles from the city center and encircled by highways, parking lots and swamps—and nobody has been able to supply them with a good answer. It’s just how we like it!

Read: Airport chaos is leading people to ride the Amtrak

One reporter asked Kolluri about the 2014 Super Bowl, held in the same location, also with approximately 80,000 people in attendance. NJ Transit did not raise fares anywhere near as much for that game, he pointed out. “First of all, do you know what happened in the Super Bowl?” Kolluri snapped. “I think you’re the only guy who may not know what actually happened.” What happened was widely reported travel chaos: Long lines and delays, and at one point, a request that people stay inside the stadium until some portion of the crowd dispersed from the train platform. The event went so poorly that the agency commissioned an independent investigation of its failures. Kolluri described all of this as having caused “PTSD,” and said that the situation was a reason to do things very differently this year. “People think about that moment and say we can never let that happen again,” he said. (People did, in fact, let that happen again in 2019, when thousands of fans got stuck waiting for hours in the darkness for a NJ Transit ride after a WrestleMania event.)

The $150-a-ticket pricing, Kolluri argued, was only what would be needed to prevent catastrophe. “I think that’s a defensible claim,” says Zoe Baldwin, the vice president of state programs at the Regional Plan Association, a nonprofit focused on economic development and quality of life in New York, New Jersey, and Connecticut. “We have a very old system that is in desperate need of overhaul, let alone maintenance.” Equipment failures are more common in the summer, she told me, so NJ Transit will have to spend on back-up crews and engines in case any trains are taken out of service. She seemed optimistic about the agency’s ability to handle the tournament crowds, and she emphasized that the trip out to the stadium would be a great opportunity for people all over the world to get a look at one of the country’s biggest and most fascinating urban wetlands. When I asked her whether those same people might be horrified by the look they get at New Jersey’s tangle of unwalkable roadways and parking lots, she protested: “What are they going to think when they go to L.A., then?”

To her point, the most public drama over World Cup transportation until now has occurred in a region that has better public-transit options than any other part of the United States does. The railway infrastructure throughout the Northeast may be old and shoddy—for example, Amtrak service between New York and Boston was recently suspended because pieces of a highway on-ramp had fallen onto the tracks—but at least it exists. Just two World Cup host cities in the U.S.—Seattle and San Francisco—have an Amtrak station anywhere near their stadium. In Houston, where fans can take the city’s light-rail system, two of the relevant lines run only once every 12 minutes. In Los Angeles, the matches will be accessible via shuttle-bus service from designated Metro drop-off points. Even back East in Philadelphia, where SEPTA service goes directly to the stadium, the system will be strained: A spokesperson estimated that that line can transport 15,000 people an hour, but twice that many are expected to take a train to each match.

When I asked Jim Mathews, the president and CEO of the nonprofit Rail Passengers Association, about his impressions of the various host cities’ transportation plans, he complimented the Los Angeles strategy on the grounds that it would be affordable and temporarily link several independent transit systems. But he did not agree with the triple-digit price tag for NJ Transit rides, or the $80 fares for those who take a train from Boston to a match at Gillette Stadium. “You’re taking this moment when the spotlight of the world is on you, and you’re making it stupidly expensive,” he told me. “It just shows you what happens when you go for decades underinvesting in capacity.”

Mathews said he’s worried that visitors from overseas will be shocked when they arrive in the U.S. and get a look at its trains. Although some cities here now have more transit options than they did a few years ago, tourists may still be disappointed by the scarcity of options. And despite Americans’ dramatic increase in interest in soccer over the past three decades, he expected we’d be embarrassed on the field too: “We are still going to exit in the first round.”

The Most Emacs Bzr Saga | Thanos Apollo

Wed, 13 May 2026 17:52:59 +0000

Since I had no interesting books to read today, nor interesting films to watch, I decided to scavenge for the most intriguing content one can find online. I ended up reading the Linux kernel mailing lists, but those discussions seemed to be 18+, so I settled for the comparatively civil emacs-devel.

For those unfamiliar, emacs-devel is the primary development discussion list for GNU Emacs – where design decisions get made, patches get reviewed, and occasionally where people spend 200 messages arguing about version control software. This is the story of that last one.

2008: “This question is over and decided”#

In March 2008, Emacs was migrating from CVS (yes, CVS) to something more “modern”. The two contenders were Git and Bazaar.

Git, created by Linus Torvalds for the Linux kernel.
Bazaar was a GNU project, maintained by Canonical

A 236-message thread erupted on emacs-devel. People benchmarked both tools. The results were not subtle. Andreas Schwab, one of the core developers, reported his first impression:

My first impression is that bzr is slow, so slow that it is completely unusable. How can it come that a simple bzr log takes more than a minute to even start? Even cvs log is instantaneous in comparison, although it has to request the log from the server.

David Kastrup found it equally puzzling:

I find this surprising: “git log” is pretty much instantaneous, and git recalculates a code piece’s history in the process. In contrast, one has to tell Bazaar when one copies or moves or renames files, so it should have the information available right away.

The actual numbers:

git log | head -1 0.012 seconds. The same command with Bazaar took 21.5 seconds.
Committing a single-file change: 0.08 seconds with Git, 17 seconds with Bazaar.

The benchmarks kept coming. Stefan Monnier, the head maintainer, set the bar low:

I don’t care if Bzr is slower or faster than Git, but in order to switch to Bzr, we need it to be ‘fast enough’. Currently it is not. At the very least the ‘bzr diff’ should not take more than a couple seconds.

Meanwhile, Jonathan Lange, an actual Bazaar developer from Canonical, was in the thread doing tech support.

His recommended workflow for the initial checkout:

An even better way to do the initial download is this:
$ wget http://bzr.notengoamigos.org/emacs.tar.gz
$ tar xzf emacs.tar.gz
$ bzr init-repo emacs-bzr
$ cd emacs-bzr
$ bzr branch ../emacs trunk
$ cd trunk
$ bzr pull –remember http://bzr.notengoamigos.org/emacs/trunk/

Compare that to git clone!

Someone in the thread finally asked the obvious question:

To the emacs maintainers and decision makers: What more information is required to convince bzr is not the right tool at the present moment?

Richard Stallman’s reply:

This decision is not a decision for the present moment. It is a long term decision. So it would be better to wait a few months while Bzr developers improve it, than to make some other “temporary” decision that would probably be hard to reverse.

And in case anyone missed the point, in a separate message:

This question is over and decided. We will use GNU Bzr, because it is a GNU package.

When someone pointed out that this political decision was “wiping away all technical arguments,” RMS replied:

The rule that GNU packages should support each other helps make the GNU system as a whole work better.

Someone asked the obvious follow-up: “Why can’t we just make Git part of the GNU system?”

RMS:

We could include it in the GNU system, but its developers are not likely to want to make it a GNU package.

To be fair to RMS, there’s a real principle here. If the GNU project doesn’t use its own tools, it sends a message that those tools aren’t good enough, which undermines the whole idea of a self-sufficient free software ecosystem. He’d been making this argument for decades, and it had served the project well in many other cases. The problem wasn’t the principle. The problem was that Bazaar couldn’t live up to it.

The 236-message thread, the benchmarks, the Canonical employee’s workarounds; none of it changed the outcome. The decision was political, and the politics were settled.

2008-2012: The long tail#

The rest of the world moved to Git. GitHub launched in 2008 and exploded. Emacs contributors, meanwhile, had to learn Bazaar, a tool they used nowhere else, just to submit patches.

Threads like “Help me unstick my bzr, please” and “Can NOT bzr the emacs repos (may be bzr has a memory leak)” became regular occurrences.

Then in 2012, Canonical laid off the Bazaar development team.

2013#

In March 2013, a year after Bazaar’s development ceased, John Wiegley posted what everyone had been wanting to say:

We have often debated the merits of Git vs. Bazaar, and which one the GNU project should use for Emacs development. I think now is an appropriate time to revisit this decision.
My main reason for bringing this up again is that Bazaar development has effectively stalled. There are major bugs which have been in their bug-tracker for years now – bugs affecting Emacs development, such as the ELPA repository – which have been ignored all this time.
So, to Richard as the undisputed Czar of all things Emacs: can we now, pretty please, switch to Git?

200 messages followed.

RMS’s first response:

The maintainer says he is fixing some bugs, and I asked him just yesterday to fix the ELPA branch bug. I’d like to give him a reasonable time to do this.

The bug was 1.5 years old. Dmitry Gutov called it out:

Isn’t this a bit late? The bug is 1.5 years old. Was he not aware of it before?

Then RMS revealed his hand:

I am trying to determine whether Bzr is effectively maintained or not. I’d rather get a Yes answer than a No answer.

That’s a remarkably honest thing to say publicly. He wasn’t hiding his preference. He genuinely believed in the principle of GNU projects supporting each other, and he was hoping reality would cooperate.

Joakim Verona, a longtime Bazaar user and Emacs contributor, described the reality on the ground:

I have done my best to be a constructive user of the tool, and I have had many technical difficulties. When I try to find solutions to the issues I notice the following:
The bzr community is very helpful. This is good.
There are many well known bugs. There are also many well known patches for these, some of them provided by Emacs developers. They never enter upstream. By “never” I mean years. This is bad.

RMS replied:

I don’t have time to read the Bzr mailing list. Or any development mailing list. The only such list I am on is this one. You might as well tell me to fly to the moon as tell me to read something on the Bzr list.

And when asked whether the users of Bazaar should have a say in whether Bazaar is sufficiently maintained:

When I have to decide whether a maintainer is doing an adequate job or needs to be replaced, I pay attention to whatever relevant information I get. However, to give users “a say” in the decision seems improper, so I don’t do that.

Karl Fogel, a veteran open source developer (author of Producing Open Source Software and one of the original Subversion developers), delivered the sharpest critique in the thread:

Well, really, you don’t have time to pay close enough attention to Bzr development to competently decide whether it’s still a good choice for Emacs. That’s fine – no one has time to do every important thing, and you do many other important things.
But then why do you think you still have the time & mental bandwidth to make this decision well? Why not delegate it to the Emacs maintainers on the grounds that you no longer have time to do a good job of this evaluation?

He pointed out that asking one person about one bug is not a proxy for project health, and that others in the thread had already done more thorough research than RMS could, given his time constraints.

RMS’s reply:

Because more than Emacs is at stake here.

One line, but it’s the core of his worldview. If the flagship GNU project abandons a GNU tool, what signal does that send to every other GNU package? He wasn’t wrong about the stakes. He was wrong about the tool.

Karl pushed once more:

You should either devote enough time to evaluating Bzr’s maintenance state to get a reliable answer, or delegate to someone who can do so. Instead, you’re asking the maintainers to rely on your investigation… yet you clearly don’t have time to do a good job. This is a poor use of everyone’s time.

RMS:

I already have a plan for how to proceed on this, and I am doing it.

No details. No timeline. No delegation. Trust me.

Meanwhile, Stefan Monnier, posted exactly one substantive message in the entire 200-message thread:

Just like I didn’t fight Richard’s choice of Bazaar, I don’t care very much whether we keep using Bazaar or we change to Git, Monotone, Darcs, Mercurial, OpenCM, Fossil, younameit.
The only thing I care for now is to move away from Bazaar for the ’elpa’ branch because Bazaar can’t handle it properly.

Leo Liu summarized what everyone was thinking:

Most GNU projects aren’t using BZR as you might be aware.
While helping BZR fixing bugs might be a gain for BZR, it is a loss as a whole for GNU. Volunteers spend their spare time on GNU projects and if 20% of that time is taken up by wrestling with BZR, it becomes costly to the point discouraging people from joining.
For the greater good of GNU, move off BZR seems like the only sound choice.

The thread ended without a clear resolution. RMS had a plan. He was working on it. The 200 messages didn’t produce a decision, but they did make the community’s position unmistakable.

2013: The ELPA branch breaks#

Later that year, Stefan made the practical move that set the stage for everything that followed. The ELPA branch was broken on Bazaar, a bug that crashed on checkout, with no one left to fix it. Stefan moved it to Git, and his announcement showed exactly the kind of careful leadership that had kept Emacs development running through all of this:

I’m not terribly happy about this change, since it means we’ll be using two different tools (Git for ’elpa’ and Bzr for ’trunk’), but I really see no other way out.
I don’t want this to be a discussion about the merits/pitfalls of Git vs Bzr, and this is not an occasion to discuss the use of Git for the ’trunk’ either.

He knew exactly what everyone was thinking “if Git is good enough for ELPA, why not for trunk?” and he headed it off. One problem at a time.

2014: ESR pushes the button#

By August 2014, Eric S. Raymond had the conversion scripts ready.

He’d been working on it quietly:

You haven’t heard much about it because the hard work is all done. I have the scripts ready to go and need only about eight hours’ notice before pushing the button.

The actual migration happened in November 2014. On November 13th, ESR posted a seven-word message:

Commits are open. Have at it.

Which some even described as heroic.

Six years of debate, 236 messages in the 2008 thread, 200 messages in the 2013 thread, Stefan’s years of quiet maintenance, countless “help me unstick my bzr” pleas, one dead version control system, and it ended with seven words.

The aftermath#

The days following the migration were educational. Half the core contributors had never used Git:

“This Is The Git Help Mailing List”
“git pull fails with merge conflicts. How can this possibly happen?”
“A simple git workflow for the rest of us”
“need help adjusting workflow to git”
“Good book on Git”
“Obscure error/warning/information message from git pull” (124 messages)

These were people who’d been developing one of the most important text editors in the world for years, asking basic Git questions, because they’d been stuck on Bazaar while the rest of the world moved on.

What a bzr saga! Anyway, better than any film I could have watched tonight.

