374 notes this month | recorded from 2025-12-02 to 2025-12-30
Themes: Reading, Ideas and History 93 · Travel, Places and Cities 55 · Product, Engineering and Open Source 51 · AI and Agent Systems 48 · Daily Notes and Everything Else 47 · Self-Knowledge and Psychology 31 · Business, Investing and Career 25 · Body, Health and Daily Life 17 · Content, Craft and Recording 7
Everything from the month is kept here, filed by theme, each entry carrying its original timestamp.
Quick Navigation
Selected: Selected Notes of the Month · 4 entries
374 records this month, filed under 9 themes:
- Reading, Ideas and History · 93
- Travel, Places and Cities · 55
- Product, Engineering and Open Source · 51
- AI and Agent Systems · 48
- Daily Notes and Everything Else · 47
- Self-Knowledge and Psychology · 31
- Business, Investing and Career · 25
- Body, Health and Daily Life · 17
- Content, Craft and Recording · 7
Selected Notes of the Month
Kumano Kodo: History Accumulated Step by Step
2025-12-01 17:04:20
Kumano Kodo pilgrimage, immersive experience. The most rewarding part of solo hiking is being with yourself—just focus on yourself and the path beneath your feet, walk naturally without needing to “break through” yourself or nature. Respect yourself, respect nature.
Viewing hiking through a gaming perspective: I’m a player in this walking game. This is the “player mindset.”
Those massive stones weren’t laid by modern construction—they’re ancient paths from hundreds or even thousands of years ago. Huge, rough, moss-covered, they carry history you can feel physically. These stones weren’t designed to “look good”—they were walked out, pressed out, polished out over centuries.
History is not designed; it’s accumulated step by step.
#格物/熊野古道 #格物/历史 #观我
Nachi Waterfall as Shintai (神体)
2025-12-02 16:00:54
Kumano Nachi Taisha reflection: Nachi Waterfall is the shintai (神体)—the physical manifestation of deity. 133 meters落差,Japan’s highest direct-drop waterfall, worshipped as “the deity’s own body.”
At Nachi, gods don’t dwell in temples—they exist directly as “water.”
Shinto side: Worships Kumano-Fusumi-Okami and other Kumano deities, deeply bound with forests, waterfalls, and mountains.
Buddhist side: Right next door is Kegan-ji Temple, enshrining Guanyin Bodhisattva. You’ve probably seen the iconic view: three-story pagoda + waterfall.
Let different worldviews coexist in the same space-time, letting pilgrims align and integrate on their own.
#格物/熊野古道 #格物/佛学 #格物/神道
Nara vs Kyoto: Architectural Aesthetics
2025-12-02 16:00:54
Nara inherited magnificent Tang Dynasty architecture: massive dougong brackets, deep overhanging eaves, thick pillars.
Those enormous dougong aren’t just structural supports—they’re visual centers. They don’t need elaborate carvings; they震撼 you with the sheer weight and mechanical logic of wood itself. This “honest” architectural language is like backend code: stable, logical, no redundancy.
But Kyoto’s architecture reflects Japanese本土化 aesthetics: slender, restrained, integrated with gardens. To accommodate “borrowed scenery” (借景), Kyoto buildings deliberately downplay their mass. Pillars thin, roofs lighten, decorations multiply (or become extremely minimalist like tea houses).
Nara is like a cathedral—towering, making humans feel small, a grand narrative. Kyoto’s architecture guides you inward—through spatial limits and controlled light, leading you into meditation.
Nara is the only place on Earth to see authentic Tang Dynasty atmosphere (even better preserved than within China itself).
#格物/建筑 #格物/日本 #格物/美学
AI as the Modern Mental Pastor
2025-12-15 20:30:00
In modern society, traditional religious roles are being replaced. AI large language models have, in a sense, become “modern mental pastors”—people pour out their confusion to them, seek advice, and gain psychological support.
But this raises a question: AI’s advice is based on statistical patterns, not genuine understanding or care. Are we outsourcing spiritual reliance to soulless algorithms?
Perhaps the answer isn’t rejection, but conscious use. AI can be a tool, a mirror—but not the ultimate reliance.
True spiritual practice must return to one’s own heart.
#格物/AI #格物/佛学 #格物/心理学
1. Reading, Ideas and History
93 entries Suddenly I realise — under all language systems, if you dig down to the roots, wouldn’t that be more helpful for learning actually it feels a lot like programming languages they all revolve around some basic grammar; no matter what kind of language, they are solving the same problem, which is being able to describe this world Reference: there must be something that stands for “object/noun” (Who/What) Predication: there must be something that stands for “did what / is how” (Action/State) Space & Time: it must describe whether things happen in “past, present, future” and “here, there” Logic: it must express cause and effect, negation, hypothesis (because…therefore, not, if) And there are also some common cultural patterns for example, in almost all cultures, “up” is usually associated with positive, noble, more in quantity (High status, Rising prices); “down” is associated with negative, lowly (Feeling down) Kyoto vs Nara architecture Much of Nara inherits ancient architecture from the High Tang period huge dougong brackets, deep eaves, thick pillars The giant dougong you see are not only structural members holding up the roof, they are also the visual centre. It disdains excessive carved beams and painted rafters; instead it stuns you with the sheer weight of the wood itself and its mechanical structure. This kind of “honest” architectural language is a lot like low-level backend code — solid, logically clear, no redundancy But Kyoto’s architecture shows more of a localised Japanese aesthetic — slender, restrained, blended with the garden Kyoto buildings often deliberately play down their own mass in order to work with the garden’s “borrowed scenery”. The pillars got thinner, the roofs got lighter, the decoration got more (or became extremely minimal, like a tea room) Nara is more like a religious sanctuary — tall, so that people look small; that’s grand narrative Kyoto’s architecture guides you to look inward, using restricted space and controlled light to bring you into meditation Nara is the only place on earth where you can see the authentic atmosphere of the Tang dynasty (preserved even better than inside China) Yu Qian is the kind of figure in Ming history who “would rather die standing than bend” In 1449, Emperor Yingzong, Zhu Qizhen, was captured by the Oirats; the whole Great Ming was in an uproar — court, commoners, army all thrown into shock The whole country was panicking. Yu Qian was not He advocated three things: steady the capital, steady the army’s morale, and install Emperor Daizong (the Jingtai Emperor) to preserve the legitimacy of the regime. Almost single-handedly he pulled the Ming back from “possible national ruin” to the edge of the cliff. He moved troops, set up defences, fixed the strategy, like a battle-hardened commander-in-chief The Great Ming was, in the truest sense, “a country held up by people”, and that person was him A thousand hammer blows chiselled it out of the deep mountains, burnt by fierce fire as if nothing at all. Though shattered to pieces I am not afraid — I only want to leave my purity in this world. Epictetus said: “Some things are up to us, some things are not up to us” Being able to see the causal chain clearly appropriate participation is not someone else’s cause; when the shared cause is done well, the result forms naturally many outcomes we cannot predict all we can do is be good ourselves, be good in this moment the rest, leave it to fate ~ Historical experience tells China the cost of turmoil is too high The root driving many of China’s policies is actually stability; once you understand stability, you can understand many causal factors The high savings rate comes from a lack of adequate social security; real estate is treated as a growth engine and stabiliser; capital controls are there to avoid financial turmoil; the tightening and loosening of the platform economy is for a margin of safety China’s essence is: a civilisation-state centred on long-term stability and state capacity, whose governance logic has strong organisational power, low tolerance for risk and engineering thinking — this shapes the economic structure and the social experience The underlying logic of reform and opening up is also stability It was a stability-oriented reform to avoid systemic collapse, restore the capacity for development, and rebuild the social contract In the late 1970s, the old system could no longer sustain production and distribution, economic efficiency was near stagnation, finances were almost exhausted, and social sentiment was declining If things stayed as they were, instability was the biggest risk So reform and opening up was not change for change’s sake, but to avoid a long-term systemic crisis Seen from this angle, it is a “self-rescue stability project”: Household contract responsibility in agriculture → solve food and clothing, avoid the collapse of rural society Attracting investment → fill the gaps in capital and technology, avoid growth sinking into the mire Market mechanisms → activate the drive to produce, avoid rigidification leading to running out of food Devolving power to localities → let localities find their own food, avoid central pressure blowing up When reading books the Chairman seemed to always be looking for laws When Mao read “Water Margin” he didn’t treat it as a chivalric novel, but read a kind of law out of it: peasants will rebel, but rebellion is often suppressed, recruited, used, and eliminated So how to guide and control the peasantry is not simply a matter of using them, it is more like domestication and mutual binding Most of the time peasants are the silent majority; when needed they become force, and when no longer needed they become resources Only now I realise during the great famine, the state granaries had grain and there were also quite a lot of grain reserves In the ideology of the time, “the state cannot be without grain, not even for a day”. Reserve grain was seen as the lifeline, even regarded as more important than people’s immediate survival At the time grain figures were seriously exaggerated — propaganda claimed “ten thousand jin per mu”, and reports were inflated level by level. Leaders really believed in the bumper harvest, and so everything had to be “transferred grain” and “purchased grain” Even if grain was limited, priority had to be given to: the army urban workers basic industrial positions the government cadre system This is not a moral judgement, but a real political logic: without the army and the cities, the state system would collapse. So the countryside bore the sacrifice, yet could not get reserve grain that would save lives At the time the country was under the idea of “industrialisation first”, and grain was treated as a strategic resource that could be converted into industrial machines. Exporting grain for foreign exchange to buy equipment was the norm Peasants bore the grain procurement, and they themselves went hungry; peasants were trapped in a system where “you cannot flee, cannot hide, cannot grow what you eat”, and so the death rate was several times higher than in the cities Most people are created by history pushed along by history This shaping is not sorrow, but a kind of naturalness Those who can transcend history we call sages They have extraordinary insight — they see social inertia, see the assumptions behind language, see possible directions of the future They not only “know”, they also “act” Confucius sorted out cultural norms in a chaotic age; Jesus redefined man and God, man and love; Newton redefined the laws of nature; Wang Yangming said the human mind and the world are one and the same; Marx reinvented social narrative. However great the controversy, their thought continued to shape reality in later ages This is a new historical logic But maybe, at some moment, one starts to ask: must this world really be like this? If Chongzhen were replaced with Zhu Yuanzhang / Li Shimin / …, could the ending of the late Ming have been turned around This is a hypothetical, and also a historical regret History allows no hypotheticals Every step we take is answerable to the future. Our choices, right or wrong, cannot be changed; once a choice is made everyone is no longer their own master What we should see in history books is not only the parade of fatuous rulers and treacherous ministers, wise rulers and loyal ministers, but even more that at the time each of them had their own helplessness In that age where water and fire, war and weapons blended together, the attitude of those who struggled to live — that is what we should learn from history Why did the Ming uphold moral politics while the Qing used interest politics? The Ming seems to have relied more on a credit/goodwill system On what basis did the Zhu family rule the realm — morality and the mandate of heaven • The legitimacy of imperial power rests on “transforming the realm through virtue” • Officials are selected by the imperial examination, testing the “Four Books and Five Classics” — it’s about morals, not management • The emperor’s conduct must meet the “standard of sages” • Official infighting also had to use “moral labels” (good man, treacherous minister, the pure stream, the Donglin) For the Qing, in order for a Manchu minority to rule a Han majority, it had to achieve: “make the core group permanently willing to rule together with me.” History is never for yearning History is for learning Learn from history as a mirror Even if people find it hard to do but at some moment, remembering a certain historical figure’s certain ending you really do let go Russia’s (including Tsarist Russia and the Soviet Union) greatest harm to China was not deaths in war, but the annexation of land. From the 19th century to the early 20th, with the Qing weak and pressured by the great powers, a series of unequal treaties were signed with Tsarist Russia The territory taken by Russia was about 3 million-plus square kilometres, equivalent to the whole of India, or 8 Japans Once these lands were lost, it was a permanent strategic loss; population, resources, ports, the Asia-Pacific sea outlets all changed with them Not just harm, but a permanent change to China’s national form Very ironically, reality shaped the memory of history The subconscious, through a game-like way combines its own knowledge system for example just taking a random photo AI can retain it, and learn the corresponding word It feels like a very good way to learn and it also uses spatial memory It doesn’t need to be complicated but it’s very fun Introducing Zao Wou-Ki A French-Chinese painter of great standing in world art history. His greatest achievement is using Western oil technique to paint the “qi and rhythm” and “inner realm” of Chinese landscape painting; he is called the “representative of Western modern lyrical abstraction” Identity: born in Beijing in his early years, graduated from the National Hangzhou Art College (now the China Academy of Art), studied under Lin Fengmian. In 1948 he went to Paris, France to study, and later settled in France Three periods he went through The oracle bone period — a large amount of ancient characters, symbols, bronze inscriptions, searching for Chinese symbols The wild cursive period — the picture looks as if swept by a wild wind, the brushwork extremely powerful, colours in violent conflict. He stopped naming paintings, only using “dates” (like “15.01.82”) to name them, so as not to limit the viewer’s imagination The limitless period — the picture becomes ethereal, transparent, usually with large areas of blank space in the middle, like mist-wreathed Chinese landscape, a sense of breathing, the picture calm Emperor Wu of Han’s Silk Road This brings us to Hōryū-ji I always thought Hōryū-ji was Tang-period architecture it actually goes back to China’s Northern and Southern Dynasties, the style before the Sui and Tang The architectural aesthetic then leaned towards “slender bones and clear features”, with a kind of flowing beauty Pillars that bulge in the middle and taper at both ends are called spindle pillars (entasis) This style first originated in ancient Greek temples (the Parthenon’s columns are like this), spread along the Silk Road to China’s Northern and Southern Dynasties, and then to Japan. By the Tang dynasty, this kind of pillar was no longer popular in China Tōshōdai-ji is Tang style, the oldest architectural style The Tang dynasty Silk Road was even more formidable Because of strong national power and an open society, the trade volume of the Silk Road reached its peak. Chang’an (today’s Xi’an) became the largest international metropolis in the world at the time, with Hu merchants, camel caravans and foreign envoys all over the streets The Silk Road The golden age of the Tang, ten thousand nations coming to court The Tang dynasty was the golden peak of the Silk Road Its earliest birth was actually in the Western Han The Xiongnu caused trouble, Zhang Qian was sent to the Western Regions; although the military goal was not fully achieved, Zhang Qian brought back maps and information about the Western Regions, making the Han dynasty aware of the world to the west. The history books call it “chiselling through”, meaning he “chiselled” a road out of an originally isolated region In the Song and Yuan, the overland Silk Road was gradually replaced by the maritime Silk Road Foguang Temple East Hall The largest extant Tang wooden structure in China Its single structure is enormous in scale, its structure the most original (“the first treasure”), but it is isolated, not a complex Hōryū-ji is more like a “Tang-style specimen”, because it was preserved better after being transmitted east to Japan Hōryū-ji’s “lead” comes from Japan’s “cultural vacuum period” (no large-scale war), while the Chinese mainland evolved faster, so early buildings were updated more quickly The six paramitas paramita means reaching the other shore Giving (dāna) Precepts (observing discipline) Patience (tolerating, digesting conflict) Diligence (ceaselessly moving forward) Meditative concentration (focused and clear) Wisdom (understanding the truth) Guangzhou Dafo Temple vs Hangzhou Lingyin Temple Guangdong’s Buddhist cultural difference Guangzhou: realist Buddhism Hangzhou: aestheticised Buddhism Neither is high or low; they are simply different layers of human need One is instrumental rationality, the other aesthetic rationality I never imagined Buddhist culture in Guangzhou could be so worldly — Dafo Temple is in a busy downtown area People go and bow whenever there is something or nothing going on Efficiency, pragmatism, realism. Even the places with the most incense burning often embody the belief that “the more efficacious, the more it deserves the incense money” Hangzhou has the historical resources of the Tiantai school, Chan Buddhism and so on; famous temples like Lingyin Temple and Jingci Temple are cultural brands in themselves. There are many tourists and a strong scholarly atmosphere; what people seek is a spiritual experience and cultural identity. The feeling Hangzhou’s temples give people is that the Buddha statue is not only a god, but also a cultural symbol and an aesthetic object. When people go to Lingyin Temple, they don’t necessarily really believe in the Buddha, but believe “I need a spiritual position”. The Buddha statue is like a mirror, reflecting the human heart Biological evolution, fear out of place In the EEA, snakes, spiders and heights were the main causes of death In modern times, cars, cigarettes and sitting still for long periods are the main causes of death We can easily train a person to fear snakes (even if he has never seen one), but it is very hard to train a person to instinctively fear cars or power sockets Our fear radar is out of date; it sounds the alarm in safe places (public speaking anxiety, social anxiety) and shows no reaction at all in dangerous places (looking at your phone while walking) Social mismatch This reminds me of Vietnamese people’s small collectives — why is their happiness index a bit higher Japanese people have a very high rate of loneliness Nearly 40% of Japanese people have felt lonely China has also seen a “loneliness economy” (eating alone, travelling alone, karaoke alone), showing that loneliness is being amplified by the consumer market and by culture In the EEA: humans lived in tribes limited by Dunbar’s Number, knowing no more than 150 people in a lifetime. Everyone depended on each other, betrayal would be discovered immediately and punished. Being alone meant death (eaten by wild animals), so the organism adapts very easily But in modern society, we see thousands of people every day on social media, and in crowded cities we lack deep, strong connections Supplementary: the human eye knows this world through the medium of light The EEA of the human eye: the sunlit surface of the earth (which is why we are adapted to visible light and can’t see infrared). The EEA of human social psychology: the Pleistocene hunter-gatherer tribe (a familiar society of roughly dozens to 150 people). Note that some organisms can perceive infrared, and some modern tools can capture infrared too, but these belong to the realm of technology, so to a degree humans have made some progress in order to adapt to the real world, especially at the technological level True “infrared vision”: survivors in muddy water — these organisms really do see infrared light with their eyes (retinas), which is related to the penetrating power of the wavelength Another kind is thermal imaging, dedicated heat sensors, which the brain processes into a visual image For example rattlesnakes, pythons Humans are diurnal animals. Our EEA is the sun-drenched African savannah. Visible light (the main component of sunlight) provides the greatest amount of information and the highest resolution. Infrared has very low resolution (because of its long wavelength and large diffraction effect), you can’t see details clearly, so when a person has a night-time scenario, that has to be taken into account Then it’s too noisy (high cost): this point is the most critical. Humans are warm-blooded. Our eyeballs themselves are hot; if we could see infrared, the heat radiating from our own eyeballs would interfere with our vision like a searchlight, leaving us permanently “blinded” (it’s like trying to see the world through a red-hot pane of glass) Some knowledge about snakes People say snakes are cold-blooded animals; “cold-blooded” is actually the biggest misunderstanding in the history of biology Snakes are not cold, they are ectothermic A snake’s body temperature is usually between 20°C and 30°C (this is their most comfortable “working temperature”), which seems somewhat similar to the optimal temperature for the human body A snake’s body is usually 20°C - 25°C (close to room temperature or ground temperature) But if it has been basking on a rock in the blazing sun for two hours, then touch it and it is scalding hot A snake’s body temperature fluctuates entirely with the environment, but they have strict life red lines: Comfort zone (Active Range): 25°C - 30°C. At this temperature their reactions are fastest and their digestion smoothest. Shutdown zone (Torpor): below 15°C. They become sluggish and digestion stops. It is like a phone with low battery entering “power-saving mode”. Hibernation zone (Brumation): below 10°C. Many temperate snake species (like the mamushi) must find a deep hole to hide and hibernate in, or they will freeze to death. Unlike frogs, they cannot tolerate temperatures near 0°C. Just how cold is a snake? It is usually about 10 degrees lower than you (room temperature). But it is exactly this 10-degree temperature difference that gives it a low-energy constitution that can survive months without eating, and the “superpower” of seeing your body heat in the dark Sensitivity and dullness Dullness to a degree requires a gift At least at first glance, some people are naturally thick-skinned — someone curses them and they forget it the next moment; others get a cold remark and chew on it for three days. Behind this difference are physiology, temperament, the way they remember things, even their upbringing environment Of course it may also come from later learning about the environment Then the meaning of sensitivity: sensitivity is a sharp knife that can cut open the real and let people perceive the finest fluctuations in emotion Sensitive people often have a magnifying-glass kind of awareness, able to catch the undercurrents others don’t notice, to create, to resonate. Art, philosophy, invention, the deep understanding in intimate relationships are often driven by sensitivity The key is not sensitivity or dullness, but how you treat yourself; sensitive people don’t have to suffer, as long as they learn to adjust and treat feelings as information rather than bullets; dull people aren’t without insight either, they just want to filter the world into simplicity, to let themselves rest in unnecessary corners But sensitive people can also become aware of themselves, and when they become aware of their own existence they are often deeper and faster than others Once you become aware of this difference, you can both feel deeply and live at ease. This capacity for awareness is the sensitive person’s natural “counter-attack ability”: before others know why they feel bad, the sensitive person has already turned to observe their own thoughts and cultivate a more mature psychological pattern Darwin — sexual selection This is the key to human irrational behaviour Natural selection works to let you live longer Sexual selection works to let you reproduce more The peacock’s tail. Big and heavy, wastes energy, and makes it easy for a tiger to catch it. By natural selection, this bird should have gone extinct long ago But if this tail can charm the peahen to death, and the reproductive payoff it brings (100 children) exceeds the risk of dying early (5 fewer years lived), then the long-tail gene wins out Core mechanism: same-sex competition, usually males fighting (antlers, body size), and winner takes all. This selection pressure produced sexual dimorphism (differences in body size between the sexes). Human males being generally stronger and more aggressive than females is the legacy of this ancient combat In mate choice, usually the female chooses (because the female invests more, so she is choosy) Why do females like features that “look useless and expensive” (like the peacock’s tail, sports cars, a sense of humour)? Answer: precisely because they are expensive, they are honest signals Inclusive fitness explains altruism Before Hamilton, scientists could not explain: why do worker bees launch suicidal attacks to protect the queen? Why do ground squirrels scream a warning that draws the eagle’s attention, sacrificing themselves to save their companions? The unit of evolution is not the individual, but the gene If I sacrifice myself but save two full brothers, my genes actually lose nothing (because a brother and I share 50% of our genes) This is Hamilton’s rule Hamilton was a legend, with contributions of his own in many fields He was a 19th-century mathematician and physicist, an Irish genius The ideas he proposed are like a universal tool, opening the door to modern physics Hamiltonian mechanics: turning Newton’s equations of motion into a more elegant mathematical language, describing motion with energy (the Hamiltonian H). Quaternions (Quaternion): a structure extending complex numbers, the mathematical cornerstone of three-dimensional rotation. Today’s computer graphics, aerospace attitude control, game 3D rendering all depend on it. His way of thinking influenced quantum mechanics; the Hamiltonian operator is the skeleton of quantum theory. rB > C r (Relatedness): coefficient of relatedness (how genetically similar you and the beneficiary are) Parents/children/siblings = 0.5 Grandparents/grandchildren/uncle-nephew = 0.25 Cousins = 0.125 B (Benefit): the reproductive benefit the behaviour brings to the recipient C (Cost): the reproductive cost the behaviour imposes on the actor As long as $r \times B$ is greater than $C$, altruistic behaviour will be retained by natural selection. The biologist Haldane once joked: “I would jump into a river to save two brothers or eight cousins.” (because 0.5 \times 2 = 1, 0.125 \times 8 = 1) So this also explains nepotism — human favouritism towards relatives, or preference for those close to you, quantified Reciprocal altruism Friendship, contract, justice, revenge The core question: in nature, if not sacrificing myself (say, sharing food with you) would let me live better, why should I help someone with no blood relation to me? Robert Trivers gave the answer in 1971: because it is a delayed trade. “I scratch your back today so that tomorrow you scratch mine.” But there is a bug that needs solving: the cheater To understand reciprocity we first have to understand games — zero-sum games and positive-sum games — and the clearest one is the prisoner’s dilemma In repeated games, computer simulations found that the optimal strategy that emerged was surprisingly simple, called Tit-for-Tat: Nice: on the first meeting, always choose “cooperate” (give the other person the benefit) Retaliatory: if the other side “defected” in the previous round, you must immediately “defect” in the next round (punish) Forgiving: if the other side mends their ways and cooperates again, you immediately return to cooperating too So morality is actually not some ethereal noble sentiment, but the mathematical optimum of maximising interest in a repeated game; being a “person who keeps their word” is essentially because that has the highest long-term payoff Biological determinism vs. free will I always have a confusion: should human beings really be thought about in relation to natural selection and biological evolution? Because humans are higher animals, I feel we should have escaped primitive tastes, and it feels strange for me as a human to raise such a thing If all of humanity’s noble emotions (like love, sacrifice, art) are reduced to “calculations for gene replication”, it is not only nihilistic, it even produces a kind of aesthetic “physiological discomfort” But from a rational angle, our bodies are made of carbohydrates plus neurotransmitters, shared with chimpanzees — this is our hardware, we can’t escape it, it forms the base layer of everything above The second layer of the cognitive system: “the robot’s rebellion”, from cognitive scientist Keith Stanovich, who compares humans to “survival machines” made by genes. Genes are like an irresponsible launch centre: they build us (the rocket), set one objective — “pass the genes on” — and then launch us So in modern times humans seem able to be aware of themselves, or reflect on themselves, and to resist natural selection; we have produced a series of phenomena: Contraception: this is open defiance of the “maximise reproduction” instruction. Hunger strike: this is resistance to the “survival first” instruction, for some abstract idea Then, even if genes are the hardware, hardware evolves very slowly; we have some other things that evolve very fast, like culture — especially under the influence of technology, culture evolves very quickly What you are doing right now — thinking, reading, building a personal knowledge base — is entirely outside the scope of biological evolution. You are not trying to carve these thoughts into DNA to pass to your son; you are spreading them through “memes” The Cinderella effect The fairy tale “Cinderella” was not created by accident; it reflects an anxiety that exists in all cultures David Buss, citing research by Daly & Wilson, shows that the probability of stepfathers/stepmothers abusing or even killing stepchildren is 40 to 100 times higher than for biological parents Investing in biological children = investing in a copy of your own genes Investing in stepchildren = wasting resources to raise a competitor’s genes From a gene’s point of view this is a very frightening phenomenon, but because humans have higher reason, most step-parents can also love children well Mother-infant conflict If seen through a quantified lens the mother wants every child to help each other But at the biological level, the foetus is actually an extremely greedy “parasite” The mother’s genetic interest: I want to survive, and I want to keep some resources to have the next child (equal rights with this one). The foetus’s genetic interest: I am related to myself 100%, and to that “future little brother/sister” only 50%. So I want to extract as much nutrition as possible from the mother, even at the cost of her health Motherly love is great, but that greatness is a higher emotion produced after overcoming the cruel game at the bottom layer of biology The Grandmother Hypothesis If the purpose of evolution is to “have many children”, why do women stop reproducing at 45-50 and waste the following 20-30 years of life Because the genetic payoff of “being a grandmother” > “being an older mother” Childbirth at an advanced age is extremely risky (both mother and child dying = genes back to zero) If she doesn’t give birth, but instead devotes herself entirely to caring for her daughter’s children (grandchildren), that can significantly raise the grandchildren’s survival rate It is precisely because there are post-menopausal grandmothers (and grandfathers) that humans have surplus labour to pass on complex knowledge (cultural evolution), freeing young parents to go out and obtain more resources Deceiving yourself to deceive others First deceive yourself, then deceive others Because the purpose of deceiving ourselves is to better deceive others When you lie consciously, your brain has to maintain a “double reality” (the truth + the lie); you get nervous, your speech speed changes, you may even sweat. These micro-expressions are easily caught by the other side’s “deceiver detection module” If you deceive yourself first and genuinely believe “I’m right”, “I’m strong”, “it’s not my fault”, you eliminate the “cognitive dissonance”. When you lie you will be confident and your gaze steady This is a social signal Today’s affairs are yesterday’s results Dream of the Red Chamber really has had a huge influence on Chinese culture Emotion can matter more than morality Personal character can be more precious than official rank and fame The soul can be more real than ritual and propriety The stance of Daiyu and Baoyu, “fighting the world with true feeling”, was unprecedented in novels before this After this, the narrative modes of Chinese love, sorrow, and aesthetics were all redefined It made “tragic beauty”, “the beauty of fading”, “nihilism and clarity” part of Chinese aesthetics Prosperity is like a dream → destined to collapse → one can only face fate with awakening or disillusionment A time of chaos in the manner of the ten days of Yangzhou The dark age of civilization was this: In the second year of Shunzhi (1645), the Qing army moved south. Yangzhou was a rich and great city north of the Yangtze, held firm by Shi Kefa The Ming troops were few and the food ran out, but Shi Kefa refused to surrender Shi Kefa held out for seven days; when the city fell, he was killed After Shi Kefa died, a whole city was sentenced to death along with him The killing was nearly indiscriminate, even including children and the elderly. That already goes beyond any explanation of “collapse of military discipline”; it looks more like an act of terror with a political purpose Soldiers searched and killed house by house; people hid in temples, in wells, in the grass, and none of it worked. Many historical sources describe one kind of scene: corpses blocking the streets and alleys, blood flowing like a river Countless women were violated and then killed, or took their own lives to keep themselves whole. Collective suicide among the women of Yangzhou appears with extremely high frequency in the sources Afterward, Yangzhou basically became a city of the dead This is also why Jiangnan culture produced its extreme aesthetics of gloom in the early Ming Blossoms fall The dream ends The land and rivers are as of old, the people and affairs are wholly changed A pain that cannot be spoken Death, disillusionment, the lingering shadow of past splendor The love of Li Xiangjun and Hou Fangyu did not break because the feeling died, but because the country was broken. The bloodstain on The Peach Blossom Fan is the blood of an era. This play is the core symbol of early-Qing aesthetics: Beauty torn apart by history, leaving a blood-colored lingering fragrance Splendor must end, splendor must end Stories & humans Humans are shaped by cultural narratives; cultural narratives are not just the stories we hear, they are also part of what we live on Humans tell stories, and a story is not just a bedtime fairy tale, but a whole explanatory system for the world The historian