What's with all the slide decks?

Wed, 13 May 2026 17:41:44 +0000

dynomight · May 2026 · writing

News from the world of real jobs: Apparently, sometime between 10 and 20 years ago, it became standard for people to communicate by sending slide decks around. These slides are never presented. They aren’t intended to be presented. They’re born, they’re sent around, and they die. What?

I stress, the question is not why (or if) people give bad presentations.

The mystery is why everyone is using presentation software for communication that is not a presentation.

Theory 1: Everybody dumb

Is it because we’re all dummies? I’m putting this theory first because I suspect that you, beloved readers, will favor it.

True, if you ask people why they make slides instead of writing, they’ll usually say, “because nobody wants to read”. So there’s that. But I don’t consider this much of an explanation. Dummies though we may be, we’ve been like that a long time. If we entered the Slideocene 15 years ago, why then? Why not before?

Theory 2: The decline of reading

Did we get worse at reading? The Discourse seems to have decided this is true, but is it true, or just moral panic?

Since 1971, the US has tested 13-year-olds to measure long-term trends in reading ability. This shows a slow improvement until 2012, then a slow decline, and finally a post-COVID drop. The declines seem too small and too late to explain our mystery.

Since 2000, PISA has tested reading performance in 15-year-olds around the world. This shows a decline on average, but it’s smaller in rich countries and nonexistent in the United States. (It’s the same story for science and a bit more negative for math.)

Among adults, data is scarce. Basic literacy is generally improving, and American time use data shows a decline in reading for pleasure from around 23 minutes per day in 2003 to around 16 minutes per day in 2023. But this seems to miss time people spend reading on their phones.

So it’s unclear if people got worse at reading. It feels plausible that people now spend less of their adulthood grappling with complex written arguments, and so got worse at that. But there’s little firm evidence.

Theory 3: Technological change

Another obvious theory is that we now have computers and software and the internet. Without these things, it would be impossible to email slides to each other. This seems relevant!

Yes, but we had those things for a while before slide culture really took hold. And think about the situation before computers. Photocopiers were ubiquitous in corporate offices by the mid-1980s, and mimeographs were around decades before that. If slides were really that great, people could have made them by hand. But no one did.

Of course, making slides by hand is inferior. But it’s not that inferior. So slides can’t be that big of a win.

What actually happened?

And… that’s pretty much the end of the obvious theories. None of them are very satisfying. So let’s take a step back. Historically, how did the slide-as-document displace the memo?

As best I can tell, this was driven by management consultancies. If you go back to 1960, they delivered detailed written memos. The memo was the product. They’d likely give a presentation as well, but that was a separate ancillary thing, likely done using flipcharts or chalkboards.

In the 1970s, the memo was still the product, but consultancies started to enforce a top-down logical structure (the Pyramid principle). Presentations shifted to acetate transparencies. Both memos and presentations often included hand-drawn graphics like the nine-box or growth-share matrices.

In the 1980s, the memo was still the product, but presentations became increasingly lengthy and polished. Expensive computers like the Genigraphics started to be used to generate charts.

The 1990s were when things started to shift. By then, PowerPoint was everywhere, and junior analysts were expected to create presentations themselves. Consultancies gradually started to notice that (1) clients didn’t always read the memos; (2) clients loved slides and passed them around long after the presentation was over; and (3) creating a memo and a polished presentation was a lot of work. They put more and more effort into the slides. McKinsey especially evolved towards treating slides as the primary product, and mostly stopped writing long memos. Other consultancies followed.

During the 2000s, slides became even more ornate. Consultancies evolved their formatting rules, and created fancy data-dense charts. They learned that a 200 slide deck made clients feel like they got a lot for their money. Gradually, they oriented their entire business around slides. Projects would start with managers creating a template presentation with “ghost slides” and assigning different parts to junior analysts. Soon, this spread outwards, both from people who interacted with consultants and from the ex-consultant diaspora. People everywhere started thinking and communicating in slides, and now everything is slides, yay!

Alternative history

That story makes slides-as-documents sound inevitable: People liked them, so they became popular. But there’s an alternative timeline in which we resisted the slide into slide maximalism. That timeline is Amazon.com, Inc.

In 2004, Jeff Bezos famously instituted a no-presentations policy at Amazon. His logic was that slides hide poor reasoning and are a tool to persuade rather than inform. Instead, everyone involved with strategic decisions at Amazon needs to learn to write a six-page memo. Meetings begin with everyone sitting and silently reading one of these memos.

Presentation software is not banned at Amazon. The ban is only for using it for internal meetings and decision-making. They use slides for external communication. There is no policy that prohibits someone from making slides and emailing them around.

And yet, people don’t make slides and email them around, because it’s not part of Amazon’s culture. In effect, Amazon is a counter-movement. Most of the world decided that slides are good, because slides are easy. Bezos decided that writing is good because writing is hard.

There are millions of articles explaining why Bezos’ policy is pure genius. They claim that constructing a narrative requires deeper analytical thinking and exposes flaws in logic. I want to believe those theories. I now realize they’re very similar to some of my arguments for why writing with too much formatting is bad.

I’m not sure if writing is the secret to Amazon’s success. But Amazon is successful. This demonstrates that slide life is a choice, not technological destiny—institutions can choose writing over slides and flourish anyway.

OK so then what’s happening?

Warning: If you like your theories simple and mono-causal, you aren’t going to like this.

Slides are a win, but a small one. The shift to slides wasn’t a “mistake”, it happened because people like it. But if sharing slides outside of presentations became illegal, this wouldn’t cause per-capita GDP to crash. That’s why people didn’t scratch slides into mimeograph stencils back in the 1950s. It wasn’t worth the modest effort.
When computers and software showed up, it became easier to share slides. But people didn’t immediately shift to slides-as-documents because the win isn’t that big, because culture changes slowly, and because everyone had pre-existing skills for reading and writing documents.
Consultancies happened to be in the economic niche with the strongest selection pressure to evolve towards slides-as-documents. So when making slides became cheaper, they shifted. Slowly, that norm spread outwards, people got used to communicating in slides, and here we are.
Institutions can resist that norm and still be successful. If you take modern people and force them to read and write, they do just fine.
Humans evolved to learn and communicate in a fragmented, interactive, and visual style. It’s hard to argue that any shift in that direction is a catastrophe.
Except blogs. The decline of the blog must be arrested.

The Cathedral, the Bazaar and the Kitchen · blog.vrypan.net

Tue, 12 May 2026 09:01:47 +0000

Eric Raymond’s “The Cathedral and the Bazaar”¹ described two fundamentally different ways of building software. The Cathedral represented centralized, carefully planned development directed by a small group of maintainers. The Bazaar represented open collaboration: large communities, public iteration, distributed labor, and software evolving through many contributors.

For decades, the Bazaar became the dominant cultural myth of open source. Public repositories, pull requests, and community participation were not just practical tools, but moral ideals. Good software was expected to emerge from openness and collective effort.

AI-assisted software development is changing the economics and the ergonomics behind that model. Implementation becomes cheap, and coordination becomes expensive. A single developer equipped with modern tools can now produce systems that previously required teams.

At the same time, software is becoming increasingly personalized: tailored to one person’s workflow, infrastructure, preferences, and habits. Instead of building generalized tools for the widest possible audience, developers increasingly build software that fits their own environment perfectly. Others may still read the code, fork it, or borrow ideas from it, but local modification often becomes cheaper than upstream coordination.

The new model is a kitchen.

Every kitchen evolves around the habits of its cook. Tools sit where they are convenient. Ingredients are substituted freely. Recipes are modified on instinct. Two people may start from the same dish and end up with completely different results.

Unlike the Bazaar, a kitchen is deeply personal. Recipes are shared freely, but kitchens rarely converge into a universal standard. Visitors may admire another cook’s techniques, yet still return home and prepare the dish their own way. In the Kitchen model, open source becomes less like public infrastructure and more like published craft: software as personal utility, openly visible, endlessly adaptable, and increasingly authored by individuals rather than communities.

In the Bazaar model, openness was mainly a way to coordinate people. You opened the codebase to attract contributors, spread work across many developers, avoid duplicated effort, and slowly build shared infrastructure.

In the Kitchen model, openness serves a different purpose. It provides visibility, learnability, and independence. Value shifts from “others can help build this” to “others can understand, adapt, and reclaim this.”

Source code starts to resemble recipes more than public construction projects. Most people do not submit patches to a cookbook, yet recipes remain enormously valuable because they transfer techniques, preserve knowledge, and provide foundations others can adapt to their own tastes and environments. The code is open not necessarily so everyone can co-author it, but so anyone can study it, modify it, and make it their own.

This also changes the meaning of forks. In the Bazaar, forks were often viewed as failures of governance or coordination. In the Kitchen, forks become normal and healthy:

“I adapted this for my setup”
“I removed features I don’t need”
“I rewrote this around my workflow”

Forking becomes analogous to modifying a recipe² at home.

Software evolves through local adaptation rather than centralized consensus. A developer will remove features, rewrite workflows, or optimize entirely around their own infrastructure because doing so is now cheaper than negotiating a generalized solution acceptable to everyone.

This model still depends on public circulation of ideas. Even if code contributions decline, people still copy ideas from each other constantly through imitation, recombination, critique, and inspiration.

Much of programming history already worked this way:

Unix customization culture where ingredients were expected to be mixed in different ways
shell workflows and personal scripts
dotfiles designed to show others how a system is configured but rarely adopted one-to-one

These are often highly personal systems shared publicly, not collaboratively engineered products.

Open source remains essential because it preserves agency: the ability to inspect, repair, continue, and reshape software independently of its original author. The result is a world where software is increasingly personal, but where ideas, techniques, and tools still circulate freely between individuals, much like recipes passed from kitchen to kitchen.

Eric S. Raymond, The Cathedral and the Bazaar (1997), https://en.wikipedia.org/wiki/The_Cathedral_and_the_Bazaar ↩
Side note: There is rarely a single canonical version of a dish. The same recipe evolves into countless variations shaped by region, available ingredients, habits, and personal taste. Instead of converging into one standard, we distinguish them by origin or authorship: à la provençale, à la milanaise, à la grandma, à la Jacques Pépin. The variation itself becomes part of the identity of the dish. GNU grep and BSD grep were different variations of the same tool, but these were more like publisher or distribution variations. The Kitchen model pushes personalization much further, toward software shaped directly around the habits and preferences of individual developers. ↩

I’ve banned query strings — Chris Morgan

Sun, 10 May 2026 07:55:49 +0000

I’ve banned query strings

🗓️ 2026-05-08 • Tagged /url, /web, /opinions, /meta=only

I don’t like people adding tracking stuff to URLs.
Still less do I like people adding tracking stuff to my URLs.

https://chrismorgan.info/no-query-strings?ref=example.com? Did I ask?
If I wanted to know I’d look at the Referer header; and if it isn’t there, it’s probably for a good reason.
You abuse your users by adding that to the link.

https://chrismorgan.info/no-query-strings?utm_source=example&utm_&c.?
Hey! That one’s even worse, UTM parameters are for me to use, not you.
Leave my URLs alone.

So I’ve decided to try a blanket ban for this site: no unauthorised query strings.

At present I don’t use any query strings.
If I ever start using any query strings, I’ll allow only known parameters.
(In past times I used ?t=… and ?h=… cache-busting URLs for stylesheet URLs;
and I decided I’m okay breaking such requests; there shouldn’t be any legitimate ones.)

Want to see what happens if you add a query string? Go ahead, try it.

It’s my website: I can do what I want with it.

And you can do what you want with yours!

This is currently implemented in my Caddyfile.

What British people really mean when they say 'sorry'

Thu, 07 May 2026 02:56:40 +0000

What British people really mean when they say 'sorry'

Getty Images/ BBC/ Javier Hirschfeld

Sorry. Sorry to bother you. Sorry for the delay. Sorry about the weather. Sorry for all of the above.

In the UK, sorry is not simply an apology, it's a cultural reflex – a five-letter pressure valve used to soften requests, smooth over awkwardness, fill conversational gaps and avoid the national horror of seeming rude. It is perhaps no coincidence that such famously polite characters as Paddington and Mary Poppins are British.

Brits say the word on average nine times per day – more than 3,000 times a year. But for visitors, the puzzle is not how often they hear it, it is working out what sorry actually means. Because in Britain, sorry can mean regret. It can also mean excuse me, move over, I disagree, hurry up, you're blocking the aisle, I didn't hear you or I am trying very hard to not sound annoyed.

While these uses are not unique to the UK, the frequency, tone and the tiny social calculations often are. Britain is often known as a conflict-avoidant society, and sorry has become one of its most versatile tools – a way to manage space, soften disagreement, avoid confrontation and enforce rules without appearing openly impolite.

Essentially, sorry is a politeness code. This one word offers a fascinating glimpse into the many cultural quirks that make the Brits who they are – and for visitors, learning to decode it can be the difference between a friendly exchange and a baffling British misunderstanding.

1. "Sorry!" on the street

What it sounds like: An apology.

What it often means: You're in my way, I'm in your way, we have both briefly existed too physically near one another and must now neutralise the awkwardness immediately.

This is less about fault than the UK's deep discomfort with accidental intimacy: the horror of brushing a stranger's coat, blocking a pavement or occupying the same small patch of public space for a second too long.

Someone may say it when they bump into you, when you bump into them, or when neither of you has done anything wrong beyond brushing shoulders and misjudging pavement geometry. It can mean "excuse me", "after you", "please move" or "let's pretend this tiny collision never happened". The point is not blame, but social repair; a quick word that keeps things moving while sparing all involved the indignity of open confrontation.