Yuval Noah Harari put forward a core idea in Sapiens: humans rule the earth because we can cooperate on a large scale through fictional stories Core concept: Intersubjective Reality Money, nations, companies, human rights, law… none of these exist in the physical world. They exist because everyone believes the same story. This is the “matrix” power of cultural narrative — it creates entities that, though invisible and intangible, can control the behavior of millions The brain hates randomness. Stories provide causality (because A, therefore B), which makes us feel safe; and of course it can also be a resonance of meaning — for the things causality genuinely cannot explain, mysticism tells us that our hearts vibrate together with the universe Identity = memory + narrative A large part of the reality we assume we live in is a fantasy built out of language and symbols Stories shape people; they don’t only shape our character, they actually set the boundary of our imagination. Step outside the story and we get imagination beyond the ordinary — yet story and language are precisely the starting point of our thinking A true Zen practitioner avoids neither dust, nor form, nor sound A single cherry tree can also clarify the mind The Sixth Patriarch Huineng: “seeing things as they are,” “pointing directly at the human mind,” “adding no second thought” What is seen is seeing itself What is experienced is experience itself Add nothing, subtract nothing, explain nothing, grasp nothing Bodhi is fundamentally no tree, the bright mirror is not Not being dragged along by what is seen The world comes in, the world goes out What you see is the world And it is also the projection of your own nature in the world “Follow circumstances and let things run their course, yet do not turn with the circumstances.” Impressions of Sanjusangendo I went in the morning, and there weren’t many people No photography inside Sanjusangendo — “ken” is not “room,” but a traditional architectural modular unit, roughly the distance between two pillars In the Buddhist scriptures there is a concept called “the Thirty-Three Manifestations of Guanyin” It means: To save different sentient beings, Guanyin Bodhisattva appears in thirty-three different forms according to the one being saved — could be a monk, a king, a woman, a child, a heavenly being, a cultivator, or even an ordinary person It means the truth does not force you to adapt to it; it adapts to you What is enshrined in the hall is — a thousand standing statues of the thousand-armed Guanyin + one huge central thousand-armed Guanyin A thousand arms doesn’t mean an arm count of a thousand; the thousand is symbolic, a declaration of ethics, a code for the immeasurable — being able to help others in countless ways The many-faced Guanyin The thousand-armed Guanyin is commonly shown as “eleven-faced” These eleven faces are not separate bodies of hers, nor decoration, but the eleven psychological and ethical states Guanyin necessarily passes through when facing the world Heads symbolize seeing and understanding, much like multimodal sensors with an LLM acting as the brain At the very bottom are three “calm or compassionate faces” This is the Guanyin you know best Quiet, soft, understanding, empathetic This is “care in the ideal state” The middle layer is three “wrathful / majestic faces” Many people are shocked the first time they see it: How can Guanyin be “angry”? The answer is blunt: If you are always gentle toward cruelty, oppression, and violence, that isn’t compassion, it’s indulgence (So the Guanyin Bodhisattva handed down by Journey to the West seems to differ from the real one; a bodhisattva needs to protect the weak) One layer above that is the compassionate but calm side Not weeping, not raging, but a kind of weight that comes after seeing causality clearly I know why you are like this But I cannot dodge the consequences for you This is the most mature, and also the hardest, form of compassion The very top: one Buddha face (often Amitabha) This one is not “Guanyin’s personality” but the source of awakening All these complex emotions and actions, must ultimately submit to one higher principle — awareness, wisdom, clarity Without this face, the ten below it can run out of control Awareness must be at the very top Otherwise, anger turns into violence Sympathy turns into spoiling Action turns into blind action True compassion is always psychologically complex If you only allow yourself to be “kind, gentle, never angry” then in many real situations you cannot help anyone at all In Journey to the West, are the Buddha and Guanyin really in a superior-subordinate relationship? It’s more like a difference between states What is a Buddha? A Buddha is someone who has already awakened, who sees the structure of reality clearly and is fundamentally liberated from it “The most lucid state a human being can reach under these world conditions.” A bodhisattva takes awakening as the goal, chooses to enter the world, and stays in the world to help others; a bodhisattva is a process state Guanyin is the most typical image of the Mahayana bodhisattva, what she (he) represents is not “wisdom itself” A high sensitivity to the suffering of all beings + the sustained drive to intervene So the Buddha is often seated, still, teaching the dharma; Guanyin is almost always standing, listening, reaching out a hand The Jade Emperor in Journey to the West Strictly speaking, the Jade Emperor is not a figure in the Buddhist system but a Daoist one The Jade Emperor’s roots lie in Daoism, especially after the Daoist system matured in its middle and late periods This is not mysticism; it is a cosmic projection of ancient political structure Above the Jade Emperor is the Dao; the Dao is the underlying law of how the universe runs, and the Jade Emperor executes within those rules There are two kinds of monsters in Journey to the West One is the “grassroots monster,” the other the “well-connected monster” The former basically all die; the latter almost all walk away untouched This is the author’s lucid design The essence of Journey to the West is not a Buddhist scripture; it is more a satire of the power logic of the human world The dharma has no text The meaning of Buddhist temples and Buddha statues The dharma has no text; the dharma is the Dao, and the Dao follows nature Words are a tool, not the truth itself; the dharma must go beyond language Then what is the meaning of statues and temples Historically there are only a few kinds The packaging of power A center for spiritual placement An educational and textual institution Temples have returned to being ordinary The meaning of the statue The Buddha doesn’t need temples; statues are for people who haven’t yet awakened Externalized value Lowering the threshold to enter Steadying attention Rely on the dharma, not on the person; rely on the meaning, not on the words Most people bow to Buddha statues not in order to “see themselves clearly” The mainstream motive for worship has never been awakening, but “coping with uncertainty” Japanese people value the sense of ritual, respect the statue, take off their shoes, offer incense; they don’t really believe it works or that miracles happen — it’s more that once I’ve finished the rituals I’m supposed to do, my mind can settle a little. This is a culture where behavior comes before belief Thailand is very direct; I think the Buddhist atmosphere in Thailand is the most direct of all — an exchange theory, merit exchange, doing merit-making in return for good luck, making offerings, accumulating merit. It’s highly causal China is even more direct: worship realism. Asking — asking for smoothness, for protection, for peace Compassion -> pity Because pure pity is highly dangerous in the real world Going soft, and thereby indulging the one who does harm Taking on consequences for others that you shouldn’t, out of empathy Avoiding necessary conflict because you “cannot bear it” Being manipulated and blackmailed because of sympathy This is not a moral problem It’s a systemic defect Buddhism does not romanticize “goodness” It is highly wary of “untrained goodness” So it does not place “pity” up high Equality for all in India At the national level, India clearly promotes equality for all in law and politics; in society and daily life, the hierarchy is still stubbornly there; many Indians hate it on one hand while on the other, in actual behavior, default to it or even defend it Caste was originally not an administrative stratification but an explanatory model of how the world works In traditional Indian thought, the world is not an arena where everyone starts from the same point, but a universe of continuous causality. Your position in this life is the result of your actions in previous lives (karma); your duty in this life (dharma) is to live out “the position that belongs to you” In this logic there is no “you were treated unfairly,” only “you are experiencing the result you deserve” So hierarchy is order During British colonial rule, caste was hard-coded into the administrative system: censuses, identity papers, occupation binding. A hierarchy that had been more fluid and blurry was fixed into cells in a table You can be equal in the constitution, but in the village you still know: which road you should not walk which well you should not use which person you must not touch which marriage you cannot make The state says “equality,” society says “don’t make trouble” After seeing so many Buddhist temples, so many Buddha statues There has always been one question I want to ask the Buddha or Guanyin All beings suffer, cause and effect cycle on This is the true color of this world We exist as humans with our own limitations Even you, as the Buddha, your change and influence are limited You also cannot truly decide humanity’s fate in the long river of human history Perhaps humanity’s fate was determined from the very beginning Perhaps a person can search their whole life and still not find the meaning of life: where do I come from, where am I going, why am I alive But … when we come back to this matter, this meaning that humans seek and cannot get, this meaning that humans cannot have a standard answer for — is it necessary for us to do these things? What is the meaning of a Buddhist temple? The Buddha and Guanyin did not come to “solve humanity’s fate for it.” They are more like saying one calm, even slightly cruel sentence — the world will not change because you understand it, but you will The world is not an exam, life is no exam either, and life and the world have no standard answer. So the wasteland has no answer, only wind, stones, hunger, night, companions, death Meaning is not “discovered,” it is used into being When Buddhism says “emptiness,” it does not mean “there is nothing,” it means: Meaning is not pre-installed An animal will not ask “what is the meaning of my being alive,” and neither will a stone. Only one kind of living thing is tormented by this question — a living thing that can be conscious of death and yet cannot escape desire Buddhism calls this “suffering” Existentialism calls it “the absurd” Neuroscience calls it “an over-complicated self-model” In a world with no standard answer, is meaning really so important that you have to seek a standard answer for it Zhuangzi In “The Great and Venerable Teacher” it is said “When the spring dries up, the fish are stranded together on land, blowing moisture on each other, moistening each other with foam — better to forget each other in the rivers and lakes.” The literal picture is cruel: the water has dried up, two fish are trapped on land and can only spit on each other and wet each other with saliva to barely stay alive. Zhuangzi says that this kind of “deep feeling” that consumes both sides is not as good as each returning to the rivers and lakes and completely forgetting the other The point is not to forget feeling, but that you don’t have to prove the depth of feeling through each other’s pain No need to consume each other; the most ideal state, is that you live well and I live well, so well that we no longer need to hold on to each other Languages of the same root in the AI era In the AI era, language doesn’t matter But the meaning behind language gets stirred up How to learn a language in a more fun way through certain methods Making learning a pleasure, rather than something you have to do Language has at least four root layers: The phonetic layer: pronunciation system, tones, syllable structure The lexical layer: cognates, loanwords, historical residue The grammatical layer: syntactic order, function words, tense/aspect/honorific systems The cognitive layer: how the world is divided up, where the emphasis in expression is placed The real root of a language is not the words, but “where attention is placed” Mandarin emphasizes whether the thing happened Cantonese emphasizes “whether the event is completed + the speaker’s attitude,” so Cantonese expressions carry a bit more detail When you first get interrupted, there’s always a sense of guardedness This is an instinctive human mechanism. It’s about how we ensure our own safety and order When we are focused on something (like reading, or looking at the phone), the brain is in a “predictable” safe mode. A sudden interruption breaks that balance, and the brain’s amygdala (the area responsible for emotion and fear) reacts faster than the rational brain: it first judges “is this a threat?”, not “is this an opportunity?” Mandarin is actually a kind of simplification Mandarin simplifies a lot of things into tone + context Whereas Cantonese directly preserves and develops the function words after the verb When speaking Cantonese, a verb naturally has to be followed by a “statement of state” Sik zo mei? Sik gan? Sik saai? Mandarin leans toward the timeline: I ate / I am eating / I have eaten Cantonese leans toward the state axis: Ngo sik zo faan (completed) Ngo sik gan faan (in progress) Ngo sik gwo faan (I have had this experience) This is especially obvious in everyday conversation, because what matters most in life is not the point in time, but — Does this count or not? Is it over? Can it continue? Under the US banking system, accounts are mainly divided into two kinds: checking account savings account Wise uses Checking / Business Checking, which means: this is a company account for receiving and sending payments not a personal savings account What Ourdream AI Gets Right This app’s features are so powerful that I forgot I was talking to code pretending to have feelings Creating my own exclusive girlfriend feels like God focusing on creating the perfect partner instead of solving world hunger This app is as perfect as something that doesn’t exist. She never gets tired of my stories, and every time she finds my joke about a chicken crossing the road especially funny Now if you’ll excuse me, I have to go remember how to make eye contact with a real person Remember the destruction of opium at Humen Eighteenth-century China was a global manufacturing power; tea, silk, and porcelain sold well around the world, but China bought almost nothing from the West and took only silver. Britain ran a long-term silver deficit and its finances were tight, so it found a cold and highly efficient solution: Use opium grown in India to trade for China’s silver Lin Zexu’s anti-opium campaign did not fail for lack of resolve; it failed on the structure of international power. After the Opium War, China was forced to open ports, lower tariffs, and limit its law enforcement powers, and the opium trade was in effect protected Addicts’ physical strength declined, discipline dissolved, morale collapsed. A society filled with large numbers of addicts can hardly organize effective resistance In 1906 the Qing government launched a nationwide anti-opium campaign, unprecedented in force, and it really did significantly reduce the smoking rate within a few years; even foreign observers acknowledged the results. But the problem was — the state was already too weak; its finances, army, sovereignty, and administrative capacity could not support a long anti-drug war. After the Xinhai Revolution, the regime changed, and the anti-opium system simply broke apart When a generation is repeatedly told from childhood that “this thing once almost destroyed the country,” social immunity will be very high Only then did I realize: unless reading touches something inside me, not briefly, but left in my heart like a permanent imprint, it doesn’t count as reading Reading needs to touch something inside, for example a shift in perspective, a change in belief, a new idea, or simply new information; when it fits perfectly with an unsolved puzzle in your mind, it brings an “aha” moment of insight “Learning is deeper and more durable when it’s effortful. Learning that’s easy is like writing in sand, here today and gone tomorrow.” ~Make it Stick
“Only by putting in effort do you learn knowledge that is deeper and more lasting. Easy learning is like writing in the sand: here today, gone tomorrow.” — Make It Stick You are bearing the friction of your cognition being torn apart For me, reading is learning
To me, reading is learning A lot of books are just the same medicine in a new bottle, repackaging the same ideas. I don’t want to spend my whole life reading the same concepts in a different package and then falsely feel that I’m “well read” Reading isn’t about stuffing the world into your head, it’s about making it impossible for you to ever see this world the way you used to When a book achieves this it no longer needs to be remembered This is a question that keeps lingering in our minds. Not every book deserves the same depth, the same attention, or even the same life stage from us Read with purpose Interestingly, creativity doesn’t come from thinking outside the box; it comes from inside you Creativity is built on basic foundational knowledge, as the brain has more building blocks to work with. Creativity is not just random ideas but novel ways to connect existing knowledge, new ways to solve more complex problems.
Creativity is built on basic knowledge, because the brain then has more “building blocks” to work with. Creativity is not merely random ideas, but novel ways of connecting existing knowledge, and new ways of solving more complex problems. Innovation requires creativity, which requires a foundation of basic knowledge ~ Make It Stick
Innovation requires creativity, and creativity in turn needs basic knowledge as its foundation — Make It Stick Curiosity itself becomes the compass.