Getty Images/ BBC/ Javier Hirschfeld

Brits say "sorry" an average of nine times a day, but the word often does more than apologise (Credit: Getty Images/ BBC/ Javier Hirschfeld)

2. "Sorry?"

What it sounds like: A request to repeat something.

What it often means: I didn't hear you – or I did, but I need a moment to process what you just said.

This beloved apology – with a subtle upward inflection at the end – is one of the English language's most useful conversational tools. It can mean "Pardon?" or "Please, can you repeat that?" or simply "I just need a second". Because "what" can sound startlingly blunt, "sorry?" becomes the softer, less confrontational alternative.

For visitors to the UK, it is handy in places like pubs and train stations with fast-moving conversations – and especially useful in areas with strong regional accents, like those from Liverpool, Newcastle and Glasgow.

Used with a cooler or more incredulous tone, however, it can shift to a distinctly British warning shot: I heard you, but I'm giving you the opportunity to rethink what you said.

3. "Sorry, can I just…"

What it sounds like: A polite request.

What it often means: I need to take up a tiny bit of space and I am apologising in advance for the inconvenience of my existence.

This is the apology of British self-minimisation. You'll hear it on trains, in cafes, at theatre seats, in hotel lobbies and anywhere someone needs to ask a perfectly reasonable question.

"Sorry, can I just squeeze past?" "Sorry, is anyone sitting here?" "Sorry, could I ask…?"

The speaker is not really sorry. They are softening the act of asking, entering, sitting, reaching or existing too visibly in public. In a more direct culture, "Is this seat free?" might do. In Britain, sorry often gets there first, as if occupying an empty chair requires a small act of contrition.

4. "Oh, sorry…"

What it sounds like: An actual apology.

What it often means: I'm objecting, but I'm going to make it sound like an apology.

This may sound like a sincere apology, but it usually isn't. In the UK, where directness can feel horribly awkward, a clipped "Oh, sorry…" is what you might hear when someone needs to reclaim their place without sounding openly confrontational. "Oh, sorry, I think I was next"; "Oh, sorry, that's my seat"; "Oh sorry, I was using that."

The apology gives the speaker cover; the pause after "oh" does the damage. It lets them object while remaining technically polite – a very British compromise between saying nothing and saying exactly what they think.

Getty Images/ BBC/ Javier Hirschfeld

In the UK, "sorry" can also mean "excuse me", "hurry up", "you're blocking the aisle" or "I am trying very hard to not sound annoyed" (Credit: Getty Images/BBC/ Javier Hirschfeld)

5. "Sorry, but…"

What it sounds like: A polite, throat-clearing interruption before a blunt contradiction.

What it often means: Try as I might to agree with you, I can't. I'm about to explain why you're wrong and I don't care what you think.

This is the pre-emptive apology: a small cushion placed before a disagreement lands. In a culture where open disagreement can feel socially abrasive, "sorry, but…" lets the speaker object while maintaining the appearance of civility.

More like this:

• How not to embarrass yourself in a British pub

• You can sleep over at Oxford's colleges - here's what it's really like

• The tiny slice of 'America' in England

It allows the speaker to challenge, contradict or correct while signalling they're not trying to start a fight – even when they are absolutely about to. Depending on the tone, it can sound conciliatory, mildly exasperated or one step short of saying "Sorry, but I'm not sorry." For visitors, the trick is to listen to what comes after the "but". In Britain, that's usually where the real message begins.

6. "Sorry…" in a queue or pub

What it sounds like: An etiquette reminder

What it often means: I'm trying not to make this awkward, but this isn't fair; you've broken the rules.

The blood runs cold at the thought of queue-jumping in Britain. Here, the queue is sacred territory – like Westminster Abbey or Wimbledon – and a politely interjected "sorry…" acts as an etiquette reminder that everyone must adhere to the rules instead of hustling for territory. In this scenario, sorry is code for "get to the back", "don't push in", "keep your distance" or "queue jump if you dare".

In the pub, the same phrase can mean "just behind you", "I think I was next" or "please don't pretend you didn't see me waiting". It's a correction dressed up as courtesy – which is often the most British correction of all.

If you liked this story, sign up for The Essential List newsletter – a handpicked selection of features, videos and can't-miss news, delivered to your inbox twice a week.

For more Travel stories from the BBC, follow us on Facebook and Instagram.

100,000 Glyphosate Lawsuits: Why Roundup Does Not Kill Your Weeds Like It Used To – Economist Writing Every Day

Tue, 05 May 2026 18:29:56 +0000

I don’t like wasting time bending over and pulling out weeds, one by one. Much more efficient to go squirt squirt and eliminate lots of weeds at a time. But I realized in the past year that the Roundup I spritzed on the weeds in my mulch beds and sidewalk cracks just wasn’t killing them like it used to. The weeds would shrivel a bit, but then many would bounce right back. So, when I went to Home Depot to buy some more this week, I looked at the ingredients on the label. What?? Where is the glyphosate? For decades, “Roundup” was synonymous with glyphosate.

Glyphosate has several desirable properties as an herbicide. You spray it on the leaves, and it kills the plants right down to the roots. However, it has minimal residual toxicity in the soil, so it is unlikely to kill any plants you did not spray, and you can replant quickly in a soil patch that you had cleared with glyphosate. Farmers love it, because you can buy genetically engineered strains of crops like corn that are immune to glyphosate, so you can spray your fields to kill weeds without harming standing crops.

The glyphosate story is much bigger than homeowners bending over to pull weeds. The chemical has become indispensable for current agriculture. Global glyphosate sales are about $10 billion per year, and its impact on crop productivity is enormous. A 2017 study (apparently not paid for by Monsanto) predicted dire effects of discontinuance:

World prices of all grains, oilseeds and sugar are expected to rise, especially soybeans (+5.4%) and rapeseed (+2%). The welfare impacts are mostly negative, with global welfare falling by $7,408 million per year. Land use changes will arise, with an additional cropping area of 762,000 ha, of which 53% derives from new land brought into cropping agriculture, including 167,000 of deforestation. These land use changes are likely to induce the generation of an additional 234,000 million kg of carbon dioxide emissions.

What’s not to like about glyphosate? Well, maybe it causes cancers in humans. This is a contested claim, and I don’t have the expertise to penetrate the arguments. Because glyphosate makers like Monsanto and its successors Bayer have deep pockets, lawyers on contingency have swarmed like killer bees to file lawsuits, over 100,000 of them, of which about 60,000 remain active globally.

National security issues have muddied the waters here. For instance, Robert F. Kennedy, Jr. led a landmark legal case against Monsanto in 2018, securing a $289 million jury verdict (later reduced on appeal to $20.4 million) for a school groundskeeper who developed non-Hodgkin’s lymphoma after prolonged exposure to Roundup. That case energized a bazillion further lawsuits. But now Kennedy is going along with the current administration’s position that it is strategically necessary to maintain production and responsible access to glyphosate: farmers demand it, and Bayer operates the only plant in the U.S. producing significant amounts of elemental phosphorus, which is a vital material for defense and, increasingly, for lithium batteries.

Naturally, Bayer denies that glyphosate is particularly harmful. The firm continues to sell the product to farmers and landscape professionals, but it has removed it from retail bottles of Roundup you see on Home Depot shelves, in an effort to reduce exposure for further litigation.

What have they substituted for good old glyphosate? I found a brew of three other chemicals. I can report reliably that this mixture is much less effective, especially on grasses and on well-established weeds. The Internet backs up my observations. The Iowa State garden extension has a great table of the real-world effects of all common herbicides.

So, what to do? For grasses in my backyard gravel patch, I am spraying multiple times. If that doesn’t work, I may try covering that area with a black tarp for a month to kill the grass. I have considered buying a propane flamethrower weeder, but that seems only effective on the same things the current wimpy Roundup kills (small/young broadleaf weeds).

For mulched areas, I am incentivized to keep up with fresh mulch to keep weeds from growing in the first place. For larger weeds, I have now found myself bending down low, grasping them close to the ground, and actually pulling them out by hand.

Why airlines are always going bankrupt - David Oks

Mon, 04 May 2026 17:56:12 +0000

All photos by Mike Kelley from his “Life Cycles” series

It might not be the most important story in the world right now, as our species takes its first halting steps into a brave new world of technological power whose contours are still to us mysterious and weighted with fearful portent, but lately I’ve been spending a good bit of time reading about the death of Spirit Airlines. Spirit, for those lucky enough to have never flown on one of its planes—I have a few memories of terrible Spirit flights from New York to Miami in my teenage years—is, or rather was, one of the ten or so largest airlines in the United States, and, after its more popular rival Southwest, the most prominent of the budget airlines. (JetBlue is somewhat larger, but can’t be considered a “true” budget airline.) And, for the last few years, Spirit had been hurtling toward insolvency.

Spirit had last turned a profit in 2019; things turned disastrously bad with the COVID pandemic in 2020—as was the case for every other airline—but whereas larger flyers generally recovered, things went from bad to worse for Spirit. Corporate leadership pursued a merger with JetBlue, but this was blocked by a federal judge. And so in November 2024, Spirit filed for Chapter 11 bankruptcy protection; then it filed again, less than a year later, in August 2025. But these filings did little to save Spirit. There was talk of liquidating the company. The Trump administration raised the prospect of a capital injection that would leave the federal government with a 90 percent stake in the airline (the first time in American history that the federal government has owned a passenger airline outright), but the talks collapsed, and so in early May 2026 Spirit announced that it was shutting down for good.

The collapse of Spirit was unique in that in its death throes it managed to solicit a bailout offer from the U.S. government; but it was not unique among its fellow airlines in going broke. Airlines are a bad business: a really, really bad business. The International Air Transport Association, the trade body of the global airline industry, has documented for years that airlines as a sector destroy investor value in the aggregate. The IATA’s 2026 outlook, looking forward to a quite strong year—this was before the Iran war broke out and oil prices surged—projected an average return on invested capital of 6.8 percent, against a weighted average cost of capital of 8.2 percent. As the IATA’s report said, “the airline industry collectively does not generate earnings that cover its cost of capital.” This has been the case for a long time. From its deregulation in 1978 to the end of 2025, the airline industry has cumulatively lost money: its net profit over those 47 years sits at negative $37 billion.1

Given these grim economics, you won’t be surprised to hear that airlines have a bad habit of going insolvent. This includes many of the most famous names in the history of aviation. Pan Am, long the unofficial flag carrier of the United States, ceased operations in 1991; Eastern Air Lines liquidated the same year; TWA, the carrier of Howard Hughes, was absorbed into American Airlines after a third bankruptcy filing in 2001; Braniff died in 1982. And those are only the most famous names; countless aviation startups have come and gone. (Have you ever heard of Trump Shuttle?) Even airlines with the backing of a national government go bankrupt all the time: Alitalia, Italy’s flag carrier, reported only a single year of profit since its founding in 1946 and was saved countless times by the Italian government before ultimately ceasing operations in 2021. Even those airlines that survive for long periods of time are perpetually in financial distress. Between 1978 and 2005, more than 160 airlines filed for bankruptcy; virtually every major U.S. carrier other than Southwest has been to bankruptcy court at least once. In September 2005, every one of the four largest American airlines—United, Delta, Northwest, and US Airways—was operating simultaneously under Chapter 11 protection.

This is very strange. There’s not really a conventional economic explanation for an industry whose long-term equilibrium is losing money: an industry that, on a purely economic level, should not exist. Warren Buffett once called the airline industry a “bottomless pit” for investor capital. “Indeed,” he wrote, “if a farsighted capitalist had been present at Kitty Hawk, he would have done his successors a huge favor by shooting Orville down.”

So why is the airline business so remarkably bad?

One answer is that airlines are particularly vulnerable to shocks. There are so many potential risks with air travel that practically anything going wrong will have some effect. The September 11th attacks, for example, had a huge effect on air travel; so did the surging oil prices of the 2000s, the financial crisis of 2008 and the resulting recession, the 2020 pandemic, and now the volatility in oil prices surrounding the Iran war. Whenever a major shock occurs you tend to see a huge wave of airline bankruptcies.

But airlines obviously aren’t the only type of business in the world that’s vulnerable to shocks. Hotels, for instance, are heavily exposed to recessions, terrorism, and pandemics; their costs are heavily front-loaded into the property, just as an airline’s costs are loaded into the plane; and yet the hotel industry doesn’t go through synchronized waves of bankruptcy each time a shock hits. Shocks might explain why airlines tip over the edge into restructuring or liquidation; but they don’t really explain why they’re so vulnerable in the first place, or why the airline sector—uniquely among all major industries—is unable to generate profit in the aggregate.

And we don’t see the same structural unprofitability in any of the other companies of the aviation ecosystem: engine and avionics manufacturers, for example, do totally fine; so do the service suppliers that sell into airlines.

Maybe, then, the answer is that airlines specifically are just poorly managed. This was the dominant view in the 2000s and 2010s: legacy full-service carriers were chronic money-losers; budget airlines, like Southwest and Ryanair, were much more profitable; and so in the future air travel would bifurcate into budget aviation for the masses and Emirates-style luxury travel for the few. But the budget airlines don’t look so good anymore. Spirit was a flagship budget airline and has now been liquidated; JetBlue and Frontier, two budget or semi-budget competitors, are also at risk of bankruptcy; even Southwest, the most durable and iconic of the low-cost carriers, has been unable to make a profit since the pandemic and is now fending off an activist challenge from the hedge fund Elliott Management. So the budget strategy clearly wasn’t a solution to the airline industry’s problems.

So explanations that cite shocks or bad management either explain too much or too little. If it’s just vulnerability to shocks, why don’t other industries have such huge bankruptcy waves? And if it’s bad management, why has no airline in the long history of aviation figured out a replicable solution to running the business profitably?

I’d like to suggest that the problem with the airline industry is much deeper than people seem to think. Losing money in the aggregate is a feature, not a bug, of a competitive airline industry. The airline sector, for reasons that go into the essential nature of the industry, cannot reach a profitable competitive equilibrium. This is not because airlines are vulnerable to shocks or because they’re poorly managed. The airline industry itself can either be profitable, or it can be competitive: but it can’t really be both.

To understand why, we have to learn a little bit about game theory.

The airline industry can be competitive or profitable, not both

Game theory is the formal study of strategic interaction: that is, the study of situations where each agent’s best move depends on what they expect others to do. Game theory originated in the 1940s, with the work of John von Neumann; and it’s most frequently associated with John Nash, the mathematician who gave us the idea of the “Nash equilibrium.” (And was played by Russell Crowe in A Beautiful Mind.) Game theory is a huge and influential field, fruitful enough to branch into many subfields.

If you’re familiar with game theory, you’re probably most familiar with the study of “non-cooperative games,” like the famous prisoner’s dilemma: this is the subfield that studies how rational agents will behave when they can’t make binding commitments to one another, and are thus in a state of permanent competition. But the branch of game theory that tells us the most about airline economics comes instead from cooperative game theory, which studies what happens when agents can form binding agreements. The central question of cooperative game theory is which arrangements among players are stable: that is, which arrangements have the property that no subset of players could break away and do better on their own.

One of the central ideas in the study of cooperative games is the idea of the core. The “core” of a game is simply the set of outcomes that no coalition of players can improve upon by breaking away and dealing among themselves. If an outcome is “in the core,” it’s stable, such that nobody can propose a side deal that makes every member of some subgroup better off; if the core is “empty,” then every arrangement is vulnerable to being undercut by some side-coalition, and the market has no resting point, no stable equilibrium. It cycles, destabilizes, and, without outside intervention of some kind, eventually breaks down.

Airlines are the classic example of an “empty core” industry: an industry that is structurally incapable of reaching competitive equilibrium. But why is it that airlines have an empty core, while other industries—ones that also have plenty of competition, but converge on healthy margins and stable prices—don’t?

Luckily we have an answer from the University of Chicago economist Lester Telser, one of the pioneers in applying game theory to economic questions. Telser’s central idea, developed across a body of work in the 1970s and ‘80s, was that the empty-core syndrome was a structural feature of particular industries. Whether an industry suffered from an empty core or did not depended on a few simple conditions.

What were those conditions? Telser identified a particular combination. Industries with an empty core, he suggested, are marked on the demand side by a lack of product differentiation and volatile consumer demand. And on the production side, they combine high fixed costs with low marginal costs and sharp economies of scale: the minimum efficient scale of a single firm is thus large relative to the total size of the market, and the efficient number of firms in the market is relatively small. By minimum efficient scale, we mean the smallest level of output at which a firm reaches its lowest average cost: below it, fixed costs are spread too thin and per-unit costs are high; and above it, the cost curve eventually flattens or even rises, as coordination costs and managerial complexity erode the gains from further growth. Total demand divided by minimum efficient scale, then, gives us the efficient number of firms a market can support.

To see why this particular combination of features is so toxic, consider a stylized example. Suppose you have an industry where the minimum efficient scale of a single firm is large relative to the size of the market: large enough that the market can support, say, two and a half efficient firms. That is to say: two firms can’t quite produce enough to satisfy demand; but three firms is one too many for all of them to operate at full capacity. Obviously you can’t have half a firm: firms come in whole numbers. So what happens?

Suppose you try to run the industry with just two firms. Demand exceeds supply, such that prices are high and there’s plenty of profit to enjoy. But that profit is exactly what invites a third firm to enter, undercut both incumbents, and still cover its costs. Now there are three firms, and supply exceeds demand. Someone has to operate below scale and bleed money on fixed expenses; eventually one of the firms will have to leave the market. Now you’re back to where you started: prices recover, profits climb higher, and the cycle begins again.

So whichever side of the integer you land on—one firm too many, one firm too few—there is some coalition of firms and customers that can profitably reorganize the market against the existing arrangement. In the language of cooperative game theory, the allocation is always vulnerable to defection by some coalition. The core is empty.

This wouldn’t be a problem if the efficient scale of a firm were small relative to total demand. If the market could support 1,778.4 firms’ worth of output, and there happened to be 1,779 firms, no one would notice the rounding error. But when the efficient number of firms is small, adding or subtracting one firm causes huge perturbations. “Lumpy” supply, in a small-numbers market, is the heart of the problem. And this is made worse by volatile demand. An industry that was barely sustainable with three firms becomes catastrophically oversupplied the moment that demand softens.

Most industries don’t really fall into this bucket. Take, for example, soap: manufacturing soap has real economies of scale, but the minimum efficient scale is small relative to total demand, products are differentiable enough through brand and recipe, and demand is reasonably steady. So the firm can settle quite comfortably into an equilibrium that is stable, competitive, and durably profitable. The empty-core syndrome only kicks in where minimum efficient scale is large relative to total demand, where products are undifferentiated, where economies of scale are sharp, and where demand is prone to swing.

You find the empty-core syndrome, for example, in the railroad industry of the nineteenth century. Building a railroad required vast capital expenditure on track, rolling stock, depots, and bridges; but once the infrastructure was in place, the marginal cost of carrying an additional ton of freight or another passenger across it was almost zero. Two railroads running competing lines between, say, Chicago and New York could not both operate at full cost recovery; so they spent the 1870s and 1880s alternately forming pools and rate-fixing agreements, then watching them collapse into ruinous price wars, going bankrupt, reorganizing, and starting the cycle over again.

And you’ll find the same dynamic in the contemporary airline industry.

Suppose you’re managing an airline that does flights from San Francisco to Tokyo. (A flight I’m considering taking, by the way.) Most of your costs, you’ll find, are fixed. The aircraft itself—let’s say it’s a modern widebody, like a Boeing 787—will cost you somewhere in the low hundreds of millions; that’s a fixed cost. So are the gate slot for your departure from San Francisco and the landing rights for your arrival in Tokyo. Labor costs might seem variable, but they’re actually not: pilot, flight attendant, and mechanic compensation in the United States is governed by the Railway Labor Act of 1926 (which was extended to airlines in 1936), which stipulates that collective bargaining agreements don’t actually expire but rather remain in force until they’re replaced. So even your wage bill is more or less fixed over multi-year horizons. The most variable major cost you’ll have to deal with is jet fuel; but given that spiking fuel prices aren’t your friend, you’d rather hedge fuel costs aggressively to smooth out cash flow. So fuel also acts more like a fixed cost.

All of which is to say: you have a lot of fixed costs.

Now let’s suppose daily demand on your route from San Francisco to Tokyo is roughly 800 passengers willing to pay a price that covers the full cost of flying. A widebody seats around 250 to 300 people. Perhaps you offer one flight per day, and you have two competitors that also offer one flight a day. This puts about 750 to 900 seats into the market: close enough to demand that fares stay healthy and the route covers its costs, plus a bit extra that you and your competitors can take as profit. Some customers don’t fly, and fares get bid up; but this is fine enough, at least for you.

But there’s a problem with this situation: there’s enough slack in the market—enough customers paying more than they could, and enough margin that you’re taking—that a new entrant could see an opportunity to enter. And once that new competitor has entered, you now have 1,000 to 1,200 seats chasing 800 passengers, which means that somebody has to lose out. The efficient number of daily flights is somewhere between three and four; but capacity is lumpy. So whereas three was somewhat too few, four is definitely too many. So you enter the instability part of the cycle. Fares collapse; margins take a hit; eventually someone has to exit. Once the weakest competitor has exited, the market consolidates again, and fares recover; but then someone sees the unmet demand and enters again.

So even in the best of times, there’s a deep structural instability to the airline industry: margins are structurally depressed and companies are unable to recoup their cost of capital. And that is in the best of times. Whenever there’s a major shock that hits costs or demand, airlines enter periods of severe crisis. Because the variable cost of flying a half-empty plane is barely lower than the variable cost of flying a full one—and not that much higher than keeping the planes grounded—airlines don’t pull capacity proportional to the decline in demand. Capacity persists and fares collapse; so margins go from slightly positive to sharply negative. And this ruinous competition is the basic rhythm of the airline sector.

That is why airlines go bankrupt so frequently. Under American law, Chapter 11 bankruptcy protection—which allows a company to continue operating while it restructures its debts under court supervision—is practically the only mechanism by which an airline can renegotiate its rigid cost structure, from aircraft leases to collective bargaining agreements. Oftentimes this renegotiation takes on a rather predatory character. When United Airlines filed for bankruptcy in 2002 in the aftermath of the September 11th attacks, it terminated its pension plan and left the costs to be absorbed by the U.S. government’s Pension Benefit Guaranty Corporation—saving United about $6.6 billion.

So Chapter 11 is a relief valve for airlines struggling under the weight of their fixed costs; but it doesn’t really do much to help the system as a whole. For airlines, bankruptcy rarely culminates with liquidation; airlines that emerge from bankruptcy proceedings, having voided pension obligations and rejected aircraft leases, can operate at a fundamentally lower cost basis than their competitors. So bankruptcy doesn’t really restore the industry to a competitive equilibrium that can cover the cost of capital: it resets the floor at a lower level, from which a new round of ruinous competition can begin.

To be profitable, the airline industry has to be uncompetitive

The economics of a genuinely competitive airline industry, then, are really bad—for the same reason the economics of any empty-core industry are bad. And this suggests that, in search of stability, the market participants will eventually try to suppress competition, if only so they can survive.

This is what’s happened with most empty-core industries, like railroads and ocean shipping. Anticompetitive measures, like cartels and mergers and vertical integration, are necessary in these industries: not because the firms involved are particularly “greedy,” but because uncompetitive equilibria are the only stable equilibria. In antitrust law, these arrangements would be considered “restraints of trade.” But they’re doing real economic work: they’re choosing, even if arbitrarily, among allocations that would otherwise be inherently unstable and frequently unprofitable.

And this is also the story of the airline industry. From the very beginning of commercial aviation in the United States, it was clear that the industry couldn’t attain competitive equilibrium like other industries did. Airlines “weren’t like the other kids.” The 1930s, when commercial aviation first got onto its sickly legs, already saw a wave of airline failures; in response, the U.S. government created the Civil Aeronautics Board, a government commission to determine the shape of the airline industry. Essentially this amounted to a government-approved cartel. The CAB told airlines which routes they could fly, when they could fly them, and what they could charge; entry into new interstate markets was effectively prohibited, such that not a single trunk-line carrier was admitted to the industry between 1938 and 1978; and fares were set to provide carriers with reasonable rates of return. Under this regime, the airline industry was profitable, comfortable, and slightly boring; they competed on service and on the glamour of their cabins, not on price.

But by the mid-1970s the consensus among economists had turned firmly against the CAB regime. Studies of unregulated intrastate carriers in California and Texas showed that they charged half the fares of the regulated lines on comparable routes, and made money doing it; and the broader environment of dramatic inflation made the aviation cartel look less like a stabilizing arrangement and more like a tax on the public. In 1978, the Airline Deregulation Act stripped the CAB of its rate-setting and route-licensing authority, and—for the first time in American history—turned a regulated cartel into a competitive free market.

The period after 1978, unsurprisingly, was a much more tumultuous one than what came before. The airline industry entered exactly the pattern of chronic instability that Telser described in empty-core industries. In route after route there were waves of entry, price wars, bankruptcies, consolidation, brief stability, and then another wave of entry and price wars. The first great wave of bankruptcies came in the 1980s, driven partly by the inability of the legacy carriers to adjust quickly enough to a world in which fares were set by markets and partly by the failure of most of the post-deregulation entrants. The 1990s were relatively stable; and then the 2000s saw another wave of insolvencies, with United, Delta, Northwest, and US Airways all declaring bankruptcy in the course of a few years. The American airline industry enjoyed a period of rare stability in the 2010s—finally seeming, for a brief and glorious moment, like a good business—before proceeding to crash again in the 2020s.

In the course of this long path of suffering, every airline has decided, in one way or another, that the competitive airline industry is structurally unprofitable, and not really worth participating in.

One response is to cartelize the industry through means other than direct rate-fixing: to recreate, by private contract, the kind of competition-suppressing arrangements that the CAB previously wrote into statute. The international alliances of which airlines are so fond—Star, SkyTeam, and Oneworld, with their codesharing and antitrust-immunized joint venture agreements—are one form of this: they allow nominally competitive airlines to coordinate scheduling, share revenues, and refrain from undercutting each other on high-value trans-oceanic routes.

The hub-and-spoke model that dominates domestic aviation is another form of this tacit cartelization. By concentrating its operations at a few major airports, an airline can turn those airports into something close to local monopolies. American Airlines, for example, carries about 90 percent of passengers at Charlotte Douglas and 82 percent of passengers at Dallas-Fort Worth, but only about 7 percent of passengers at San Francisco, where the market is dominated by United, and 2 percent at Atlanta International, which is the central hub for Delta. In effect, major domestic airlines have carved up the country into a sort of feudal map of fortress hubs, with each one operating a quasi-monopoly through which it produces the margins that cannot be earned in genuine competition.

But the other response, and perhaps the more interesting one, is to leave the airline business entirely: to treat the planes as a kind of loss-leading distribution channel for what has become the actual product. The main innovation of the airline industry of the last few decades, from this vantage point, has been the frequent flyer program. Invented in the immediate aftermath of deregulation as airlines scrambled for ways to lock in customer loyalty, frequent flyer programs have become something quite different: enormous, free-floating financial businesses, miles-as-currency operations whose value bears essentially no relationship to the cost of the seats backing them.