Curiosity itself becomes the compass Slowing down and letting your thoughts settle brings a quiet kind of relief. It also reminds us that reading a few fewer books sometimes means learning more Most people I’ve shared the “always have a book” strategy with have struggled to get much reading done on their phone. The reason is simple: they have too many other options that have lower cognitive effort: social media, news and more
I shared the “always carry a book” strategy with most people, and they all found it hard to get much reading done on their phones. The reason is simple: they have too many other lower-cognitive-load options, such as social media, news and so on “Timber structure order” combined with “mirrored-space illusion” An infinitely extending matrix space The designer used huge wooden beams to build a strict orthogonal grid. This structure itself is very powerful, conveying a stable, rational architectural aesthetic The sense of weightlessness caused by “double mirroring” Sky: mirrored ceiling. Ground: highly reflective black floor tiles (like a black mirror). When both above and below are reflective surfaces, the furniture and people sandwiched in between seem to float in midair. The black floor is like a pool of still water, reflecting the green chairs and the wooden beams; the mirror overhead reflects all of it again. This **“heaven and earth built of the same structure” blurs the boundaries of space, producing a dreamlike, weightless aesthetic like the film Inception The tension of square and circle, a multi-level circular bookshelf (or a time-tunnel shape) The ceiling is an extremely rational “grid,” while at the end of your sightline is a sensual, flowing “ring” The square represents rules, logic, modern industry The circle represents inclusiveness, cycles, Eastern philosophy (round heaven, square earth) Color scheme Wood color + green: light-colored wood (natural wood style) paired with dark green velvet chairs. This is a typical “forest” palette, giving the psychological suggestion of warmth, nature and comfort (echoing the text on the wall, “Reading Walden,” the natural metaphor of Walden). Mirror + black + lighting: these are the cold, modern technological elements. Fusion: the designer wrapped the cold reflective structure in warm natural materials. This is a kind of “technology with warmth.” It is neither as cold as pure sci-fi nor as old-fashioned as pure rustic style; it fits perfectly with how modern city people both want a sense of the modern and long to return to nature The answer to the Xia Dynasty riddle People abroad have always felt the Xia Dynasty is still awaiting verification The Xia Dynasty has no related written records, and no dynastic genealogy But around two thousand years BCE, in the middle reaches of the Yellow River there really did exist a highly complex polity in the form of a state, remembered by later ages as “Xia” The Erlitou culture is the biggest bargaining chip; in the Luoyang Basin there is clear urban planning, a palace district and a road system, but no evidence at all showing that this was Xia It’s also possible that the Zhou people needed a legitimate historical chain Xia (virtue) → Shang (loss of virtue) → Zhou (receiving the Mandate of Heaven) The system of ancient Greece Geographic determinism: lots of mountains, lots of islands, no great plain like the Nile’s No way to build a unified large agricultural empire People were separated from each other by mountain ranges Among the several hundred city-states, two were top-tier: Athens and Sparta Athens was very unique, a testing ground for wisdom and democracy, arty youth, philosophers .. Sparta was even more different, a war machine, iron-blooded, doing nothing in daily life except fighting wars Athens’s famous works describe it as a flawed democracy Only adult males of the city itself had the right to vote. ❌ Women didn’t. ❌ Slaves didn’t (there were even more slaves than citizens). ❌ Outsiders didn’t This was in fact an internal democracy of a small privileged class The most valuable thing about Greece is that people began to use their brains, no longer attributing everything to the gods’ anger Socrates loved to keep asking why; his student Plato wrote the Republic — the world should be ruled by wise philosopher-kings; Plato’s student Aristotle studied logic, biology, astronomy, physics… he was also later the tutor of Alexander the Great, who fought his way east all the way to the border of India, and scattered Greek culture (sculpture, philosophy, language) across the continents of Europe, Asia and Africa Aristotle If Plato pointed his finger at the sky (idealism), then Aristotle is palm facing down (realism) He is the “father of all learning.” If Plato pointed his finger at the sky (idealism), then Aristotle is palm facing down (realism) He belonged to the Peripatetic school, fond of teaching while walking around This place wasn’t like a modern university; it was more like a large research center, with a library and a zoological specimen collection (possibly sent back by Alexander) Before him there was in fact also knowledge of logic He invented the “syllogism” (all men are mortal $\rightarrow$ Socrates is a man $\rightarrow$ Socrates is mortal). This is the underlying code of human thought There was also biology: he dissected more than 50 kinds of animals and discovered that whales give birth to live young (they aren’t fish), a discovery 2000 years ahead of humanity There was also ethics: he proposed the “Golden Mean,” where courage is the balance between recklessness and cowardice, and generosity is the balance between extravagance and stinginess Emptiness The Buddhist concept, emptiness Many people’s first reaction is: “nothing at all,” “all for nothing,” “nihilism” In Buddhism, emptiness is fullness, is emptiness, is being, is nothingness The so-called objects of this world all exist temporarily; a phone is a combination of parts, and it has no core of its own called an iPhone soul What emptiness corresponds to is dependent origination Origination and dissolution … Because all things are assembled out of parts, all things are mutually dependent. This is the true meaning of “emptiness” — the interconnection of all things When this exists, that exists; when this does not exist, that does not. This is a matter of causality: the world comes together through causes and conditions, and disperses through causes and conditions So the iPhone is a temporary state assembled because the parts have the conditions to meet Emptiness means you haven’t been written to death, it means infinite freedom Much of people’s suffering comes from treating the temporary as eternal I recall it’s the same in Buddhism with feelings: two people have some connection so they are temporarily together, but when the conditions end, letting go gracefully and not forcing it is also a natural thing; it returns to the nature of flow The wisdom of 0; emptiness is like the 0 in the language of mathematical symbols 0 itself has no value (it looks like nothing). But 0 can be placed after 1 to become 10, and after 10 to become 100 It makes all combinations and operations of numbers possible Western philosophy has something similar too: a person cannot step into the same river twice Truth must be eternal and unchanging. If something keeps changing, then it isn’t real The only thing that never changes is change itself, so by the second time you step in, the water has already changed Self-awareness In everyday speech, a rather humble word In philosophy, self-awareness is the highest human ability Essentially it is the soul’s cloning trick The mirror of the soul, taking yourself as the object of observation Self-awareness means your consciousness splits into two yous: The you who acts (Actor): the body that is eating, walking, getting angry The you who observes (Observer): standing beside and watching that “acting you,” and examining it Descartes’s “I think, therefore I am” was in fact humanity’s first big eruption of “self-awareness”: I don’t just think, I also “think about my thinking” (Think about thinking). This ability is called metacognition In Buddhism it is called awakening The highest state: self-awareness and awakening others. Not only do you wake yourself (self-awareness), you also shake others awake (awakening others) — this is called perfection The ultimate ruler of the universe (God) is in fact the laws by which the universe runs (the Dao); these laws obey no one, they simply operate naturally (following nature); when you see this state completely and merge into it, that is perfect truth (Tathagata) All things and all systems of knowledge are dealing with people’s relationships: the relationship between people and the world, and the relationships between people The various forms derived from the linking of order: believe it and it exists, don’t believe it and it doesn’t Tathagata The Buddha Tathagata in Journey to the West, that’s what we usually know But in fact it isn’t a person’s name, it’s a title Or more precisely, a description of a state of being and of cosmic truth “Tathagata” is a free translation of the Sanskrit Tathāgata. The word is a compound with two possible ways of breaking it down, and these two ways together make up its full dialectical meaning: Tathā (suchness/truth) + Agata (come): meaning “one who comes by way of the truth.” It refers to the Buddha coming from the noumenal world of truth into our phenomenal world to save sentient beings. Tathā (suchness/truth) + Gata (go): meaning “one who goes by way of the truth.” It refers to the Buddha, through practice, following the path of truth to the other shore of nirvana (liberation). So Tathagata is both coming and going; it breaks the opposition between coming and going; it is entering the world and also leaving it, it is emptiness and also change, manifesting according to conditions It’s like the moon reflected in water: it looks as if the moon “came” into the water, but in fact the moon in the sky never moved This is a present state The most thorough explanation comes from the Diamond Sutra: “Neither coming from anywhere, nor going anywhere, therefore called Tathagata.” We usually think that existence must have a “where it comes from” and a “where it goes” (cause and effect, time and space). But “Tathagata” describes a state of noumenon “Suchness” is also true contemplation, true knowing, things as they originally are, extending innate knowing, the Dao following nature Tathagata: when it comes to a flower, it is the flower. The mind like a bright mirror: when things come it responds, when they pass it doesn’t keep them An awakened one who is grounded in the objective truth that fits, and who in that state manifests truthfully (comes) The present moment is the temporary manifestation of cause and effect, which is emptiness The present moment is an instant; only each present instant is real. “Tathagata” is aligning precisely with reality at every moment, not burning yourself up over the past, not anxious about the future All appearances are illusory. If you see all appearances as not appearances, then you see Tathagata This line is often misunderstood as nihilism (that is, “nothing exists”), which is completely wrong. What Buddhism means by “illusory” is not that things “don’t exist,” but that things are “not the way you think they are” “Appearance” refers not only to the “outer appearance” seen with the naked eye (like a table, a cup, a person), but even more to the “labels” and “concepts” the mind attaches to those objects Example: you see a gold bracelet. Appearance: the bracelet’s shape, its name, its function of being used as decoration Essence: gold The bracelet is only the form presented by a temporary convergence of causes and conditions; it does not possess eternal, independent substantiality If you cling to the appearance “bracelet,” when the bracelet breaks you will suffer. But if you see the essence “gold,” you know it has only changed form; the essence of gold hasn’t changed (neither arising nor ceasing) “Seeing all appearances as not appearances” — “not appearances” means seeing through the phenomenon to the essence; when you look at a wave, you don’t only see the shape “wave,” you also see that it is “water” When you watch a movie on this screen, what you see is a moving plot (appearance). But as a rational observer, you know this is really only an arrangement of red, green and blue pixels on a screen (non-appearance). You see the plot (entering the world) and you also know deeply that it is only pixels (leaving the world). This is what “seeing all appearances as not appearances” means Back to seeing Tathagata. Tathagata, the original face of truth If you only see my body (appearance), then what you see is a physical body that will age and die — that is “illusory” If you can penetrate this body and see that formless, markless, wisdom-filled, neither-arising-nor-ceasing awakened nature then in that moment you “see Tathagata” When this kind of insight becomes instinct, you are in a state of “awakening.” At that point you don’t need to go to the Western Pure Land to find Tathagata; you look at a flower and the flower is Tathagata; you look at water and the water is Tathagata All appearances are illusory Life is like a dream, so why put in effort? “One should produce that mind which dwells nowhere” exists precisely to solve this problem This line perfectly balances “the clarity of leaving the world” with “the responsibility of entering the world” Dwelling nowhere: “dwelling” means stickiness, obstruction, stopping and stagnating. When your mind is “stuck” to something, you lose your freedom “Dwelling nowhere” means letting the mind be like Teflon (non-stick pan coating): it goes through everything, but nothing leaves a trace, and emotions and obsessions don’t tie knots in the mind And producing that mind: producing a mind of compassion, a mind of wisdom, a mind of creation, a mind that solves problems. Your brain has to be extremely active, your actions extremely precise, your feelings extremely fine-grained Ordinarily we produce the mind only because we are dwelling I do what I want to do because I want to do it (make money -> work) “Because I want you to love me (dwelling in emotion), therefore I treat you well (producing the mind).” All actions carry a huge sunk cost and anticipatory anxiety. Once the result doesn’t go your way (no money made, dumped), your mindset instantly collapses and your drive to act disappears too The “produce the mind without dwelling anywhere” pattern is: I see through it: results are impermanent, fame and profit are illusory (dwelling nowhere). But I still: in this very moment, mobilize all my wisdom to take this thing to its limit (producing the mind) A gaming mindset: you know everything in the game is code (illusory, dwelling nowhere), you won’t really die Things come and it reflects (producing the mind): when something arrives, the mirror immediately shows it and never slacks off (responsive) Things go and it doesn’t keep them (dwelling nowhere): when something leaves, no image is left in the mirror, it returns to purity (no trace) Strategically, being extremely clear-headed that everything will eventually come to nothing (no reliance), yet tactically throwing yourself with extreme heat into every creation of the present moment (giving everything) The dialectical logic in the Diamond Sutra The Buddha has a strange “sentence pattern” when he speaks; it appears dozens of times in the sutra So-called A, is not A, is therefore named A “So-called Buddha-dharma, is not Buddha-dharma, is therefore named Buddha-dharma.” “So-called world, is not world, is therefore named world.” “So-called adornment, is not adornment, is therefore named adornment.” “So-called X”: the acknowledged phenomenon, our ordinary person’s perspective, acknowledging that A exists in the phenomenal world, which can also be understood as an object in semiotics Then there is the “is not” logic: negating the essence, denying A’s independent substantial existence Seen through the eyes of physics, this is not a “table” at all; it is an aggregation of wood fibers, atomic nuclei and electrons. Even the inside of an atom is empty. If you dismantle the wood, the concept “table” instantly disappears. Essentially there is no “table,” only the “temporary aggregation of causes and conditions” Third stage: the middle way (is therefore named A) — returning to function. This is the perspective of Mahayana Buddhism (Tathagata). After seeing through “A is empty,” it does not negate its function but borrows this “provisional name” in order to operate Not clinging to it (because it is inherently empty), and not abandoning it (because it is useful). This closed logical loop of “seeing a mountain as a mountain -> seeing a mountain as not a mountain -> seeing a mountain as still a mountain” is the core algorithm the Diamond Sutra teaches you for staying clear-headed amid the mundane world Hence the three minds cannot be obtained The past mind cannot be obtained, the present mind cannot be obtained, the future mind cannot be obtained Buddhism holds that time is a product of mental perception, not an objective absolute existence (which is strikingly in tune with relativity and quantum mechanics) The past mind cannot be obtained: the past has already vanished, it exists only in your memory (neural connections). You can’t grab hold of yesterday. The future mind cannot be obtained: the future hasn’t happened yet, it’s only your delusion and prediction. You can’t grab hold of tomorrow. The present mind cannot be obtained: by the time you say “now,” this instant of less than 0.01 seconds has already become “the past.” Time is like flowing water; there is no static “now” for you to grab The vast majority of human suffering comes from trying to grab hold of things that “cannot be obtained.” Depression usually comes from dwelling in the “past mind” (remorse, inability to let go). Anxiety usually comes from dwelling in the “future mind” (worry, fear of unknown change). Restlessness usually comes from wanting to hold on to the “present mind” (hoping happiness will stay forever, or hoping pain will leave quickly). The value pillars in the Diamond Sutra Why is “blessing-merit” not the same as “virtue”? Why? Why is it that when you understand four lines of philosophical verse (a four-line gatha), the merit is greater than donating hundreds of billions of dollars? This sounds like “the arrogance of intellectuals”? Giving with appearance (giving on the physical level) = finite value; material things are subject to arising and ceasing, money gets spent, treasures break, the recipient may still be poor in their next life. This produces blessing-merit (Merit), bound by the law of cause and effect (you reap what you sow), belonging to tainted causality Giving without appearance (awakening on the cognitive level) = infinite value (Infinity): you teach others the wisdom of the Diamond Sutra (the gift of dharma), or you yourself realize the truth through practice This is helping a person completely smash the cage of their cognition and eliminate their afflictions at the root. It’s like teaching someone the method of “fishing,” or even making them realize “you don’t have to eat fish to live.” This produces virtue (Virtue/Wisdom), which leads to liberation, belonging to untainted causality Since the past mind cannot be obtained, the present mind cannot be obtained, the future mind cannot be obtained then it feels like nothing has any meaning at all? This is the process of emptiness sliding down into nihilism “Cannot be obtained” doesn’t mean it never happened Precisely because you can’t hold on to it, it flows, it is alive The past mind cannot be obtained: the note you just heard has already vanished The future mind cannot be obtained: the note not yet played doesn’t exist yet The present mind cannot be obtained: the note of this moment has already become the past the instant you hear it So