The greatest beneficiary of this turn has been Delta, the most profitable airline in the United States, which started a fruitful partnership with American Express in 1996 and launched a co-branded card with them in 2008. Annual spending on Delta-branded American Express cards comes out to about 1 percent of U.S. GDP. In 2025, this produced about $8 billion in revenue for Delta, accounting for more than the entirety of its profit. That means that without the American Express partnership, Delta would be operating at a substantial loss. In effect, Delta’s aviation business is a loss leader for a much more profitable credit card partnership. So to the extent that Delta is now a good business, it is because it escaped the basic instability of the airline industry by becoming less of an airline.

(We see a sort of mixed strategy from Ryanair, the most consistently profitable of all airlines. Ryanair has managed to attain by far the lowest fixed costs of any major airline in the world, due largely to its canny patronage of low-volume regional airports which give healthy discounts on gate slots and landing rights; it flies out of small secondary airports, like Stansted and Charleroi, and thus effectively monopolizes hundreds of routes within Europe; it extracts substantial subsidies from regional governments eager to attract air service; and it earns a large share of its profit from ancillary fees, treating the seat itself as something close to a loss leader. Ryanair, then, is profitable roughly to the degree that it chooses not to compete.)

Airlines are a permanently bad business

It’s not at all clear that all this instability has been a bad thing for consumers. Real airfares in the United States have fallen by roughly half since deregulation; and this decline in prices has made air travel a form of mass transit, rather than a privilege of the affluent. To the extent that people have suffered for the empty-core syndrome that afflicts the aviation industry, the brunt has fallen on equity holders, who have been wiped out repeatedly, and on the airlines’ workers, whose contracts are occasionally rewritten under the duress of bankruptcy at the trough of every business cycle.

So it’s not quite clear that airline deregulation was a bad thing: indeed, given the relative dearth of technical innovation in commercial aviation over the last few decades—largely due to the sad failure of supersonic airliners—we probably have deregulation to thank for the declining cost of flying.

But it’s still worth thinking about what the airline industry’s tendency toward bankruptcy tells us, not only about aviation but also about economics. Not all industries are able to attain a profitable competitive equilibrium. We need aviation to exist as an industry, but we’re unable to have it survive as a profitable concern; the natural tendency, then, is toward anticompetitive consolidation of some kind, or—as pioneered by Delta—toward treating airline seats as a loss-leader for something more lucrative.

In the last century we’ve groped toward stability through government-approved cartels; then dismantled them; then approved their tacit reinstatement through private contracts; and now the empty-core syndrome has at last raised the question of direct government ownership of a major airline. We’ve admitted to ourselves, by now, what we’re still not able to say aloud: that there’s no such thing as a competitive equilibrium for the airline industry. The only question that remains is who is going to be left holding the bag.

This Substack is supported by readers like you.

A calculation of cumulative net profit for the aviation industry was last published, as far as I could find it, by the economist Severin Borenstein in 2011; I had my trusty agentic LLM analyze his methods, replicate them (looking at “BTS TranStats Schedule P-1.2”), and carry them forward until Q3 of 2025. It reported that “cumulative domestic airline profit since deregulation is still about -$24.5B in 2009Q4 dollars, or about -$37.4B in March 2026 dollars.”

Security Through Obscurity Is NOT Bad - Mo Beigi

Sun, 03 May 2026 18:29:41 +0000

Escaping the crowded echo chamber

I was recently reading a post by a user on a web development forum. This user, whom we’ll call Mini, was asking the community whether it was worth using JavaScript obfuscation for some of the scripts running on their website. Their main goal was to make it harder for data-scraping bots to reverse engineer and replicate the API requests powering the page.

Then I saw it: like a solo LGTM comment on a +4,156/-1,640 line PR, a comment from another user whom we'll call Echo:

Security through obscurity is bad

What was worse was that this comment had many upvotes, likely from others who had heard the phrase once and simply channelled their inner parrot to repeat it forever.

I decided to reply to Echo's comment and share my thoughts:

Security through obscurity is NOT bad.
Security ONLY through obscurity is bad (Kerckhoffs's Principle).
Security through obscurity, as an additional layer, is good!

At first, I thought this was what Echo actually meant, but to my surprise, Echo believed that all forms of obscurity were redundant and should not be used at all. They also specifically argued that, in the modern day, AI had made getting around any sort of obscurity trivial.

In this post, I will explain why Echo is wrong and why security through obscurity has its place.

Don't show your working out

Security through obscurity is the practice of reducing exposure by keeping an application's inner workings or implementation details less visible to attackers. Unlike in mathematics, you do not want to show your working out.

It's the digital equivalent of hiding a spare key under the doormat instead of leaving it in the lock. In this scenario, a malicious actor might not bother looking under the doormat and might just leave. Congratulations, obscurity just saved you a break-in. They might still find the key, but they may check a nearby potted plant or mailbox first. That costs time, and time is money. To a malicious actor, the longer they spend chasing dead ends, the more likely they are to give up and move on.

Now, of course, proper security here would be not hiding a spare key near the door at all, but instead leaving it with a trusted family member or friend. Relying only on obscurity for security is bad. You should always secure your applications to the degree warranted, then sprinkle some obscurity on top to make the endeavour of attacking you more expensive. This is simply one part of a defence-in-depth strategy.

Obscurity in the real world

Of course, some examples might help drive the point home. Here are some specific examples I have personally encountered.

WordPress database table prefix

There is a long-standing security recommendation to change WordPress's default database table prefix to a random one. For example, wp_users becomes wp_8df7b8_users. This is often dismissed as "worthless" because it is security through obscurity.

I used to run this very blog on WordPress. Back in 2015, one of the plugins I used had an SQL injection vulnerability that allowed malicious actors to dump the databases of websites using it. These actors had bots scouring the web for vulnerable WordPress targets.

My website was vulnerable. However, I was not impacted by any attacks, and I updated the plugin to a patched version a few days later. While other sites were "nulled" and destroyed, I was spared. I later found a PoC script on GitHub showcasing the exploitation. Using that PoC on my own site failed with a generic error like Table 'wordpress.wp_users' doesn't exist. Therefore, while I was likely still vulnerable and could have been exploited with different SQL queries, the standard query targeting most users did not impact me.

I was spared thanks to that additional layer of security through obscurity. AI tooling today could keep trying different queries, and it may produce good results for malicious actors, but tokens still cost money. The more time and money the bot spends, the more likely it is to give up and move on. It's a battle of sustained resistance.

CSGO's debug symbol leak

I ran an Australian and New Zealand-based CSGO community server called Invex Gaming for several years. As a server operator, I tried to distinguish the servers I ran by adding unique custom mods. To do this, I used an amazing platform called SourceMod, which allowed you to write custom plugins and extensions. To write useful mods, you would often have to find and call functions directly in the game's binaries. CSGO ships with binaries such as engine.dll, client.dll, and server.dll. These binaries contained much of the game logic we wanted to invoke.

For example, I might want one of my mods to programmatically set a player's health. To do this, my script, running on the CSGO server, had to call the right function in the game. In this case, CBaseEntity::SetHealth is one such function, and calling it programmatically would allow me to set the health of any entity.

This is what the function signature looks like:

Now, this is a common and well-known function that SourceMod maps for us correctly. But there are many functions in the game that are less common or not well known. How do we find and use these functions in our scripts? We have to find the function in the binary by reverse engineering it with tools like IDA Pro or Ghidra. Once we have identified it, we can build a stable reference to it using signature scanning or an offset in a virtual function table.

Unfortunately, we cannot access Valve's source code for the game. When looking at reverse-engineered game code, we see a mess of compiled code that takes significant effort to reverse and document properly. Function names, variable names, and data types or structures are not included. This is because it is common practice to strip away debug symbols from game binaries. Debug symbols are metadata generated during compilation that map a program's machine code back to its original human-readable source code. They are extremely useful for reverse engineering and understanding the code.

One day, Valve accidentally pushed an update for the macOS version of CS:GO that included the full, unstripped Mach-O debug symbols in the .dylib binaries. This exposed much more of the game's internals at the time. This led to a rush of server operators using the new information to create new and exciting scripts. Unfortunately, cheat developers also used it to further develop their cheats. A classic double-edged sword.

This is a prime example of Valve valuing the additional layer of security provided by obscurity. Valve has to ship the game in binary form, because the game runs on our machines when we play it. Valve chooses to strip debug symbols from its binaries because doing so is highly effective at reducing the efficiency of cheat developers. Shortly after this release, Valve realised its mistake and re-released the same version with the debug symbols stripped.

Obfuscated code

I do my fair share of malware analysis and CTFs every now and again for fun. It is extremely common to run into obfuscated code, which is source code that has been intentionally complicated to make it harder for humans and tools to understand while remaining fully functional. The malware industry is a billion-dollar industry, and nobody relies on obscurity more than malicious actors. The more obfuscated the malicious payload, the less likely security researchers and tools are to understand what is going on.

On the flip side, enterprises like Google also use JavaScript obfuscation to hide sensitive logic in the browser. A great example is Google reCAPTCHA, where the obfuscation is often heavy and sophisticated in order to make it harder for bots to understand the checks being performed and automate solving them. Netflix also uses obfuscation in its browser-side DRM components to help protect the logic that lets your browser play the video without exposing everything needed to easily extract and save a playable copy. Riot Games also uses obfuscation around parts of the communication between its kernel-level anti-cheat system, Riot Vanguard, and its servers to make it harder for cheaters to fake a clean signal while cheats are running.

Now, there was a suggestion that advances in AI have made obfuscation obsolete. I disagree. While AI tools are good at deobfuscating code, it is still often a slow and expensive process. I do believe a strong model will eventually reach a solution, but it will take time and money. Again, the longer and more expensive it is, the more likely people are to give up and move on.

I do not have concrete data to share on this topic, but I do have some anecdotal evidence. I attempted a hard PWN-style CTF challenge last year that I was not able to solve on my own. Using an LLM, Claude Opus 4.5, and giving it all the information, binaries, and local tools needed to solve the challenge, it still failed at first.

It was not until 4.5 hours of non-stop token burning and many trial-and-error iterations later that the LLM was able to find a solution. This endeavour used 61 million input tokens and 11 million output tokens, or roughly $300 USD. While I was willing to spend that much to gauge the model's ability to solve the challenge, it is important to keep in mind that we already knew a solution existed because this was a CTF challenge with an intended solution. Would malicious actors be willing to spend that much per attempt across a large enterprise attack surface, where results are far from guaranteed? How long are they willing to iterate on one specific angle? That uncertainty is exactly where obscurity still has value.

Spread the word

It should be clear by now that security through obscurity still has its place as an additional security layer in the modern world, even with AI-assisted tooling.

So I no longer want to hear the phrase "security through obscurity is bad".

From now on, let's spread these two statements instead:

Security ONLY through obscurity is bad

Security through obscurity, as an additional layer, is good!

The text mode lie: why modern TUIs are a nightmare for accessibility — The Inclusive Lens

Sun, 03 May 2026 18:29:34 +0000

The mythical, it's text, so it's accessible

There is a persistent misconception among sighted developers: if an application runs in a terminal, it is inherently accessible. The logic assumes that because there are no graphics, no complex DOM, and no WebGL canvases, the content is just raw ASCII text that a screen reader can easily parse.

The reality is different. Most modern Text User Interfaces (TUIs) are often more hostile to accessibility than poorly coded graphical interfaces. The very tools designed to improve the Developer Experience (DX) in the terminal—frameworks like Ink (JS/React), Bubble Tea (Go), or tcell—are actively destroying the experience for blind users.

The Architectural Flaw: Stream vs. Grid

To understand the failure, we must distinguish between two distinct concepts often conflated under “terminal apps”: the CLI (Command Line Interface) and the TUI.

The CLI (The Stream): This operates on a standard input/output model (stdin/stdout). You type a command, the system appends the result below, and the cursor moves down. This is linear and chronological. For a screen reader, specifically kernel-level readers like Speakup, this is ideal.
The TUI (The Grid): This treats the terminal window not as a stream of text, but as a 2D grid of pixels, where every character cell is a pixel. It abandons the temporal flow for a spatial layout.

Case Study: The `gemini-cli` Madness

Let's look at a concrete example: gemini-cli, a tool written in Node.js using the Ink framework. On the surface, it looks like a simple chat interface. But underneath, Ink is trying to reconcile a React component tree into a terminal grid.

When you use this tool with Speakup (Linux) or NVDA (Windows), the application doesn't just fail; it actively spams you.

Because the framework treats the screen as a reactive canvas, every update triggers a redraw. When the AI is “thinking,” the tool updates a timer or a spinner. To do this, it moves the hardware cursor to the timer location, writes the new time, and moves it back.

For a sighted user, this happens instantly. For a screen reader user, this is what you hear: “Responding... Time elapsed 1s... Responding... Time elapsed 2s... [Fragment of chat history]... Responding...”

It drives the screen reader mad. The cursor is teleporting all over the screen to update status indicators, spinners, and history. Speakup tries to read whatever is under the cursor at that exact millisecond. You end up hearing random bits of conversation mixed with timer updates, making it impossible to focus on what you are actually typing.

Worse, lets pretend that you've somehow managed well with speakup so far, but that you want to do some work with nvda. Maybe paste an error you're getting on windows. So you open your terminal, ssh into your linux box, attach to your screen session and paste your text.

The result is an immediate crash of the screen reader (NVDA) or massive system instability. Why? Every time you type a character or paste text, the application triggers a state change. The framework decides it needs to re-render the interface. Because the conversation history is part of that state, the application attempts to redraw or re-calculate the layout for thousands of lines of text instantly. The more messages you have in a conversation, the more this will happen. And no, you can't just avoid this by using insert+5, the key combo supposed to avoid announcing dynamic change of content.

The Lag Loop

Furthermore, frameworks like Ink running on single-threaded environments (like Node.js) suffer from massive performance degradation when the history grows. If you paste a large block of text, the system has to calculate the diff for thousands of lines.

This causes input lag. You press a key, and you wait. You can wait up to 10 seconds for a single character to echo back. The system is too busy calculating how to redraw the screen to actually process your input.

Why The “Old Guard” Works (`nano`, `vim`, `menuconfig`)

Sighted developers often ask: “If TUIs are bad, why do you use nano, vim, or menuconfig?”

The answer is not that these tools handle the cursor perfectly by default. The answer is that they allow you to hide the cursor entirely.

1. Hiding the Cursor (`nano`, `vim`)

In tools like nano or vim, usability depends on turning off features that track cursor position. If you run nano with options that show the cursor position (like --constantshow), or if you use vim without specific configuration, the experience is broken.

When the cursor is visible and tracking is active, Speakup prioritizes the cursor's location update over the character echo. Instead of hearing the letter “a” when you type it, you hear “Column 2”. You type “b”, and you hear “Column 3”.

These older tools succeed because they allow you to disable this noise. You can configure them to suppress the visual cursor or status bar updates, forcing the screen reader to rely on the character input stream rather than the noisy coordinate updates. Modern frameworks rarely offer a “no-cursor” or “headless” mode; they assume the visual cursor is essential.

2. Single Column Focus (`menuconfig`)

Tools like the Linux kernel's menuconfig work because they enforce a strict, single-column focus. Even though there are borders and titles, the active area is a vertical list. The cursor stays pinned to that list. It doesn't jump to the bottom right to update a clock, then to the top left to update a title. The spatial complexity is kept low enough that the screen reader never gets “lost.”

3. The Lost Art of Scrolling Regions (`Irssi`)

Irssi is the gold standard for accessible chat, but not because of luck. Irssi was built over 20 years with a custom rendering engine that utilizes VT100 Scrolling Regions.

When a new message arrives in Irssi:

It tells the terminal driver: “Define a scrolling region from line 1 to 23.”
It sends a command: “Scroll up.” The terminal moves the bits up.
It draws the new text at the bottom of that region.

Crucially, it handles this in a way that minimizes interference with the input line. It relies on the terminal's hardware capabilities rather than rewriting every character on the screen manually. Modern frameworks ignore these hardware features in favor of “diffing” the screen state and rewriting characters, which is computationally heavier and hostile to accessibility.

The “Stale Bot” excuse: A Case Study in Neglect

Google and the maintainers of gemini-cli pretend to care about accessibility. “Pretend” is the operative word here. If you look at the repository, critical accessibility regressions like Issue #3435 and Issue #11305 have been left to rot. There is no discussion, no roadmap, and no fix. Even worse is the fate of Issue #1553, which was supposed to track these accessibility failures. It didn't get solved; it got silenced. It was closed automatically by a bot with this generic dismissal:

Hello! As part of our effort to keep our backlog manageable and focus on the most active issues, we are tidying up older reports. It looks like this > issue hasn't been active for a while, so we are closing it for now.”

This is unacceptable. Closing an accessibility report because the maintainers haven't touched it in months is not “tidying up”; it is hiding evidence. It effectively says that if a bug is ignored long enough, it ceases to exist. It boosts the project's “Closed Issues” metric while leaving the actual software unusable for blind users.

Conclusion

If you are building for the terminal and care about accessibility, stop using declarative UI frameworks that treat the terminal like a canvas.

The “modern” TUI stack has optimized for the developer's ability to write React-like code at the expense of the machine's ability to render text efficiently.

If you cannot guarantee that your application allows the user to hide the cursor, or if you rely on aggressive redrawing to show spinners and timers, you are building an inaccessible tool.

For the blind user, a dumb, linear CLI stream is infinitely superior to a “smart” TUI that lags, spams, and scatters the cursor across the screen.

航札（一） | JustGoIdea

Sat, 02 May 2026 20:05:36 +0000

航札（一）

02 May, 2026

按：读书遣日，偶有所感，辄随手札记。不成系统，聊备覆瓿。今略加铨次，辑而名之曰「航札」。

笔记之书

刘义庆《世说新语》、段成式《酉阳杂俎》、苏轼《东坡志林》、洪迈《容斋随笔》、张岱《夜航船》，皆笔记集结而成，读来不费力，却常得稀奇古怪之知。近现代此类佳作不多，鍾叔河《念楼学短》堪称翘楚。笔记之妙，不在体系森严，而在随手拈来，眼光到处，皆可成文。

喷嚏

今俗谓打喷嚏乃有人念我，或云「一想二骂三感冒」，本以为近代俚语耳。细读《容斋》，方知其来甚古。《诗经·邶风·终风》有「寤言不寐，愿言则嚏」，郑笺谓女子思我，我则嚏。洪迈因而叹曰，今人喷嚏便云「有人说我」，实古之遗语也。民俗之久远，往往藏于最琐碎处。

蜜人

《南村辍耕录》记「蜜人」事，云回回国有老人自愿舍身济众，不复进食，惟啖蜂蜜，久之便溺皆蜜。死后以石棺盛蜜浸之，封存百年，启封取用，可疗损折肢体。俗呼蜜人，番言即木乃伊也。读之怪诞，然古人所谓灵丹妙药，常在生死、身体、信仰之间游移。人之求愈也切，遂想象尸身亦可为药。

放火

「只许州官放火，不许百姓点灯」一语，出放翁《老学庵笔记》。田登作郡守，讳其名，州人不得言「灯」字，皆改称「火」。上元放灯，吏人出榜曰：「本州依例放火三日。」讳名本权力之小癖，一经传播，遂成千古讽刺。语言为权力所扭曲时，笑话往往最先看出破绽。

人心道心

大学士张英，有家训笔记《聪训斋语》传世，开卷第一句便是「人心惟危，道心惟微」。

此八字相传出《尚书·大禹谟》，乃舜授禹之十六字心传前半，后半为「惟精惟一，允执厥中」。然经清人阎若璩考据，《大禹谟》实为伪作。《荀子·解蔽》引「道经」亦有此语，足见其源远流长，不必系于一书。

张英以此训子女，释之甚精：「危者，嗜欲之心，如堤之束水，其溃甚易，一溃则不可复收也。微者，理义之心，如帷之映镫，若隐若现，见之难而晦之易也。」末云：「人心至灵至动，不可过劳，亦不可过逸，惟读书可以养之。」此语平正切实，可为座右。

川上

《论语》中最有余味者，或为「子在川上曰：逝者如斯夫，不舍昼夜」一句。后世写流水、写时光、写兴亡，大抵皆可追溯于此。杜甫「无边落木萧萧下，不尽长江滚滚来」，东坡「大江东去，浪淘尽，千古风流人物」，杨慎「滚滚长江东逝水，浪花淘尽英雄」，其源头皆在川上一叹。夫子此句好在不加议论，只将天地流行指与人看。

繁简

繁简之争，不可只以便利衡之。按《说文》，今之简体「礼」反是古文，「处」「与」亦古已有省体。然「葉」简作「叶」，则失其音义；「鐘」「鍾」并归「钟」，亦多生误解。繁简之间，有造型，有语源，有历史层累。文字非死物也，乃文明留下之纹理。若只以书写之快慢论之，便把文化看小了。

造化鍾神秀

杜甫《望岳》「造化钟神秀，阴阳割昏晓」，今人见「钟」字，每解作钟情，谓造化偏爱泰山。若还其本字，作「造化鍾神秀」，则义明矣。「鍾」者，聚也。大自然将神奇秀丽汇聚于斯，非钟情于斯也。钱锺书先生不愿被写作钱钟书，其意正在此。一字之差，足使诗意改道。

悼亡

自《诗经》以降，悼亡之传统深植文脉。《唐风·葛生》开其先声。

年少时表姐授我诗词，言其独爱东坡「小轩窗，正梳妆，相顾无言，惟有泪千行」之深情。然我更为贺铸「空床卧听南窗雨，谁复挑灯夜补衣」所动。常言生活中总见真着，人之大悲不在生离死别之际，实在晨昏琐碎之中。

唐宋以来，元稹以「惟将终夜长开眼，报答平生未展眉」立誓，梅尧臣借「窗冷孤萤入，宵长一雁过」造境，纳兰「赌书消得泼茶香，当时只道是寻常」化典无痕，俱为悼亡之佳构。乐天「手携稚子夜归院，月冷空房不见人」，虽稍显平实，亦自有其沉痛。

至于义山「何当共剪西窗烛，却话巴山夜雨时」，是否确属悼亡，历来尚有争议。若果为王氏殁后追思之作，则当推千古悼亡绝唱之首——以未来虚拟之欢聚，写当下刻骨之永诀，时空折叠而成之艺术张力，实非常人所能及也。

西方哀歌

西方虽无「悼亡」之固定名目，哀歌传统亦深且远。

俄耳甫斯追寻欧律狄刻于冥府，已见失爱之殇痛。及至浪漫派，个体情愫勃发，悼亡遂为诗中重要一脉。雪莱作《阿多尼斯》，丁尼生为挚友写《悼念》，磅礴之思与细腻之情兼具，探讨生死，寻求慰藉，开掘至深。

后世如哈代忆亡妻艾玛，字里行间尽是日常琐碎之回放。刘易斯《卿卿如晤》，以至诚笔触记丧偶后之悲恸与信仰挣扎，读来令人心折。

与东方之婉约含蓄不同，西方悼亡往往更为坦率直切，兼融哲学与宗教之思辨。然无论中西古今，因爱别离之锥心、对逝者之绵长思念，皆为人类所共通，于墨痕诗行中觅得永恒寄托。

隽永

何谓文字隽永？庾信《枯树赋》末云：「昔年种柳，依依汉南；今看摇落，凄怆江潭。树犹如此，人何以堪！」二十四字，千载之下犹觉苍凉。柳河东《江雪》亦然：「千山鸟飞绝，万径人踪灭。孤舟蓑笠翁，独钓寒江雪。」东坡叹其「殆天所赋，不可及也」，诚非虚语。隽永者，不必多言，而余意自生。

一语成谶

晋灭吴后，武帝司马炎谓孙皓曰：「设此座待卿久矣。」孙皓从容反讥：「臣于南方，亦设此座以待陛下。」一时以为笑谈。孰料后五胡乱华，晋室果南渡建业，改称建康，竟应其语。

寇老西少年时作诗，有「到海只十里，过山应万重」之句。后果贬死雷州，去乡万里，一语成谶。

东坡再过渑池，赠弟诗云：「人生到处知何似，应似飞鸿踏雪泥。」此句飘逸洒脱，却暗合其后半生宦海沉浮、颠沛流离之际遇。好句有时不只是好句，亦是命运之预言。

屈原食谱

论中国文人美食家，世人常举东坡、汪曾祺，鲜有人想到屈原。《招魂》《大招》虽为祭祀之辞，若当食谱读之，亦极丰盛。谷物、肥牛、鳖羔、鸿凫、甜点、冻饮、琼浆，陈列满前。楚辞之华美，不只在香草美人，亦在饮食声色。魂兮归来，先以口腹招之，此亦人情之至也。

有香气的咖啡

葡国有所谓 café com cheirinho，即加了酒的咖啡。cheirinho 者，「香气」之谓。里斯本人称 bica com cheirinho，马德拉与亚速尔群岛则称 café com música——「带音乐的咖啡」，名目妙不可言。所加之酒，或为 aguardente，或为 bagaço，或为 medronho。咖啡得酒，既暖身，又添兴，民间更谓可疗感冒喉痛。口腹之物，最能保存一方性情。

七月半

农历七月半，古有尝新祭祖之礼。《礼记·月令》云：「是月也，农乃登谷，天子尝新，先荐寝庙。」道家以此日为地官赦罪之期，佛家称盂兰盆节，又为僧团自恣之日。三月结夏安居后，僧众解夏忏悔，故又名佛欢喜日。虽已入佛门，亦须循规蹈矩，方得大自在。此语甚好。

月食

乙巳年初秋，中元甫过，适逢月全食。古人视血月为兵凶饥馑、天下巨变之兆，非中国独然。古希腊、古罗马乃至葡萄牙民俗中，皆有类似之说。葡人有所谓 mal de lua，亦有驱赶食月之兽的故事。中国谓天狗，葡国换作狮子，名目不同而恐惧相通。

不同文明隔绝万里，却同以天象之昏暗为不祥。或许所相似者，非迷信本身，而是人类面对黑暗与遮蔽时心中原型之恐惧。天象本无意，人心自有象。

福尔摩沙

欧洲拉丁语系诸国中，至今仍有人称台湾为 Formosa，或 Ilha Formosa。传十六世纪葡萄牙航海者初见台湾，叹曰：「Ilha formosa！」美丽之岛也。中文译作福尔摩沙，音义皆有余韵。地名犹如一枚贝壳，历史退潮之后，仍留海声。