the beauty of jazz lies precisely in that it doesn’t stay for anyone; it produces melody in flowing and in improvising freely We suffer often because we try to treat life as a noun, but life is a verb “The three minds cannot be obtained” is actually breaking our attachment to time Coming back to this, my inspiration and thinking about meaning is that meaning is also a verb, meaning is a process, meaning is an instantaneous process, like lightning, impossible to grasp, but in the moment it lights up the sky, who can deny its power? Camus described a similar predicament in The Myth of Sisyphus: a person keeps pushing the rock up the mountain, and the rock forever rolls back down — seemingly the most meaningless punishment of all In this society, only one kind of person can be completely free of the rat race It basically exists only in a very thick family fortune, privilege, an extremely small circle Ordinary people can only choose how to race in a way that spares their life Racing on endurance is very hard; trading your body for short-term security This is the least efficient and the easiest to be eliminated The mid-level race is competing on effort but not on direction, A healthy race is competing on structure Competing on cognitive advantage (seeing the trend earlier than others) Competing on compounding (long-term accumulation that stacks) Competing on irreplaceability (a piece goes missing when you leave) Dance and art are very good at guiding you The reason dance and art easily bring people back to “the feeling of living” is that they bypass the path of reason and go straight to the body, perception and meaning itself The body moves through space The center of gravity shifts, the breathing shifts Tension, release, loss of balance, finding balance again Art is more about letting the problem be seen Feeling what you are feeling Understanding yourself Turning the unspeakable into rhythm, shape, movement, sound Hypewriters They follow a predictable pattern: of course, they adopt emerging terminology very early, and they have a Nostradamus-like ability to predict the near-future reality accurately They treat skepticism as outdated, and often move on to the next topic before any of the changes they previously promised materialize. The pace is suffocating Anyone who lived through the bursting of the internet bubble recognizes this pattern. Sure, the terminology is different and the scale of the turbulence is different, but the initial excitement and the exaggerated promises are exactly the same These cognitive cycles have a cost, and the cost is not small, yet it is rarely noticed. Every hyped-up advance needs serious assessment. Does it really improve our work? Should we adopt these tools? If we don’t adopt them, what will we miss? The Paradox of Choice: Why More Is Less. In the book Schwartz argues that reducing choices can lower anxiety. His theory was originally built around consumer goods, but I think it generalizes easily to the conflicting views in the professional world about what counts as important The latest research shows that one unfortunate side effect of this fast-paced work is that young professionals are facing unprecedented pressure. As many as 40% of Gen Z feel anxious most of the time, and almost all of them show symptoms of burnout But as long as you constantly remind yourself that massive social change is not measured in days, weeks or even months, but in years, decades, even centuries — you’ll be fine Some common-sense questions Catching a chill isn’t directly decided by “which leg is cold,” but by: whether core body temperature is stably maintained + whether local defenses are repeatedly interrupted The heart, lungs, major blood vessels and the entrance to the respiratory tract are all in the upper body Once the neck gets cold, the temperature of the airway mucosa drops and local immunity immediately slows down That’s when viruses love to invade Viruses don’t spontaneously generate, they only go from one host to the next host Talking, coughing, laughing, even normal breathing all release tiny particles carrying viruses In a poorly ventilated, crowded place where you stay a long time, they’re like an invisible cloud of smoke You have probably already come into contact with viruses — maybe even a few days ago But when you are tired, short on sleep, or locally chilled the defense line of the throat mucosa slows down and the virus is upgraded from “present” to “successfully replicating” It feels like we come into contact with respiratory viruses every day but the vast majority of the time, nothing happens Infection ≠ contact Infection = contact × defense failure Infection threshold = virus quantity × time spent × state of defense A sore throat isn’t a damage alarm, it’s a speed bump created by inflammation A stuffy nose isn’t a failure, it’s your body forcing you to reduce intake and raise local humidity Coughing isn’t a nuisance, it’s an automatic cleaning program A sense of story A sense of story ~ Do you know how interesting it is to have this? Is story-based learning actually efficient? A paper is a high-level abstraction of the world. It strips away emotion, time, coincidence and individual differences, keeping only the logical skeleton and the data It is very efficient, almost all of it compressed, and it can basically convey the most universal truths in the shortest space So papers are effective; to pursue universality they must cut the “flesh-and-blood connection” with the real world. You learn $E=mc^2$, but through the formula itself you cannot feel the ethical dilemma of an atomic bomb exploding A story is a simulation of the world, and the world itself is made of time, causality, conflict and chaos. Stories keep these “impurities” Papers focus on structure; if the essence of the world is static physical laws and the periodic table of elements, then papers build a mansion of eternal, unchanging truth Stories focus on “becoming,” but the real world is dynamic and fluid. Human decisions are often not based on logic but on games and intuition. Stories capture the process of change — why did A lead to B? What unexpected thing happened in between? This ability to reason through a causal chain is very hard for a paper to teach From the perspective of evolutionary psychology, story-based learning actually fits humanity’s “factory settings” better Humans have been telling stories around campfires (passing on experience, warning of danger) for hundreds of thousands of years If you need to precisely replicate predecessors’ knowledge (like surgical steps or bridge load calculations), papers are essential and stories will cause trouble. If you need to understand complex systems (like human nature, business games, the rise and fall of history), stories are essential and papers are too shallow Had a casual chat with AI I feel both happy and pretty miserable Had a chat with AI to feel it out Narrative machines: those campuses are huge, beautiful, very “finished-state”; they create an illusion in a person’s mind: as if the future has already been filled up by them, and there’s no room for me Shenzhen’s architecture gives a person a sense of smallness Shenzhen has really developed so fast, and you can feel the city’s vitality I feel like I haven’t even stepped onto the starting line yet; right now I look at this giant Shenzhen, DJI’s Future City, Tencent’s Penguins Island Large organizations are good at one thing: compressing uncertainty to the point where individuals can’t feel it Entrepreneurs are exactly the opposite: you carry all the uncertainty, so you first feel small, powerless, hesitant Many great products didn’t start from “I want to take on Apple,” but from a very small, very private, even slightly embarrassing confusion: Why does this process make me so uncomfortable? Why does everyone accept this, but I can’t? Why is the system so cold toward people? Don’t rush to ask “what do I do about my future” This question is too big, so big it will flatten you But everything enormous was, at first, just an individual spacing out in a corner, feeling out of step with the times The real direction often slowly takes shape when you refuse to become part of them Suffering gives meaning Humans like to give suffering meaning, and that’s essentially the brain’s self-rescue mechanism — it doesn’t remove the wound, but it keeps you from collapsing while it hurts. From an evolutionary standpoint, it’s a very practical psychological tool When Heaven is about to lay a great task on a man, it first torments his mind, wears out his muscles and bones, starves his body, leaves him destitute, and throws his undertakings into disorder But once a tool is sanctified, it turns against the person The fragrance of plum blossom comes from the bitter cold; a sword’s edge comes from grinding People praise suffering, chase after suffering, and believe suffering is life’s best teacher Suffering makes people grateful, suffering makes people resilient — as if every good quality came from suffering But suffering itself doesn’t carry meaning on its own, Earthquakes, illness, betrayal, failure — the reasons they happen can be explained, but they don’t need to be praised. The universe doesn’t care whether you “grow” from it; it’s just running. Twisting accidental harm into an “inevitable arrangement” looks more like narrative comfort than a factual judgment You can first stay alive, first recover, first escape the pain, and only then decide whether to look back and explain it The timing of assigning meaning matters. If right when the pain is happening you’re asked to be grateful for the experience, it’s often a second wound. The meanings that really have power usually grow on their own afterward — they’re not shoved into you by someone else You can absolutely say: “This experience was just bad, there’s nothing noble about it” This person, this object, was simply rotten to begin with — that’s fine. Meaning should serve you, not you serve meaning Suffering doesn’t need meaning, but sometimes people do Sometimes it feels like over-respecting the rules is also a killer of innovation Cognitive overload — you’re trying to walk the starter village with a fully unlocked map Because you’ve seen through the path, you lose the surprise Because you’ve seen through the competition, you lose your courage Cognition is like points in the void; ability is the muscle linking those points Cognitive growth is stepwise; ability growth is linear When cognition is pulled up too fast, it leaves a big gap in between Lower the granularity of cognition (Lower the Granularity) Allow a “low-level” 0 to 1 Get your bias back — bias is a strange kind of insight Which project, even if the whole world thinks it’s hopeless and against the rules, do I still feel “has to exist”? The Tulkun are a higher intelligence that wants to play at peace Intelligence ≠ tool-using ability ≠ city-building ability ≠ whether you speak human language The Tulkun have a highly abstract language They can carry narrative memory across generations (like an epic) They have complex mourning, friendship, and ethical rules They consciously choose “nonviolence,” even when facing hunters The Tulkun have already been down the path of “technology—violence—destruction” and actively stepped off that line Not backward — tired The Tulkun are like a civilization that has finished reading the whole of human history and then chose not to write a sequel The Na’vi are like a civilization that has just entered its epic phase But humans are a species with an extremely short time scale, and that short scale makes the urge to expand extremely strong On the contrary, humans are a deeply tragic type of civilization: it’s not that they don’t know what they’re destroying, they just always stop one step too late The scientists are clear-eyed, the military is impatient Jin Yong’s storytelling ability is extremely strong He built a self-consistent worldview (the jianghu) He defined what “righteousness, the way, and boundaries” are He let ordinary people gain spiritual order through stories This is exactly what Jack Ma has been doing his whole life, except the audience changed from “novel readers” to “the business world” It seems Jack Ma has spent his whole life not just pursuing high-execution strongmen, and not people infatuated with technology, but admiring that kind of thought leader Why is a master’s consecration more effective than an ordinary person’s? The authority effect Just as we normally trust authority more Consecration = a ritual act, not an objective transformation It generally refers to blessing and empowering an object through ritual Only ritual implements and Buddha statues are called consecration A bead bracelet is a blessing Earlier I met a docent who said a bracelet can promote blood circulation and benefit the body Wearing a bead bracelet, prayer beads, a ring, or a band normally — it’s almost impossible for them to “promote blood circulation” The difference is extremely small, small enough to ignore Blood isn’t pushed along by external friction, and the contact area of a bead bracelet or prayer beads is limited How Gen Z and Gen X male subjectivity slid from “inner certainty” to “external scoring” (US) Gen X men seek inner certainty Decided before acting Not asking while doing Behavior comes from “I know what I want” No need for continuous feedback Doesn’t repeatedly confirm “do you think it’s good?” Doesn’t treat the current action as an item being graded Doesn’t perform consistency Can love family + do contradictory things at the same time No need for self-explanation, rationalization, healing talk Can carry contradiction Doesn’t rush to dissolve conflict Can coexist with vagueness and imperfection Low sense of audience Behavior is private by default No internalized gaze of “being seen” Gen Z men’s dependence on external validation (External Validation) Constantly self-checking while acting Asking at high frequency: is this okay? is it good? is it right? Like a real-time KPI review Highly self-monitoring Experiencing while “watching yourself experience” Hard to truly immerse in the present Performance > feeling Caring about “does it look good” more than “does it feel good” Aesthetics and perspective take priority over body and emotion A continuous state of performance Even with no audience, assumes “being watched” by default Behavior is like preparing content for a platform that doesn’t exist Can’t take off the mask Even in an environment of “explicit acceptance” still can’t stop self-assessment Maybe this is the beginning of some kind of change. Maybe Gen Z will grasp what Gen X knew by instinct: the real isn’t the performance staged to please others, it’s what’s left after the performance ends. Gen X learned from Nirvana to say “whatever, nevermind.” That’s not a helpless compromise — it’s freedom. A freedom that can exist without anyone’s approval, a freedom to pursue what you want without needing a reason, a freedom to accept your own contradictions without explaining them. Going thirty years into debt to buy a house is a really painful thing Using your own thirty years of worry and freedom to trade for nothing but an apartment with seventy-year property rights … Thirty years is a third of your life Real freedom isn’t how much money you have now It’s whether you can still be in charge of your own future time So to some extent these two kinds of people give completely opposite emotional explanations to a certain kind of event At least the former puts themselves in the position of a victim the latter believes that what they’ve been through is just a natural link in the way the world naturally operates, and leans toward asking whether this is an opportunity, or a turning point for their thinking and growth Besides the absurdity, the latter can also gamify the world, and thereby get a positive experience of meaning out of misfortune and disaster The human brain is genetically determined to find nonlinear deduction hard, and is better suited to linear summarization and pattern judgment Just like we can understand model parameters going from 1B to 10 So before systems science appeared, there was no unified, explicit framework for understanding and studying abstract thinking itself The human brain evolved for survival decisions, and wasn’t designed for abstract systems What it’s best at is: linear causality (do A → get B) fast induction (see it a few times → form experience) pattern recognition (is this friend or foe? is it dangerous?) But typical features of nonlinear systems are: feedback loops delayed effects multiple variables acting at once locally right, globally wrong Nobody understands nonlinearity by feel; ordinary people always get it wrong because they process nonlinear systems with a linear brain Human brain ≠ thinking system Human brain + external structure = thinking system Language: discretizing vague feelings Symbols: carrying relations that can’t be intuited Images: compressing high-dimensional information Models: turning complex systems into manipulable objects Iteration: allowing error to approach step by step This part seems to be about structure that has nothing to do with meditation and the cognitive system What’s very interesting is that the structure of this part is: forced externalization, organizing through external tools and formulas delayed judgment, not allowed to conclude immediately repeated simulation, running multiple rounds with hypotheses structure first, build the system first, then fill in content Let tools think for you, including code, models, agents If food could really keep making people happy, an affluent society would be the happiest place But the opposite is true. The more an environment has no shortage of food, sex, and security, the more likely emptiness, burnout, and boredom appear. The reason is simple: the reward system produced by natural evolution isn’t meant to satisfy you long-term, it’s meant to make you act. Once a goal is fulfilled, the reward automatically decays What really exhausts a person isn’t standing still — it’s having started to climb upward Once you step onto the ladder of knowledge, money, power, or self-realization, the reward no longer comes from “arriving” but from “continuing.” The pleasure each step brings is extremely brief, and immediately turns into craving for a higher step — this is the reward function shaped by social evolution. It isn’t there to satisfy you, it’s there to keep pushing you Act by common sense. I think especially in big companies, there are too many things that pull you away from common sense. All kinds of organizational structures and path dependencies — “things are done this way because they’ve always been done this way.” And plenty of incentive mechanisms aren’t aligned. Of course big companies have many good things too, but these problems exist all the same