朋友之义

《容斋》论朋友之义，历举《诗》《中庸》《孟子》《论语》诸书，末叹云：「本朝百年间，此风尚存。呜呼，今亡矣！」读此可知，南宋时朋友之道已为世风所伤。

又忆及反右之时，冯亦代常往章伯钧、费孝通诸人家中走动，貌为友朋，实则卧底。朋友若沦为耳目，义从何来？古人一叹尚止于世风浇薄，后世之祸，则是有人刻意以友谊为凶器。

一念

一九五〇年，张爱玲着旗袍赴沪上文代会。满座皆列宁装，独她一人异服。此一瞥之间，已窥全豹。去意渐生，两年后南渡香江，复远赴洛杉矶，终老于斯。

老舍彼时在美，写作售稿，生计无虞，亦有赴英执教之邀。然周公故交赵清阁殷殷劝归，竟从之。十余年后，太平湖畔，独坐半晌，终蹈水而殁。那日亦有相识在旁，无一人往劝。

人生抉择往往在一念之间。张爱玲之一念，老舍之一念，一去一留，天壤之别。世事之残酷，不在刀兵相加，而在归来时已无路可退。

敌在本能寺

日本战国之末，织田信长几近一统天下，威势赫赫。天正十年，驻跸京都本能寺，仅率少从。麾下大将明智光秀素为信长所倚重，手握重兵。孰料奉命出征途中，忽调矛头，率众夜袭。信长于烈焰中殒命，史称「本能寺之变」。

防外敌易，防家贼难。危亡之兆，常起于萧墙之内。故后人以「敌在本能寺」喻之。权力之巅与信任之危，从来比邻而居。古今皆然，今时今刻，恰如彼时彼刻。

笔记与著述

古人论著述，有两途焉：一曰体系，一曰札记。体系者如经学注疏，层层架构，丝丝入扣；札记者如随笔杂录，随感而发，不拘成法。今人治学问亦然，或建宏大之系统，或积零碎之心得，二者不可偏废。

然札记之难，世人多不察。以为随手可就，其实不然。信息被捕获，不过是起点；若不能化入自己的语言，接入旧有经验，生出新的判断，便只是搬运之功。读书而不思，犹饮食而不化，终为肠胃之累。

写作亦同理。思维本非线性，而文字必须线性呈现，其间之转化即是功夫所在。真正的写作，是在发散与收束之间往返：先见联系，再探本质；写而回看，重组再写。手慢一点，有时正是思想沉下去的机会。昌黎所谓「行成于思」，不外如是。

世事不宜深考

大国之间的折冲樽俎，公众解读往往两极：或曰「赢麻」，或曰「丧尽」，情绪之张力远大于理性之审思。此中固有当轴者对「从胜利走向胜利」之叙事渴求，亦有媒体对耸动标题之本能追逐。

然而棋局再宏大，终是由人来下。那些于夹缝中艰难求存之个体——卑微、渺小、羸弱——其挣扎与叹息，谁人倾听？「莫管他人瓦上霜」之冷漠，依旧是底层真实的处境。

殊不知，当前种种精巧之运筹与刻意营造之声浪，皆是掷出之回旋镖。看似飞向远方，终将循迹而归。

妙绝千古

诗词曲中，私以为堪称「妙绝千古」者，略举数联：

鸡声茅店月，人迹板桥霜。——温庭筠《商山早行》

海日生残夜，江春入旧年。——王湾《次北固山下》

无可奈何花落去，似曾相识燕归来。——晏殊《浣溪沙》

一川烟草，满城风絮，梅子黄时雨。——贺铸《青玉案》

枯藤老树昏鸦，小桥流水人家，古道西风瘦马。——马致远《天净沙·秋思》

兴，百姓苦；亡，百姓苦。——张养浩《山坡羊·潼关怀古》

此数联或以意象胜，或以气势胜，或以理趣胜，皆能以极简之笔触，拓极深之境界。妙处难与君说，惟读之再三，自能会心。

七二〇

一九六七年七月二十日前后，文革浩劫之际，江城发生「七二〇事件」。彼时毛泽东宿东湖宾馆，几为军中异己所拘系，险遭软禁之厄。周恩来星夜驰赴，方解危局。此乃毛氏问鼎后最惊心动魄之一幕，与一九七一年沪杭「龙潭虎穴」之险，实不相上下。弄权者自以为驾驭天下，殊不知反噬之力亦从权力内部滋生。

天不惜才

读王鼎钧《昨天的云》，提及许地山，忆起幼时所学《落花生》，余韵悠长。许氏卒年仅四十七，实是天不惜才。陈寅恪撰挽联云：「人事极烦劳，高斋延客，萧寺属文，心力暗殚浑未觉；离乱相依托，娇女寄庑，病妻求药，年时回忆倍伤神。」读之怆然。

又读王勃事迹，再叹天不佑英才。魂断沧海，年方二十八。传说其为文之前，先蒙头酣睡，忽然起身，笔走龙蛇，一气呵成，不复修改而华彩锦绣。或谓此等天授之才，本不宜久留人间。然终究是惋惜多于豁达。

星经

《史氏星经》为中国现存最古之星表。近读材料，有学者考其年代或可追溯至公元前四世纪，较旧说更早约二百五十年。若此说成立，则或为世界已知最古之星图。

古人仰观天象，不独为占验吉凶，亦是于无边黑暗中替人间寻秩序。星辰无言，而秩序感本身，已足以安顿人心。

身与心

孙隆基《中国文化的深层解构》有一段精妙之论，大意谓中国人非一个人便能完成自身，须经由「二人结构」——一人以心关照另一人之身，另一人亦如此回报，双方交心而安身，安身而安心。

故中国人之问候，必曰「你身体可好」「别来无恙」，所问皆身；而对方答以「你有心」「谢谢关心」，所谢皆心。一问一答之间，身心之学尽在其中。此与西方先确立独立个体再建立关系之路径，恰好相反。

凋零

港圈「黄金配角」一位又一位凋零，那个时代的风光，已如潮水退去。

戊戌年间，曾陪香港某制片人在京城看展，他说：「香港哪里还有以前的气象。」彼时不甚理解，甚而觉其言论偏激。后来目睹反送中与种种后续，方才若有所悟。

时代之一粒尘，落在寻常人头上，便是一座山。经济、政治、文化、社会，走到这一步，似乎皆属自然。然而见过维港之绚烂，听过街头之悲鸣，总有些不甘。再想，又有什么好不甘的呢？

一切皆无常。人终将老去、死去。那些绚烂过的、悲鸣过的，都曾真实存在。如此便够了。

文明与族群

今人有谓刘邦苗族、杨坚鲜卑、李唐沙陀、朱元璋回族云云，遂断言历代大一统王朝皆少数民族政权。此说貌似新颖，实则犯了根本性的概念错置。以现代民族国家之分类框架，强加于古代文明秩序之上。岂不可笑？

先秦时无现代意义之「民族」观念。所谓诸夏、戎、狄、蛮、夷，本是政治文化之标签，非血统之分类。一部族入华夏礼制秩序便是「夏人」，背离此秩序则谓之「夷狄」。《春秋》之义甚明：诸夏用夷礼则夷之，夷狄用夏礼则夏之。族属不由血统决，而由行为、文化、制度界定。此种认知框架，远比后世西方民族国家之血统论更为开放。

从秦至隋，「汉人」主要指一种政治文化认同。魏晋南北朝时，汉与鲜卑、羯、氐、羌于北方长期混居通婚，制度语言服饰互相塑形。孝文帝改革后，「汉」渐从地域性之文化认同，演为更广泛之文明共同体。「汉族」一词首见于唐代官方文献，以指称长城以内之居民，与「胡」相对；确立为法定民族身份，则已是清末民初之事。

故「汉族」本即复合型民族，是文化认同之概念，可视为「中华民族」观念之早期形态。入此秩序、认同此秩序者，皆为其成员；离之者，即为「他者」。中华文明之连续性，正建立于「文明高于族群」这一认知基础之上。

吴国光先生所提「海华族」之概念，亦可视为此框架之现代延伸。生活在中国以外之华人，无论国籍如何变化，只要认同华夏文明之价值与传统，便仍属这一文明共同体。

礼与正统

一个政治实体，若须靠精心炮制幻觉、对内进行认知作战，以维系其统治，则此实体之气数已可见端倪矣。

中国数千年之传承，核心在一个「礼」字。礼者，仪式也，秩序也，典章也，教养也。正统之最大合法性，即在于维护、遵循、发展此「礼」。守礼则正统不待自证，外人无可置喙。

犹记杨老虎于阿拉斯加对美方之言，何以豪气干云而不显突兀？何以连意识形态对立之媒体亦予正面报道？盖因彼时彼刻，杨氏占着一个「礼」字。主人失礼在先，客人据理以争，天下公论自在。美方纵有不甘，亦只能受之。礼之为用，不在声高，而在理直。

面包与理想

金庸写道：有人的地方就有恩怨，有恩怨就有江湖，人就是江湖。

理想是需要的，面包也是需要的。然而二者孰先孰后？理想之代价，极大的可能正是面包。若为理想而不顾面包，则中道崩殂近乎必然。更堪讽刺者，通往理想之路与获取面包之路往往背道而驰——为理想则失面包，求面包则弃理想。

所谓伟人，最要紧处在于能自洽地暂弃理想以得面包，复以面包哺育理想。此等腾挪功夫，常人万难为之。

然则于理想与面包之间，终须以最基本之人性为出发点。再宏大之愿景，再精巧之路径，背后皆是人，皆是恩怨，皆是江湖。伟人写下宏大的历史，常人活出具体的人生。江湖之中，能守住对人性的基本尊重，已属不易。

小民尊严

「没有大国崛起，哪有小民尊严」一语，流传甚广，细究之则谬甚。

国之强盛与民之尊严，非因果之关系，亦非充分或必要之条件。国家机器之威势，衡量的是其在国际体系中之位置；个体之尊严，系乎法律与社会结构中权利之实际保障。二者不同质，且往往此消彼长。

更堪玩味者，「小民」二字本身即预设了等级秩序，仿佛个体天然渺小，须依附宏大叙事方能获得价值。此正是以集体之名消解个体主体性之话术。

人之尊严，不待外在证成。法治之保障、权力之制衡、言论思想之自由、财产之不可侵犯……凡此种种，皆与「大国崛起」无涉，反常与之相悖。《孟子》所谓「民为贵，社稷次之，君为轻」，先贤早已洞见此理。奈何后世偏要把次序颠倒过来。

载道之辨

「文以载道」四字，世人多以为是道德教化之本旨。然细思之，其中恐怕更多的是话语权之垄断。

「克己复礼为仁」、「安贫乐道」、「存天理，灭人欲」，此类说教之本质，在于进行欲望之阶级分配：大众不宜执着个人情感，个人才华当用于颂扬圣德，「风花雪月」乃有资格享受者之专属。所要消灭的并非欲望本身，而是要把欲望之满足变成特权之标记。

古之儒者如此，后之当轴者亦如此。道统之名义虽殊，垄断之逻辑则一。

译经双杰

罗什三藏创造性地运用汉字组合表达抽象概念，以意译之法，将佛教辩证思维化为中文之表达习惯。所译经论如《金刚经》《维摩诘所说经》，文辞流丽，义理通达，至今仍为最广流通之版本。

玄奘大师则以精确音译配严谨注疏，赋予中文系统分析意识结构之能力。其所创译名如「阿赖耶识」「末那识」等，虽初见生涩，却为汉语开辟了精密思辨之新疆域。

简言之，罗什以意译化外来思维为本土语汇，玄奘以音译建概念分析之严谨框架。二公所奠之语言基础，非仅译事之功，实为中文思维能力之一次升级。

拜菩萨

世间有一桩常理，好多人一辈子不曾想透：倘若拜菩萨当真有用，便没有人会去拜菩萨了。

此语看似悖论，其实不然。正因为世事无常，求而不得是常态，人才需要一个寄托。菩萨之灵验与否并不重要，重要的是那一跪一拜之间，人暂时放下了自己无法掌控一切的焦虑。信仰之用，不在于改变世界，而在于安顿此心。

Previous |

#notes

How Big Is the American Dream House? - The Atlantic

Fri, 01 May 2026 21:39:11 +0000

This is an edition of Time-Travel Thursdays, a journey through The Atlantic’s archives to contextualize the present. Sign up here.

“When I was nine or ten and lived in a dark fourth-floor apartment in a building that had seen better days, I fantasized mansions that were more suited to my romantic nature,” Linda Lewis wrote in The Atlantic in 1978. In adulthood, she got only more covetous—of friends’ gorgeous houses, of French castles, of architectural marvels such as Monticello and Frank Lloyd Wright’s Taliesin West.

As a fellow house envier, I sympathize. I feel, as Lewis did, that I “behave differently in different kinds of rooms”; that I’m “powerfully influenced by shapes and sizes, light and color, by degrees of privacy and security and beauty.” But I’ve never wanted a mansion, let alone a Monticello. I own a row house, and when I fantasize about something more suited to my romantic nature, what I’m picturing is a slightly bigger row house in a neighborhood with more restaurants.

This is very Millennial of me. My generation of Americans is the first in decades to collectively abandon the dream of a big house. In part, that’s likely a concession to reality: Real estate is so expensive that homeownership is, for many, a fantasy. But it also reflects changing ideas about what makes a good house and a good life, for both renters and owners. According to the National Association of Realtors’ most recent survey of American housing preferences, the majority of Millennials and Gen Zers would rather live in smaller homes in more walkable communities than larger ones in less dense areas. As a country, though, we aren’t building accordingly.

In 2012, the architect Daniel Parolek coined the term missing middle housing. What’s lacking in America is the “middle scale of buildings between single-family homes and large apartment or condo buildings,” he argued in his 2020 book. Row houses, which are built in a continuous line and tend to be smaller than the average single-family detached home, are emblematic of the missing middle. Although they’re in high demand, they represent less than 20 percent of new construction, according to the National Association of Home Builders. In part, this is due to zoning issues: Parolek’s book notes that many cities’ codes for house construction usually call for “minimum lot sizes that are too large, densities that are too low, [and] parking requirements that are too high” for attached homes.