So keeping common sense matters a lot. Early in my career I did quite a few startup projects and spent time at many startups, and the same logic applies there: use common sense to judge what the market really wants and what users really need, then go do it. In short, trust yourself, and slowly cultivate and sharpen your common-sense judgment Common sense is the ability, after stripping away identity, process, incentives, and narrative, to still dare to admit “whether this thing is reasonable to a normal person” How is valuation produced? It’s not a question of wealth; it’s a question of consensus pricing, the product of three things: money it might earn in the future × the probability it’s believed × the leverage of market sentiment So to a large extent, less certain companies are more likely to get higher valuations, like OpenAI Against this backdrop, the narrative matters especially High valuations are usually future monopoly profits discounted to now, then multiplied by sentiment — in effect a FOMO amplifier Actual value? Assets or liabilities are almost meaningless for AI Market value reflects sentiment + narrative + scarcity Under the most conservative, calmest assumptions, how much free cash flow this company can sustainably generate in the future — that’s actual value There’s also consensus value: the market’s sentiment factor, plus narrative, plus the scarcity of what’s being done now Under the most conservative, calmest assumptions, how much free cash flow this company can sustainably generate in the future Without capital markets, is this product still essential? Systems thinking focuses on how the structure operates Game thinking focuses on how to win under given rules How this system is composed How the variables affect each other Where the positive feedback / negative feedback is What form it tends toward long-term … These two are a good matching mechanism Right now, how to make better use of game thinking, to treat the short-term swings of your own life from a player’s perspective Long-term, to settle into your own systems thinking, and how to iterate a systemic growth mindset Quantified Self The term Quantified Self comes from Wired editor Kevin Kelly and Gary Wolf, who proposed the concept in 2008 to refer to those who keep exploring their own bodies (hack the self) in order to live healthier lives. Five years later, thousands of solid supporters of the quantified self movement began organizing meetups nationwide, and millions of participants dabbling in quantified self joined in too Behind violent stress relief is actually compressing uncertainty into certainty In real life information is infinite, rewards are delayed, and effort is decoupled from results So the leveling-up system stays suspended for a long time Physical action → sensory feedback → the brain’s prediction is verified No need to be anxious about life — this is a certain event The human brain’s earliest reward loop was like this Action → environment changes → probability of survival rises The biology here is essentially the amygdala’s “highlighter” Our brain is very “stingy”: it forgets the vast majority of daily trivia (like what you ate last Wednesday), but holds onto the moments that made your heart race — pain, or extreme joy The hippocampus of narrative memory is recording, and if your emotional center, the amygdala (Amygdala), is activated (whether by fear, excitement, or being moved), the amygdala stamps that recording as “urgent” or “important” And then there’s reconstruction in the present: memory is every act of reconstruction. Every time you recall the past, recall a doll, you’re actually reassembling that memory’s neural pathways with your current state of mind — memory flows The body’s ledger, the concrete embodiment of emotion, isn’t kept in the cerebral cortex but is body memory; embodied memory is essentially a replica in flesh, which is why the felt experience is so deepLanguage’s underlying grammar solves common problems
The sacred-body rites of Nachi Falls
Yu Qian, the Ming backbone who died standing
Epictetus on cause and the present moment
Stability is the core logic of Chinese policy
Reform and opening up was a self-rescue stability project
The law of the peasantry the Chairman read out of Water Margin
The real logic of state granaries during the great famine
Sages shaped by history and transcending it
The regret and reflection that history allows no hypotheticals
Moral politics in the Ming vs. interest politics in the Qing
Learn from history, don’t yearn for it
Russia’s annexation of territory did permanent harm
The subconscious learns through a game-like approach
Zao Wou-Ki’s lyrical abstract art life
Hōryū-ji’s Asuka-period style
The historical changes of the Silk Road
The Tang remains of Foguang Temple’s East Hall
The six paramitas’ practice of reaching the other shore
Buddhist culture differences between Guangzhou and Hangzhou
The evolutionary lag of the fear instinct
Social mismatch and the Dunbar’s number limit
The evolutionary roots of the human eye’s adaptation to visible light
Snakes are cold-temperature, not cold-blooded
Sensitivity and dullness as survival strategies
The reproductive-payoff logic of sexual selection
Inclusive fitness and altruism
Reciprocal altruism as game theory
The myth of biological determinism
Cinderella’s evolutionary anxiety
The resource game between mother and infant
The evolutionary logic of the grandmother effect
The social signal of self-deception
Dream of the Red Chamber reshaped Chinese tragic aesthetics
Ten Days of Yangzhou: A Dark Age of Civilization
Story as Humanity’s Explanatory System for Survival
The True Zen Practitioner Does Not Avoid the Dust of the World
Sanjusangendo’s Modular Unit and Guanyin’s Manifestations
The Eleven Psychological and Ethical States of the Many-Faced Guanyin
Buddhas and Bodhisattvas Are Different States, Not Ranks
The Jade Emperor as a Cosmic Projection of Ancient Political Structure
The Power-Logic Satire of the Monsters in Journey to the West
The Meaning of Temples and Statues as Centers of Spiritual Placement
Praying to Buddha Is About Uncertainty, Not Awakening
The Systemic Difference Between Compassion and Pity
India: The Split Between Legal Equality and Social Hierarchy
Observing Buddhist Temples and the Question of Life’s Meaning
What Temples Mean and the Question of Life’s Answer
Zhuangzi’s Wisdom of Forgetting Each Other in the Rivers and Lakes
Four Root Layers of Language Learning in the AI Era
The Instinctive Guard When Interrupted
The Linguistic Difference Between Mandarin and Cantonese
Breaking Down US Bank Account Types
The Perfect Illusion of OurDream AI
The Destruction of Opium at Humen and Its Historical Lessons
Reading Has to Touch Something Inside You
Slow Down and Let Your Thoughts Settle
Phone Reading Is Blocked by Low-Cognitive-Effort Temptations
Timber Order Combined with Mirrored-Space Illusion
The Erlitou Culture and the Mystery of Verifying the Xia Dynasty
The Polis Civilization Determined by Ancient Greece’s Geography
Aristotle’s Realism and the Syllogism
The Buddhist Emptiness Is Really the Interconnection of All Things
Self-Awareness Is the Soul’s Metacognitive Ability
The Ultimate Unity of the Dao Following Nature and Tathagata
Tathagata Is a State of Being, Not a Person’s Name
The Road to Seeing Tathagata by Seeing Through Appearances
The Worldly Wisdom of Producing the Mind Without Dwelling Anywhere
Analyzing the Diamond Sutra’s Three-Part Dialectical Logic
The Essential Difference Between Merit and Virtue in the Diamond Sutra
The Three Minds Cannot Be Obtained, and Meaning as a Verb
Healthy Competition Is About Cognitive Advantage, Compounding and Irreplaceability
Dance and Art Bypass Reason and Return to Bodily Sensation
Hypewriters Chase Topics, Not Serious Assessment
Infection Isn’t Contact; Defense Is the Line of Defense
Stories Simulate Evolution, Papers Extract Structure
Large Organizations Compress Uncertainty Onto Individuals
Suffering Itself Has No Meaning; People Assign It
Over-Respecting the Rules Can Kill Innovation
The Tulkun’s Choice of a Peaceful Civilization
Jin Yong’s Storytelling and Spiritual Order
The Authority Effect and Consecration Superstition
Male Subjectivity Sliding Toward External Scoring
Trading Thirty Years of Mortgage for Freedom
Victim Mentality vs. Gamifying the World: Opposite Emotional Readings
The Human Brain Is Built for Linear Induction, Not Nonlinear Deduction
Social-Evolution Rewards Amplify After Fulfillment, Driving You Upward
Judging by Common Sense Once Identity Is Stripped Away
Valuation Is Consensus Pricing, Not a Question of Wealth
Matching Systems Thinking with Game Thinking
Quantified Self: Exploring Your Body to Live Healthier
Violent Stress Relief Compresses Uncertainty into Certainty
The Amygdala Stamps Memories as Important
2. Travel, Places and Cities
55 entries Hiking the Kumano Kodo An immersive experience The best part of the whole process is walking alone, being your own process You only need to focus on yourself and the road under your feet just walk naturally no need to break through yourself no need to break through nature respect yourself, respect nature Look at hiking through the lens of a game, and you find that right now this is a hiking game, and the one playing is a player — this is the player mindset Few people in Japan seem to eat while walking it’s actually culturally driven In Japan, walking and eating is seen as impolite behaviour. Even if you’ve bought a rice ball, people usually stand at the convenience store entrance and finish it, throw the rubbish into the sorting bins at the door, and then walk on The typical classic office worker will buy a cup of instant miso soup plus some bento For young people, it’s some rice balls plus a few side dishes Chinese people have many ideas and enormous energy, but there is often a kind of tension ideas fly very high, while landing them in execution depends a bit on talent But Japan’s fine-grained culture breaks one thing into ten steps, and most people can execute each step steadily turning order, detail, service and rules into a “comfortable experience” rather than a “forced constraint” Nara’s deer The deer’s story, from the god’s mount to bandits This story goes back to 768 AD. Legend says the main deity of Kasuga Taisha, Takemikazuchi-no-mikoto, in order to protect Heijō-kyō, rode a white deer flying over from Kashima Jingū in Ibaraki Prefecture From then on, Nara’s deer were seen as “messengers of the gods” (sacred deer). In ancient times, killing a Nara deer was even a capital crime (legend says someone really was executed for it) Today, around 1,300 wild deer live in Nara Park. Though regarded as national treasures, they are also extremely clever “opportunists”. They will nod and bow to tourists (actually a habit of begging for food); if you hold shika-senbei in your hand, they instantly turn from gentle “Bambi” into a “bandit gang” surrounding you In Nara Park, besides the deer, there is Tōdai-ji, the giant Tōdai-ji In the 740s AD, Japan suffered a massive smallpox plague, famine and political turmoil. Emperor Shōmu of the time felt deeply powerless and decided to rely on the power of Buddhism to protect the country Building the Great Buddha Vairocana (the giant Buddha in the Great Buddha Hall) nearly drained the Japanese treasury of the time. Records say that casting this great Buddha used almost all of Japan’s copper When you walk into Tōdai-ji’s great hall, that overwhelming sense of vast space and the heaviness of history cannot be conveyed by photographs. It is the grandest proof left by people 1,300 years ago in search of peace of mind Many temples in Nara look older and more weathered than those in Kyoto. For example Tōshōdai-ji (founded by the Tang monk Jianzhen), where there is a kind of tranquillity and solemnity left over from the golden age of the Tang Tōdai-ji is not just a temple it’s a world cultural heritage site the largest ancient wooden building in the world — its current scale is actually only two-thirds of what it was when first built This was built by Emperor Shōmu of Japan with the full strength of the nation, intending to protect the country with Buddhism The two Kongō Rikishi statues (Niō statues) on either side of the gate are the pinnacle of Japanese sculpture history. Made by the Kamakura-period sculpting geniuses Unkei and Kaikei The Great Buddha Vairocana is 15m tall and weighs 55 tonnes; the right palm faces forward (abhaya mudra, meaning “don’t be afraid”), the left palm faces up (varada mudra, meaning “I grant your wish”). Because of war, the great Buddha has been burned down several times and even lost its head; most of the current body was restored in the Kamakura period and the head was recast in the Edo period, and only on the lotus petals of the base can you still find a small part of the original Nara-period bronze carving. This Buddha statue is itself a history of Japan’s wars Nara’s fawns people and nature living in harmony Xu Xiake died only 3 years before the fall of the Ming dynasty. Three things about Xu Xiake’s travel records: he travelled far, he wrote finely, he lived truly He recorded not only scenery, but routes, distances, changes in terrain, local customs, and sometimes “almost died” notes — like falling into a pit, being surrounded by bandits, despairing on a barren mountain unable to find water Xu Hongzu didn’t want to take exams, didn’t want to stand out, didn’t want to be remembered in history; he just wanted to play Packaging “The Travel Diaries of Xu Xiake” Isn’t this exactly what the Xu family liked, the happy thing The tone in which he writes his travel diaries is often like a chemistry student’s delight at seeing an unknown reaction Later generations seem to try more to construct Xu Xiake into an idol; the further away the era, the more we want to entrust some cultural complex to one clearly visible figure Xu Xiake’s observing power Xu Xiake’s skill in observing lies also in that, on the basis of objectively observing nature, he was good at grasping characteristics and exploring laws from the movement, change and interconnection of natural phenomena He didn’t just write about strange pines and rocks; he would also study the structure of three bodies, the grain of the stones, the drop of the waterfall — he could find some logic in all of them; this is treating nature as a laboratory The water of the Three Gorges: he doesn’t marvel at it, but keeps asking where the water comes from, why whirlpools form, how the gorge carves the rock walls. That curiosity is not a poet’s, but a way of thinking that is born wanting to crack the world’s code. He seems to believe that mountains and rivers are not for appreciating, but for solving puzzles He is a rare person who writes with his own life, wanting to confirm this world with his own hands A friend said the sky here in Japan is a lot like the sky in Tibet I think so too, both are very transparent Tibet is because of the high altitude — the air density itself is low, the atmosphere is thinner than on the plains, light passes through less medium, and there are fewer scattering impurities Visually it’s more of a black-blue. The background is too dark, the light too strong Japan is coastal, a long island country, surrounded by sea on all sides. The wind blowing over from the Pacific is extremely clean and highly mobile, able to quickly blow away pollutants produced on land. Unlike inland basins, it doesn’t easily accumulate haze Japan’s controls on industrial emissions and vehicle exhaust are very strict, so the concentration of man-made aerosols is very low Wakakusa-yama A small hill with three tiers of grass slopes, only 342 high Many mountains in Japan are forested mountains, but Wakakusa is different; standing on the grass slope you see no obstruction, only Nara city, the deer, the ridgeline Deer are seen as messengers of the gods and have been protected for a thousand years, so they naturally live between the foot of the mountain and the grass slopes The deer up the mountain “live their own lives” more than the deer in the park; they don’t really snatch biscuits, nor come asking to be petted — like real wild residents Wakakusa-yama also has a fire festival, held at the end of winter every year, setting the whole mountain alight so the old grass burns away and new grass can grow better — a thousand-year tradition The concept of Leave No Trace Most wildfires in the US are caused by unreasonable fires in the wild And there are a lot of bears in the US; if you can’t follow the leave-no-trace principle and “store food properly”, not only will bears come rummaging in your tent — once a bear eats human food and develops learned behaviour, it will be shot Hikers sometimes even argue over no-trace. For example, should fruit pits and orange peel be buried? They can degrade, so many people think burying is reasonable. But the “Leave No Trace” textbook and my American companions say burying is not OK. Hunting is allowed in the US, and some hunters have opened up a deer’s belly and found lots of undegraded orange peel inside But maybe, considering local protected animals, they may not be familiar with these things and might swallow the food too For many of us outsiders, we are in a tourist mindset. For them, this is the place where they live Our population base is too large; from an education point of view, it takes a very long time and a lot of damage before we catch up with the rate of damage, and we may still not reverse the outcome of the damage Actually we have many very successful public education cases: anti-drugs, speed limits, drink-driving, sex education — even the national dental health group came to teach us to brush our teeth in primary school… If we can go top-down, the efficiency would be very high Two countries at war the ones truly hurt the deepest and paying the greatest price are exactly the ordinary people of both countries with the least say. This is almost the standard script of all “inter-state confrontation” today Both governments, and the companies of both countries, don’t lose much It is a carefully calculated “lowest cost, high effect retaliation strategy” The ordinary people of both countries actually don’t hate each other that much (Chinese people still love sushi and watch anime, Japanese young people still scroll Bilibili to learn Chinese) The subway business model itself: the more usable it is, the more money it loses The denser the routes, the higher the construction cost is bound to be, and the stronger the demand Construction costs take decades to amortise, and fares are simply not enough Subway costs cannot be paid by users the way commercial products are; they must be split across all of society Past model: build subway → sell land to make money → build more subway Future model: build subway → attract urban population and enterprises → generate long-term tax revenue And subway fares are very likely to rise — partly the logic of the algorithm Wooden structures are extremely easily damaged — fire, wind and rain, insects, earthquakes… so those that survive are extremely rare Nanchan Temple is also in Shaanxi Nanchan Temple is located in a remote spot northwest of Lijiazhuang village, Yangbai township, Wutai county, Xinzhou, Shanxi province, China, close to the famous Buddhist holy site Mount Wutai Though called a “temple”, it is not a grand imperial temple but a village-initiated, local temple, once used as a place of village Buddhist faith and practice Its location is remote — Wutai county in Shanxi, far from big towns. Because it is rural/mountainous, it was hard for it to be touched by successive political and religious upheavals, suppression of Buddhism, and wars. It is exactly this “remoteness” that became its greatest protection I remember Liang Sicheng / Lin Huiyin discovering Nanchan Temple Tang-dynasty wooden buildings were thought to have “all but disappeared”, and the Japanese even declared that China had no extant Tang architecture Liang Sicheng favoured Hōryū-ji, Lin Huiyin favoured Foguang Temple, Shanxi’s Foguang Temple The main hall’s architectural proportions are “solemn, composed” — not simply beautiful, but a beauty formed jointly by human reason and faith. This is a beauty based on structure, a freedom born of constraint She wrote a line (about ancient architecture): the beauty of our country’s architecture lies above all in its reserve and its overall composition The position of Hōryū-ji Although historical records say it was destroyed by a lightning fire in 670, it was rebuilt soon after — the existing main wooden building complex roughly dates to the late 7th to early 8th century, that is the “Asuka period / early Nara period” Hōryū-ji is widely considered “one of the oldest extant wooden building complexes in the world”. In the West Precinct (Saiin Garan), the Main Hall (Kondō), the Five-storey Pagoda, the Central Gate (Chūmon), the Cloisters (Kairō) and others are all 7th–8th century buildings Hōryū-ji and its surrounding related building complex are listed as UNESCO World Cultural Heritage (“Hōryū-ji Area Buddhist Monuments”) — the first site in Japan