Row houses have a prestige issue, too. Before the advent of car ownership and the suburban lifestyle it facilitated, row houses were popular with families of many classes and backgrounds; once more families began migrating farther from cities in the 1910s and ’20s, row homes were usurped by detached houses with lawns. As the latter became emblematic of comfort and success, the former came to be seen as down-market or second-class. Meanwhile, “over the course of the 20th century, government policy, the invention of cheaper, mass-produced building materials, marketing by home builders, and a shift in how people regarded their houses—not just as homes, but as financial assets—encouraged ever larger houses,” Joe Pinsker wrote in The Atlantic in 2019. The architect Witold Rybczynski observed in this magazine in 1991 that the average new single-family house had grown by more than a third from 1963 to 1989. From 1989 to today, it has grown even larger, now averaging around 2,100 square feet.

Rybczynski considered this increase in home size a mistake. In his view, row houses are an ideal design. Dividing one of them into multiple apartment units, or into mixed-use living and retail, is easy; their thick shared walls can reduce heating and cooling costs because fewer exterior-facing facades means less exposure to the elements; and they use less land than single-family homes, which means they’re usually more affordable. Of course, not everybody likes the trade-offs. Plenty of home buyers still want more backyard space, more rooms, more parking, more to show for their expensive mortgage.

Size aside, there are the matters of darkness and noise. Detached houses get more sun because they generally have windows on all four sides, and not everyone wants to rely on a white-noise machine, as I do, to drown out my neighbors watching Bravo on the other side of the wall. But with those annoyances comes what Rybczynski calls “the gregariousness of living in relatively close proximity.” Encountering a single block of row houses in isolation is rare; more frequently, they make up whole neighborhoods. As Parolek told me, the dream neighborhood is “the American dream house for a majority of American households now”—and they’re happy to live smaller, and deal with some secondhand Housewives, to get there.

When I asked Elizabeth Plater-Zyberk, an urban planner and a professor at the University of Miami’s School of Architecture, about her vision of the ideal American home, she brought up a development called Kentlands, a townhouse-centric development in Gaithersburg, Maryland, roughly a 45-minute drive from downtown Washington, D.C. The community is designed to be walkable, with plenty of shared space, meaning that “the entertainment portion of the house” is effectively outside the home, she explained.

At first I couldn’t quite imagine what she meant. Then I thought about the neighborhood restaurant where I met with a parents’ group when my daughter was an infant. I thought about the playground where we see friends weekly, the public picnic grove where we hosted her most recent birthday party. I have easy access to those places because, like Kentlands residents, I live in a dense and walkable area. The reason my fantasies don’t extend beyond a bigger row house is, I think, because I don’t want to lose that kind of access. What’s more, it strikes me as entirely possible that if I hadn’t been raised in the ’90s era of big American homes—in a country and culture that gave me the expectation that, as a grown-up, I’d have a guest room and a yard to mow—I’d never think about moving out at all.

Do Women Reproduce More than Men? - by Tomas Pueyo

Thu, 30 Apr 2026 19:19:45 +0000

Previous premium article: Why Did Islam Spread So Fast in North Africa & Spain?
Next premium article: An anlysis of all the news on the Game Theory of Sex & Relationships

Here’s a crazy fact I read: 80% to 90% of women reproduce, but only about 40% of men do. What?!

It made me think of this:

This represents the idea that, nowadays, women can easily have sex, so they gravitate towards attractive men, which is bad for unattractive men (no access to sex), and bad for all women (heightened competition for a few men only), while attractive men get all the access to sex they want and end up not partnering up.

Are these things true? Do men really reproduce less than women? Is it because historically, the more attractive men could hoard the women, leaving the least attractive men out of luck? If that’s the case, how frequent was it? How did different societies deal with this problem? Did it lead to violence? And what about today? Do more women than men still reproduce? How has technology like Tinder or contraception changed this? What can we predict about the future of sex, reproduction, and violence?

1. Men Don’t Reproduce as Much

Let’s start with the initial claim: If 80% of women reproduce and only 40% of men, that means women have been 2x more likely to reproduce than men in history. Is this true? This is an estimate of how many women1 reproduced in a studied population throughout history:

Notice how the human population starts exploding about 70,000 years ago, around the time of the 2^nd move out of Africa. Then, there’s a new explosion around 10,000-12,000 years ago, around the time of the end of the last ice age.

But this is for females. How did males reproduce in comparison?

I had to collapse the vertical axis for males because in the original paper, the researchers used different scales, which defeats the purpose of comparing them!

Assuming that 80% of women reproduced at any given time, what does this tell us about the share of men who reproduced?

This is bonkers!! It means historically only about a quarter of men reproduced!?2

This maps tells you where:

In Europe and Eurasia, the male source individuals were 85-95% fewer than the female source individuals! Red numbers: estimated female effective population size; blue numbers: estimated male effective population size; top number near each triangle: “current” effective size for that regional sample; bottom number near each triangle: ancestral/founding effective size for that regional lineage; black oval “26 / 15”: estimated effective size of the initial out-of-Africa bottleneck: ~26 females, ~15 males; black arrow dates: inferred divergence times in years before present; arrows schematic ancestry/migration splits, not literal routes. Source.

How is this possible?

Homosexuality

About 3% of men consider themselves fully homosexual, so it can explain only a tiny part of this gap.

Polygyny

The most logical explanation for this would be polygamy (one person marrying several people of the other sex), which generally means polygyny (one man, several women).3 If the average man who marries has 3 wives, it means about ⅔ of men won’t have a wife.

The Sexual Violence Problem

This causes a lot of conflict. Because you get a lot of men who can’t access sex and reproduction. So societies had three options.

One was to just be violent, with men killing other men and kidnapping and raping women. That is not very stable at all, so such societies quickly learn to focus the violence outwards: They organize in clans and tribes where you’re not supposed to kill each other, and should raid your neighbor instead.

The societies that managed to do this at scale were able to spread extremely fast. Examples include the Vikings and the Arab Muslims: In both cases, they had serious polygyny, and men were recruited to go raid foreigners and take their women as partners, concubines, or slaves. The Muslims had the added brilliant idea of telling recruits: “Even if you die fighting, don’t worry, you’ll get your women in heaven.”

For example, we know from DNA analysis that 80% of the Icelandic male DNA was originally from Scandinavia, whereas 60% of women’s is Gaelic.4 So the most common way that Iceland was settled was with Scandinavian men taking Gaelic women in the British Isles and settling in Iceland. Something similar happened in Arab territories:

The Y chromosome is mostly handed down from father to son without much modification, which helps us track male lineages across many generations. Similarly, mitochondrial DNA is handed down from mother to children with no modification.

So very certainly part of the difference between male and female reproduction is that some men hoarded women, so other men either didn’t have access to women and didn’t reproduce, or had to go to war or raid neighbors to get women. They would either die trying, or succeed and extinguish the male line of the conquered.

It doesn’t always need to be violent though. For example, the Swahili Coast in Africa has in some places up to 80-90% of male ancestors from Persia, India and Arabia, with nearly 100% local female ancestors, from Africa.

Swahili coast, in Africa

This was from trade: Rich Muslim traders would come and marry local women. But slaves were part of the traded goods, and so some of the partnerships would have been coerced rather than free.

Conversely, some Arabic ancestry comes from African women, but not men. Given that Arabs traded African slaves, the most logical explanation is that they didn’t let African men reproduce in Arabia,5 but some of the African women slaves did reproduce with the men.6

Age

This one blew my mind. There’s a second way a polygynous society can manage the sex imbalance: Growth plus age differences. This explains why old men marry young women in Muslim countries.7

If you have 100 men, and they each have on average two wives, you need 200 women. How can you get them? You can solve that if the men are 40 years old and the women are 20 years old.

The women have four children each, for a total of eight children per father—approximately four boys and four girls. That means that our total population of 100 men and 200 women becomes, 20 years later, 800 men and 800 women.

The women marry immediately and have children,8 but not the men! The 800 men have to wait 20 more years. By the time they do, there’s been a new generation. Now there are 1600 men and 1600 women aged 20: Enough women so there’s two of them per man!

Another way to think about this: If every generation is twice as big as the previous one, every man can simply marry, when he is older, two younger women of the next generation.

Purple is women, blue is men. In generation 1, there’s one 40 year old man and two 20 year old women who marry, and have eight kids in the 2nd generation (four kids per woman), four girls and four boys. The four girls marry when they are 20 with two 40 year old men from Gen 0. They have together the 3rd generation, made of 16 children (again, four kids each), eight girls and eight boys. The eight girls in Gen 3 marry the four boys from Gen 2, and have together 32 kids. And so on.

So by marrying women young enough, remaining fertile, and having lots of children, you can actually maintain a highly polygynous society while all men still have women!

Something like that has happened in high-growth polygynous societies across the world. It’s one of the reasons I believe old men marry young girls in some Muslim countries. I don’t like it (and with children it’s abhorrent), but that’s the logic.

This wouldn’t explain why men have had more children than women though—in fact, it increases the mystery, because it could explain away polygyny without the need for women reproducing more than men.

Monogamy

The other way to handle this problem is by forbidding polygamy. If everybody is monogamous, everybody can pair up and have children. Yay!

This has a big advantage: Instead of pouring resources (men) into violence against each other to fight for women, who are then raped and / or kidnapped, you can focus all the energy on building stuff and accumulating wealth.

This might be one of the reasons why the Greeks (and later, Romans), who were uniquely monogamous, were so successful. They imbued this into Christianity9, which then spread the concept around the world.

Islam and China were not technically monogamous, but in practice the vast majority of marriages have been, I assume for the same reason.

A study looked at the ratio of men to women in several societies, and assuming 80% of women have had surviving descendants, then in East Asian societies about 72% of men had surviving descendants; in Europe it’s 62%, and in Nigeria (Yoruba) it’s 57%. So we can see in the data that monogamy might have indeed had a strong impact in balancing reproduction and descendants—which gives credence to the idea that polygyny was originally one of the causes of this imbalance.

But if monogamy has been so widespread, why do we see such an imbalance in male-female reproduction across the world until so recently?

2. Patriarchy against Men

The most common structure for clans has usually been patrilineal: All the men in a family stay together, and the daughters leave to pair up with men in other clans. This means all men in a clan are related by blood, but not all women.10 So if a clan disappears—whether it’s wiped out, it starves, or it slowly shrinks because it doesn’t have access to resources or status—then the entire male line goes extinct. But not the female line, which is much more mixed! So, in effect, patriarchy is genetically much riskier for men than for women.

Meanwhile, this effect was compounded by the fact that successful clans would wipe out many other clans, further reducing genetic diversity.

And then, these successful clans would grow too much and would need to split. When they did, they tended to do so along lines of close relatives. So for example each successful son would move out with his male descendants. The result is that every split would increase male line genetic concentration.

One paper tried to figure out how much the effects described in this section could explain the gap between male and female reproduction, and it found that it could theoretically explain it all. In other words, you don’t theoretically need polygyny or violence between men to explain this gap. The clan structure could be enough.

My guess is that it’s something in between:

Polygyny meant that many men could not get wives and never reproduced.
Many men killed other men and took the women, further reducing the share of men who would ever reproduce (or killing off their offspring). This was more common in polygynous societies than in monogamous ones.
Even if some males reproduced, clan structure could eliminate their entire male line, if the entire clan was wiped out.

What does all this tell us about today? Here are the consequences we can draw for dating apps, incels, polygamy in the West, immigrant profiles, Ukraine and Russia, and AI and robots.

You Don't Want to Make Things, You Want to Have Made Things | blog.spu.io

Thu, 30 Apr 2026 00:27:16 +0000

You Don't Want to Make Things, You Want to Have Made Things

28 Apr, 2026

One thing I’ve learned about myself over the years is that I want to watch movies, but I want to have watched tv shows. Of course it isn’t black and white like that, there are exceptions on both sides, but broadly speaking, with movies I’m looking forward to watching them, with shows I’m looking forward to finishing them - to having watched them. TV shows are just so long, man. Yes, I’m interested in this show, but am I interested enough to want to watch all of this?

I think that’s how a lot of people feel about making things.

For different reasons.

Some want the feeling of having made something. Without the hassle of making it. "I made this" meme meets "but it's hard" Scott Pilgrim gif.

Some think that making something is fun in the same way using it is. You see this a lot in video games. "I love playing games I'd love to make one." And yes, making games is fun, it's just a very different kind of fun than playing them.

Some want to have the thing, but they don’t enjoy the process of making it. This is different from "I made this", it's addressing a specific need. If someone else was making that thing they wouldn't even be thinking about making it.

And then there's, of course, the delusion you could make quick and easy money. When reality is neither quick, nor easy, nor money. For most people in the arts or entertainment, including people you've probably seen or heard of, having to quit their profession to become a gig economy food delivery driver represents a raise of hourly wages and overall income. And I'm talking about people who actually managed to make a living at some point - then there's also the ones who never get a paid gig in the first place.

You can read a lot of this pretty clearly between the lines (and sometimes straight up on the lines) of how tech companies are trying to sell generative “AI”.

I hate that so called "AI" now has to be part of this topic. In the past you'd try something, realize it isn't for you, come to terms with it and move on. Which is BIG! Admitting to yourself that something isn't for you, that there's something you can't do and being able to live with it and move on requires character. A healthy relationship with your ego. It's not easy. And now you have these snake-oil salesmen trying to tell you otherwise. Keeping you from growing as a person.

It was already hard enough to convince people that it's okay not to be good at everything, now it's next to impossible.

I guess this is somewhat related to yesterday's What People Get Wrong About Discipline

Previous Next

#thinking