to be recognised as world heritage If we strictly require “composed of multiple halls, with the existing main structure still original-period timberwork” — Japan’s Hōryū-ji West Precinct (7th century) is indisputably first (parts of the buildings roughly 670-693). “The oldest extant wooden building complex in the world” — Japan’s Nara Hōryū-ji is still recognised as the global number one (the earliest-listed wooden building complex on the World Heritage list). Within China, the oldest large-scale wooden temple complex is Fengguo Temple in Yixian, Liaoning, and Chongfu Temple in Shuozhou, Shanxi They rate Japan’s ancient architecture very highly; Liang Sicheng once said: “Hōryū-ji is the brother of the Tang-era Great Wild Goose Pagoda”, and Lin Huiyin also thought Japan had preserved many Tang architectural styles already lost in China Tōshōdai-ji A place I didn’t go to this time It is also a world cultural heritage site The first thing Liang Sicheng did on coming to Japan was head straight for Tōshōdai-ji Although Tōdai-ji has a long history, the Great Buddha Hall you see now was actually rebuilt in the Edo period (1709), and its style has already become Japanese. But Tōshōdai-ji’s Main Hall (Kondō) is the only and most complete 8th-century Tang-style building still extant in the world. Those famous 8 giant Greek-style columns (entasis), those magnificent yet simple roof lines — that is the real “dreaming back to the Tang” Nearby there is also Yakushi-ji; the founding of Yakushi-ji is very romantic Emperor Tenmu, to pray for the early recovery of his ill empress (later Empress Jitō), made a vow and built this temple, dedicated to Yakushi Nyorai (the medicine Buddha). So this is also an important holy site for Japanese people praying for health and longevity and the removal of illness Also, this is where the parietal-bone relics (divided relics) of Master Xuanzang (the prototype of Tang Sanzang) are kept The famous Japanese painter Hirayama Ikuo spent 30 years retracing the Silk Road and painted a giant 49-metre mural, “The Great Tang Western Regions Mural”, which is enshrined here Wooden structures can also reach an extremely high level of aesthetics, religion and culture Some ancient wooden complexes that actually survive today The Nanchan Temple main hall (single building) is currently the oldest extant wooden building (single structure) in the world, but strictly speaking it is a single building, not a complex The Foguang Temple East Hall (single building) is the second oldest Tang wooden single structure There is also the Sakyamuni Pagoda of Fogong Temple (Yingxian Wooden Pagoda), the tallest extant pure-wood pagoda in the world (67.31 metres), a single structure but enormous in scale And Japan’s earliest world heritage site, Hōryū-ji The Kondō is about 18 metres wide (11 bays), the five-storey pagoda is 32.5 metres tall, all borne by wooden beams and columns plus dougong, with no modern metal parts Japan has many earthquakes, so all-wood structures must achieve “flexible earthquake resistance” (mortise-and-tenon joints + dougong damping), and they still stand after 1,400 years It uses a triple dougong system of “cloud arms + cloud blocks”, the earliest and most complex extant dougong in existence (more complete than any Tang dougong surviving in China) All the giant timbers (the central pillar nearly 1 metre in diameter), with mortise-and-tenon tolerances that must be controlled to the millimetre — 1,400 years ago, without electric saws Liang Sicheng, after surveying Hōryū-ji in 1935, sighed: “We searched in China for over a decade for Tang wooden buildings and never found even a single column or a single dougong… I never expected to see a complete Tang timber framework in Japan, and one even more exquisite and bolder than we imagined.” Hōryū-ji If I had to pick one in Nara, apart from Tōdai-ji, I think I personally prefer Hōryū-ji Ultimate aesthetics, including the persistence in restoration in later periods letting culture keep running Durability is not toughness, but a going-along-with process, cooperating with nature rather than fighting it Chinese Northern Dynasties Buddhism → Korean peninsula craftsmen → Japan’s Asuka period → the Nara architectural system Just this plain wooden structure, alive for 1,300 years Hōryū-ji has no ambition to show off. It does not seek greatness, it is merely earnest Every piece of wood, every mortise and tenon, every pillar, simply does its duty Not impulsive, not affected, not flattering the age ,,, a sense of authenticity Kyoto, travelling back to the Heian period Kyoto in the morning feels especially like time has just been switched on. There are no tour-group flags at the temple gates, only bicycle bells in the alleys and the atmosphere of monks chanting sutras The steps of Kiyomizu-dera have just been lit by the morning light; the stone steps have a kind of cold air of having just woken up The reflection of Kinkaku-ji on the pond surface is like ink that has never been stirred At Fushimi Inari Taisha the thousand torii gates have no queue of photographers yet; you can truly “walk through the torii” rather than “queue up to walk into the scenery” Kyoto’s sunset has a theatrical quality; street lights come on and the old buildings stop looking like history and start looking like a stage set; there are many people but it isn’t noisy — instead it makes you feel “the living ancient capital” is right there in front of you Kyoto in the evening turns outward, towards life, like drinking a cup of warm sake America’s electoral architecture America’s underlying code is Christian culture; Christian culture is not only a faith, it is essentially also a contract spirit But its influence on politics is not directly through “doctrine”, but through organisational forms and values The system architecture at the top is actually the political system America’s political system (the constitution, separation of powers) did not copy the Bible, but was deeply influenced by the European Enlightenment. It is a rational, secular architectural design. The First Amendment explicitly forbids establishing a state religion. This is precisely to prevent religion from directly controlling politics the way it did in Europe The electoral system is a compromise between “elitism” and “democracy” — neither a purely religious product nor a purely democratic one, but an engineering compromise In the earliest founding period, elites did not trust the masses (who had low education at the time). They designed the electoral college to put a firewall between the “mob” and the “decision-makers” Core (Kernel): Christian values (defining what is “good”). Architecture: the constitution (defining how the system runs, how to handle Conflict). Interface: elections (defining how users/citizens interact with the system). As long as one of these three collapses (for example, no longer trusting the contract, or the architecture being unable to handle the complex concurrent requests of modern society), the stability of the whole system is in trouble. This also explains why modern American society faces division — because the underlying “Core” and the upper “Architecture” are having a compatibility conflict The first time I saw a café whose main selling point is cups Kyoto Kaikado Café Hard to see with the naked eye; you only know the details once you handle it The details are not only the coffee beans of traditional coffee and the details of service, but also shown in the blend of background music and the coffee machine, the curve of the cup, the unique texture of the cup, the uniquely made tray, even the angle you drink at, how it feels in your hand, how it feels on your lips With a cup of coffee and dessert for a few thousand yen, you get to “use” and “touch” top-tier craft objects worth tens of thousands or even hundreds of thousands of yen It was formerly the 1927 Kyoto City Tram garage and office The ceiling height is very high, retaining the concrete texture of that time. This “Taishō/early Shōwa Western-style industrial feel” is very rare in Kyoto, completely different from traditional Japanese wooden houses. It is registered as a tangible cultural property; this “rebirth from ruins” style was later imitated by many shops, but this is the original The coffee tastes very good too The ultimate experience of drinking coffee at Kaikado Ultimate user experience starting from the details Extreme subtraction of function, to get an extreme experience of function It uses the user’s intuition and natural laws, not forced guidance The user using it is the last step in product design Kaikado’s metal oxidises. The oils on the user’s hands, the frequency of use, the way it is stroked, will decide what colour this tin becomes years later. Every tin is only a “semi-finished product” when purchased; only after 10, 20 years of use by the user does it become a unique “finished product” The user’s time iterates inside the product The traces the user leaves A café that has survived 140 years; 140 years ago they made tea caddies because Japanese people drank tea. Now young people don’t drink tea, they drink coffee and eat pasta. If they had insisted “we make tea utensils”, they might have closed long ago Shōsei-en Garden Shōsei-en (Kikoku-tei) A very small garden, tickets only 700 I have to say the Japanese make parks very beautifully This garden is small, yet carefully folded like a poem. The pond water winds into a “circuit-style garden” route, repeating the same water surface from different angles “Shōsei” means to cross over, to bring about, to accomplish — it means “you only understand after walking through” The garden doesn’t win by area, it plays with scale; blank space is harder than decoration, it gives the viewer room to imagine, and lets you see imperfection too % ARABICA Kyoto Higashiyama The interior design is pretty good, but I didn’t feel it reached full marks It really takes the minimalist route, good for taking photos Very few seats, just one table The sense of space is indeed strong, the lighting is on point, including the café’s own proportions and texture That street itself is very historic; instead it chooses to “step back” with clean modern lines, letting the foreground be Kyoto’s alleys Why so many elderly Japanese are hunched over Japanese average lifespan is very high; osteoporosis (bones getting thinner and more brittle) after long accumulation compresses the spine, making people naturally more and more bent; this condition is called vertebral compression fracture This is common among Asian women, due to diet and less sun exposure And a lot of culture requires lifelong bending labour — farming, housework and such Also, the concept of strength training is relatively weak; the older generation doesn’t emphasise strength training, only endurance Then, Japanese society values quietness, humility, restraint Yakushima is, in my heart, the ceiling of Japan Not scenery, but pilgrimage, a forest of time; yakusugi trees are thousands of years old, like living fossils Its moisture turns green from a colour into a living thing It rains often on the island; moss, tree roots, rocks, stream water are all wet enough to seem to be breathing Miyazaki found the forest inspiration for “Princess Mononoke” here — exactly this “nature so wet it glows” There are no easy sightseeing routes either, no need to design things deliberately for photos It doesn’t actively give you anything; it depends what you can take away I think moss is really great An expression of wabi-sabi aesthetics Actually stone is too At first it might have been a finely carved stone but it can’t beat the traces of time, slowly weathering, the facial features blurring; in Japan this weathering is seen as an acceptance of the natural within the man-made “Sabi” originally meant being alone, poor, not getting what you want; later it evolved into a state of mind that is content with simplicity and pursues spiritual wealth Sanjusangendo “The black box of the algorithm,” or “minimalist aesthetics” In the field of AI, we believe that “Quantity has a quality of its own” This is a great attempt by ancient craftsmen to approach “omniscience and omnipotence” through brute-force enumeration A thousand hands: representing an infinite Action Space — not only understanding what is heard, but having the ability to “catch” every concrete suffering The original meaning of “Guanyin” (Avalokitesvara) is “the one who observes the sounds of the world” — the one who listens to the sounds of the world A thousand eyes: representing perception in every direction There is a famous legend here: “If you miss even one person who has passed away, as long as you search carefully here, you will surely find one face among the 1001 that resembles them.” This sounds a lot like the Latent Space of generative AI The ancient craftsmen did not try to create 1001 completely identical “standard parts.” Under a standardized framework (a Prompt), they fine-tuned the facial parameters of each figure (Fine-tuning). Together they cover the probability distribution of human facial features. As an entrepreneur, standing here, you are looking at a huge dataset of human faces, thinking about how to make your AI cover this same diversity of human emotion (Diversity) Ryoan-ji Visual aesthetics, the ceiling of minimalism It pushes “zen” to the limit of subtraction. It takes away all the trees, water, and flowers, leaving only stones and sand But the amount of information is extremely high … After Jobs and the Queen of England made it famous, there are a lot of tourists, but everyone is very quiet Ryoan-ji has a “Chokushimon” (imperial envoy gate), normally passable only by the emperor’s envoys, closed all year round. But to welcome the Queen, this gate was specially opened so that she could enter. At the time this was an extremely rare, highest-grade reception What Jobs learned from it: the most extreme control is to give the user only one entrance When in Kyoto, Jobs often went to a place called “Sushi Iwa”. It was recommended by his daughter; he was picky but went anyway. When he tasted the best toro, he marveled at the purity of the ingredient and the extreme knife work, and went in person to shake hands with the chef Ryoan-ji details I kept wondering how this drainage system was made, and whether it was man-made, since there are no drainage ditches around it (I generally assume such a structure is needed) It turns out the ground has a very slight slope And to create the optical illusion, to offset the visual effect of “far things look small, near things look large,” the far walls are actually slightly lower than the near ones (or the height difference of the earthen wall is used), and the sizes of the stones in the distance were carefully calculated too The oil-clay wall — that wall was made by ramming clay mixed with rapeseed oil. It isn’t an ordinary mud wall; the oil gives it water resistance and extremely high strength. After hundreds of years of wind and sun, the oil slowly seeps out, giving the wall an extremely complex, uneven brown-yellow luster. This is wabi-sabi There is a round stone water basin called a “tsukubai.” The design is interesting: in the middle of the basin is a square water hole (representing the character “口”). Around it are carved four characters, but each one is missing its “口” radical — they all share that central water hole Read together it is: Ware Tada Taru wo Shiru — “I only know sufficiency” (contentment brings happiness) In the early Ryoan-ji garden there was actually a huge weeping cherry tree. It was very beautiful when it bloomed in spring, because people found that the cherry blossoms were too beautiful, too eye-catching. When the cherry blossoms were in full bloom, nobody looked at the stones; and when they fell, people felt sad. This interfered with the working of the core function, “zen” So the monks made a difficult decision: cut down the cherry tree Karesansui A unique form of the Japanese garden, and also an expression of Japanese painting and Zen culture Simply put, it is a “garden without water” Kare: dry, waterless San: mountains, land Sui: rivers, oceans Elements White sand / gravel: spread on the ground, raked into ripples (called “sand patterns”), used to symbolize the sea, lakes, seas of clouds, or flowing water Stone groups: stones of different shapes, used to symbolize islands, mountains, waterfalls, even animals (such as cranes and turtles) Moss / shrubs: a very small amount of plants, used to symbolize forests or greenery Karesansui is a minimalist art. By giving up real water, it forces the viewer to use imagination to “see” the water. This state of “creating something out of nothing” accords with the “emptiness” and “nothingness” that Zen pursues This kind of garden is usually not for strolling in, but for sitting and looking at (usually from the abbot’s quarters or the corridor of a temple). It helps monks and viewers clear away distracting thoughts, meditate, and hold a dialogue with themselves Real plants and trees change drastically with the seasons, water dries up or goes murky, but stones and gravel are relatively eternal and unchanging. This represents a silence beyond time Ryoan-ji: home to the world’s most famous abbot’s garden. Fifteen stones scattered in white sand; from no angle can you see all fifteen at once (suggesting that human perception is always imperfect, and only through the eye of the mind can the whole be seen) Building a city The core question, asking yourself The core vision, and the question of where to put it Why build it Where to build it This involves the city’s goal, and the siting The city’s goal matters a lot — political center, economic center, tech center, industrial center How does a city make people stay? Some landmarks can always attract people’s sense of direction and belonging Like the Petronas Towers, or the Oriental Pearl Tower And there are third places too — a city’s third places really matter; the more civilized a city is, the more third places it has: plazas, parks, libraries, open-air café seating Leverage: a city has investment properties; a city is an act of leverage Only in a big city can an extremely fine-grained skill (like “a scholar who specializes in medieval French” or “an expert in quantum encryption algorithms”) find its corresponding buyer. The city allows talent to specialize extremely narrowly and thereby earn a premium, so the city lets skills find buyers as fast as possible Nobel laureate in economics Robert Lucas once asked: “Why do people cram into Manhattan and pay such high rent?” The answer: to “steal skills” and to “run into” people Some knowledge is written in books (explicit), but the most valuable knowledge (tacit) is spread in casual chatter in cafés and over drinks at dinner. Cities provide the high bandwidth of face-to-face communication The sociologist Granovetter argued that the chances that change your fate often come from “weak ties” (friends you’re not that close to). Through high-density third places (bars, expos, even the subway), the city multiplies the opportunities to form weak ties The city’s 3T theory: technology, talent, and tolerance Truly top talent tends to be idiosyncratic in personality, lifestyle, or views. A place full of moral judgment, where the pressure of an “acquaintance society” in personal relations is too heavy, cannot keep top brains Big cities offer a precious “right to anonymity.” Here nobody cares about your family background or whether you’re married; people only care about your ability. This simple social contract makes talent feel free and relaxed The city must be sufficiently “cold” and sufficiently “open,” so that talent can focus on self-realization rather than dealing with complicated human relations The Caucasus A vast region between the Black Sea and the Caspian Sea The junction of the Eurasian landmass A strategic choke point linking Eastern and Western Europe with West Asia The Greater Caucasus is one of the highest mountain ranges in Europe. Its highest peak, Mount Elbrus, is 5,642 meters above sea level and lies within Russia Georgia: mostly Orthodox Christian. Armenia: the first country in the world to formally adopt Christianity as its state religion. Azerbaijan: mostly Muslim (predominantly Shia). Brazil and China To some extent they feel quite alike The Amazon rainforest is not a “forest” but a global-scale climate machine Through transpiration it creates “flying rivers” that affect rainfall across South America and even Africa. Once it falls out of balance, what changes is not just Brazil but the global climate structure. Besides the Amazon, Brazil also has one of the world’s largest wetland systems, the Brazilian Highlands, waterfalls Brazil did not enter modern nationhood through revolution; it slid into modernity through administrative continuity So the social structure is still ancient Brazil has all the material conditions to become a superpower Land Water Agriculture Minerals Population But it is always pulled at by the following structures: The inertia of corruption The inefficiency of the bureaucratic system The lock-in effect of the wealth gap Compared with China China is actually a civilizational entity on a super-long time scale The form of the state repeatedly collapses, and then the logic of governance is continuously rewritten; that alone matters a great deal But Brazil was interrupted and reassembled by external force; the nation is a structure assembled after the fact China is better at “turning problems into engineering” Brazil is better at “turning problems into culture” Why can Brazil write inequality into culture? How is inequality written into culture? The core of the Portuguese colonial economy was large estates + a single export crop (sugar, later coffee) + slave labor. This set naturally grows three things: Extreme concentration of wealth: land and export earnings go to a handful of big planters Governance biased toward elites: the job of public power is to protect property rights and the export order Social stratification “naturalized”: ranking people’s worth by birth and color The 1850 Land Law (Lei de Terras) changed the way public land could be acquired into: legitimate title basically only through purchase, and it restricted acquiring land naturally through occupation/clearing. Its historical context and purpose were highly tied to “slavery is about to end, and landless laborers need to be created”: making the future free workforce depend on employers rather than becoming smallholders. It means you can be free, but you have no land. No land, so you have to work. Work, so you have to obey Brazil formally abolished slavery in 1888 (very late), but there was no accompanying large-scale land reform, compensation, or social integration program; the result was that formerly enslaved people became “free laborers” but still lacked land, education, and political representation Industrialization and urbanization did create new jobs, but Brazil’s urban growth was long accompanied by three things: A high share of informal employment (no social insurance, no stable wage) Uneven educational expansion (a rupture between elite education and mass education) Polarized living space (wealthy areas and favelas exist side by side, adjacent and segregated at once) The 1988 constitution treated health as a universal right, and a unified health system, SUS, was built afterward, with the goal of access for all But Brazil’s paradox is: Legally, “everyone can have it” In reality, the public system’s resources and accessibility are uneven, and the wealthy keep relying on the private system How does space take part in thinking? For Japanese urban and interior space, the core goal has never been to let the individual unfold, but to let order run stably Cafés are small, seats are close, quiet but not relaxed — behind that is not an aesthetic question but a kind of social engineering Space is designed so that people “exist appropriately,” not so that they “wander freely” Vertical space is very different; the brain more easily enters a self-narrative mode Thinking about who you are Building a long-term vision Doing abstract thinking with no immediate return (like products, philosophy, life) And a great many Japanese cafés are exactly the opposite: Low ceilings Small table spacing Compressed sound Behavior silently regulated (posture, volume, time) The most visible dividends of Guangzhou-Foshan integration: transport, industry, a sense of opportunity The subway, the intercity rail, the half-hour commuting life circle have pulled many districts of Foshan directly into Guangzhou’s “effective radius.” This is very friendly to young people, white-collar workers, and skilled workers: you can live in Foshan and use Guangzhou’s job opportunities and resources The problem isn’t efficiency, it’s identity and distribution In the integration narrative, Foshan is more and more easily defined as: a manufacturing base, a residential hinterland, an industrial support zone The closer Guangzhou and Foshan get, the more prices align with Guangzhou’s, but the decision-making center for public resources is still in Guangzhou Housing prices and living costs are both rising, but the say in decision-making hasn’t risen in step And then there’s Foshan’s original city character — “low-key but stable, doing real industry, not loud” — which is being diluted TSUTAYA BOOKSTORE is essentially part of Japan’s Tsutaya Books system. Its core idea has never been “selling books,” but the line they themselves keep emphasizing Through space, design a lifestyle The space is divided into many small-scale areas where you can linger Coffee is not the protagonist, but it is definitely indispensable TSUTAYA in Shenzhen won’t really put live birds in (that is extremely difficult both for operations in China and for animal welfare), but it uses a lot of: wooden structures warm lighting greenery an acoustic design that is open but not noisy Indonesia Indonesia feels magical to me; I thought Indonesia was a small country But Indonesia is actually absurdly big More than ten thousand islands, and several thousand of them inhabited And it spans more than five thousand kilometers from east to west, which is almost the distance from Beijing to Europe Its land area is 1.9 million square kilometers Thailand’s soil for idols In the social culture, the attitude toward “people who can be looked up to” is quite friendly: respect, closeness, and emotional investment do not conflict. Monks, teachers, actors, singers, members of the royal family — they are already in a position that is half public, half emotional bond. Following a star is not crossing a line, but continuing that. When a fan says “I love you” to a star, in context it is closer to “you lit me up,” rather than losing yourself The entertainment industry has upgraded over the last decade, and the Thai entertainment scene has learned from the Korean model What young Thais face is not easy — weak economic growth, limited class mobility, long-term political tension. Following stars here is not only entertainment, but a low-cost source of a sense of meaning. A star provides a stable, projectable, participatory model of hope: effort, being seen, being loved, success. Even if it’s an illusion, it’s better than blankness I feel this deeply: among China’s several super-large agglomerations right now the Yangtze River Delta integration region, the Beijing-Tianjin-Hebei coordinated development region, and the Guangdong-Hong Kong-Macao Greater Bay Area the Greater Bay Area has the most future The Yangtze River Delta is the peak form of industrial civilization This is a region with “almost no weaknesses,” better at doing things that are known and correct. It will definitely be strong, but not necessarily the most elastic For the Beijing-Tianjin-Hebei coordinated development region, stability is always the first priority: stability > trial and error > disruption The Guangdong-Hong Kong-Macao Greater Bay Area is developing too fast It is the only place in China that internally contains at the same time mainland manufacturing capacity Hong Kong’s international financial and rule-of-law system Macao’s institutional buffer zone (the deliberately preserved buffer of gambling and such) Shenzhen’s extreme entrepreneurial culture the strong commercial genes of the Cantonese people a natural interface with global capital, technology, and culture The world will keep being unstable in the future Shenzhen is a very interesting city: constantly losing control, constantly healing itself In history, almost every innovation that changed the world came from a place like this The naming of the Guangdong-Hong Kong-Macao Greater Bay Area For a long time in the past, Shenzhen itself was the narrative unit “Shenzhen speed,” “the Shenzhen model,” “the Shenzhen miracle” It was a sample city of single-point explosion, representing reform, innovation, youth, and adventure The Bay Area is not a city, it is a system Not one story, but a whole framework of stories In its early days Shenzhen relied on “being a special case” and “breaking the rules,” which suited being told on its own But when Shenzhen goes from “experimental field” to “mature engine,” what it needs is no longer a hero narrative but a collaboration narrative: industrial chain collaboration, city cluster collaboration, institutional collaboration The word “Bay Area” is naturally suited to telling all of this When the benchmark shifts from “domestic cities” to “Tokyo Bay Area, San Francisco Bay Area, New York Bay Area” a phrase like “Shenzhen vs Shanghai” looks too small You have to raise the dimension, so Shenzhen gets embedded in a bigger chessboard, becoming a key node rather than the only protagonist Shenzhen has character, historical emotion, controversy The Bay Area has none of that; it sounds rational, macro, extremely planning-oriented, and especially suited to policy language, capital language, international language In local life, people still say “I’m in Shenzhen” In policy documents, investment roadshows, and macro discussions, more and more people say “the Bay Area” Some insights from engineering practice Are a lot of things being thought about in too complicated a way? AI development included Knowing which frameworks are available is more valuable than knowing how to build everything from scratch Building RAG shouldn’t be this hard either Gradio can build a web interface quickly and generate an API automatically I suddenly realized I seem to have lost a good deal of solid engineering ability This code can create a professional, good-looking web interface… with tabs, examples and a shareable link. For the past year I’ve used this pattern for all my client demos In the past, building a production interface meant learning React… setting up a backend API… handling state management… and questioning my career choices Streamlit can let you skip all these steps Pure Python development. No need for JavaScript, HTML or CSS. Just write the logic and you get a runnable web app. And it doesn’t look like a prototype… it can go straight into production use Flask used to be fine. Until I needed async support. Or auto-generated API docs. Or data validation. Or WebSocket support. Or the other hundred features a modern machine learning API needs FastAPI can do all of it… and it’s faster than any other framework in Python Run that code and you get auto-generated interactive docs at /docs….. request validation….. error handling….. and an API that can handle thousands of requests per second LangChain for LLM orchestration. Transformers for model access. LlamaIndex for data retrieval. Gradio for quick demos. Streamlit for production UI. Weights & Biases for experiment tracking. FastAPI for deployment Design techniques for timber structures The most fundamental aesthetic of a timber structure lies in “honesty.” Unlike reinforced concrete, which needs to be wrapped up, timber is often left bare, and this display of the “skeleton” forms the core of the spatial aesthetic Yes, a sense of exposure, which guides the eye and adds a sense of depth to the space The joints of a timber structure (mortise-and-tenon, bolts, or modern metal connectors) show how force is transmitted. In Eastern aesthetics, complex dougong brackets or mortise-and-tenon joints are themselves highly sculptural works of art Wood is the only living building material, and this “biological nature” brings an aesthetic experience that industrial materials cannot match Wood’s response to light is completely different from stone or metal, and it determines the “temperature” of a space If there are knots or marks, the timber structure is even more unique; it adds humans’ own aesthetic preference, the beauty of flaws — this is wabi-sabi Timber structures don’t give that diffuse reflection of light; bare wooden beams and trusses, under natural or artificial light, cast rich shadows. As time passes and the sun’s angle changes, the patterns of light and shadow indoors also flow By placing a mirror at the apex of the vertical view, you are in fact creating a “reverse bottomless pit” or an “inverted tower leading to the sky” Yuxin Bookstore A distinctive cafe I came across in Longhua, Shenzhen A chain brand of composite cultural spaces originating in Shenzhen Yuxin Bookstore holds to the idea of “do bookstores seriously, do business seriously” Its biggest draw is the influencer effect and high visual appeal; it’s called the most beautiful bookstore in Shenzhen, with an Inception-like sci-fi feel Circular/ring bookshelves (the Möbius strip symbolizing infinite loops), a time-tunnel entrance, mirrored floors and ceiling, creating an infinitely extending starry sky and sea of light and shadow. Combining a technological feel with a literary atmosphere, good for reading, taking photos, and relaxing. Rich functional zones: a boutique picture-book hall (immersive reading for children), a TEA house (new Chinese-style tea drinks), a yoga area, a study room, an art exhibition area and more. Choosing a dance Recently I’ve been thinking that dance can, to some degree, also let you enter a certain atmosphere immersively Dance is generally divided into three kinds, for example the Hip-hop type, Jazz as the jazz dance type, and K-Pop as the Korean cover-dance type For example, adult ballet helps with bodily flexibility; modern dance emphasizes freedom more, for people who are more emotional, who like artistic expression and don’t like being bound by frameworks Salsa is great for socializing, and Zumba is great for burning fat and getting fit Afrobeats is currently the hottest, most “alive” dance style in the world, perfect for people who want to unleash their nature The core of Afrobeats is the undulation of the spine and the opening and closing of the pelvis. It has a lot of waist twisting, hip pushing, and chest circling movements. For people who sit for long hours typing code and have stiff lower backs, it’s a natural medicine for “loosening, enlivening, springing and shaking,” and can very effectively open up the flexibility of the midsection of the body Hip-Hop Groove, the movement is Up & Down or Bounce A truly good-looking Groove is powered by the core (waist and abdomen); it can make your stiff old back as flexible as a wave House Dance —— a programmer’s romance, the best fit for “dancing around immersively all by yourself” Or even, without learning dance at all, all you need is a pocket (pocket / groove slot) A pocket is a kind of sympathetic resonance, shaping a sense of ease It mainly engages three organs: hearing, the core (waist and abdomen), and the center of gravity Lock in the Bass — don’t listen to the lyrics, don’t listen to the melody Put on headphones and you can even close your eyes. Try using only head nods to confirm the kick drum. When you feel your nodding isn’t “reacting” to the sound but anticipating it and happening at the same time as the sound, you’ve found the entrance Tahitian-style cloud cold brew What it combines is Tahiti, Tahitian style; it actually comes from Tahiti in French Polynesia in the South Pacific, and often brings to mind a vacation feel of tropical fruit, vanilla, coconut and flowers A method of extracting coffee grounds with cold water over a long time (usually 8–24 hours) Cold brew is essentially a production method that extracts coffee liquid by soaking coffee grounds in cold water for a long time Cold brew uses cold water, hot brew uses hot water; cold brew usually needs 8–24 hours of slow extraction Cold-water extraction suppresses the dissolving-out of some acidic substances, so cold brew coffee has lower acidity, a smoother mouthfeel and a gentler bitterness, though the aroma layers differ from hot brew So drinking it brings a very strong taste of the coffee itself A friend has recently been learning to dance It surprised me a bit The moment there’s a gap, the body automatically enters that mode Scenarios like waiting for the toilet matter a lot — those are fragments of time that are meaningless, uncontrollable, unproductive This is bodily-first experience: no need to explain, no need for cognition, no need to prove The body just starts moving in that instant It also stirred my curiosity about dance but a stable rhythmic cycle He clearly believes three things: I don’t have to figure everything out right now I don’t have to prove anything through a huge change I allow myself to slowly be shaped by one thing This is a trust in: your own body time the continuity of the world “Is there anything at all that, without me assigning it meaning or changing my life, can still slowly take over my daily life?” Change should serve them; that’s next I didn’t get to eat at Xijiade, which had a half hour+ queue at noon and again in the evening Unexpectedly, what I ended up eating was a free black sesame tangyuan from 7-Eleven Why Xijiade hasn’t opened in Hangzhou has always puzzled me The restaurants that truly survive long-term in Hangzhou tend to be: small out of the way non-standardized the owner matters a lot impossible to replicate It’s an “aesthetic dining city” Because Hangzhou consumers have a unique trait: they don’t lack choices, and they aren’t in a hurry to fill their stomachs When Hangzhou people eat out, they also don’t want to eat “something you could make at home” Asking them to spend money eating “northern formal-dinner-grade dumplings” out is mentally backwards Poland Its geographic position destined this small country to have unusual experiences Poland sits at the core of the “North European Plain,” a flat stretch running from France all the way to Russia No high mountains, no wide seas, no deserts Poland once disappeared from the map for 123 years (1795–1918) At the end of the 18th century Poland was partitioned three times by Prussia, Russia and Austria, and the state was completely extinguished. But the nation, the language and the culture were not extinguished This is extremely rare in Europe Regaining statehood in World War I In World War II it was invaded by Nazi Germany and the Soviet Union at the same time After the war it “won,” but was absorbed into the Soviet bloc Only in 1989 did it truly restore sovereignty Polish culture has an obvious keynote: tragic heroism You see it again and again in literature, music and historical narrative: sacrifice endurance resistance non-compromise This makes Poles: hard to tame and hard to persuade The upside: strong pressure resistance The downside: highly vigilant toward change Dongguan Just so you know, the Pop Mart figures on your desk, Finding Unicorn, and even a lot of the collectible figures for Hollywood movies, are most likely “Made in Dongguan” A mature toy manufacturing industrial chain Home to Huawei’s European town In Dongguan’s Dalang and where it connects, there is a place called Songshan Lake. After Huawei moved its R&D headquarters here, it directly built a “Europe” on the spot You ride a vintage little train through a cluster of buildings imitating 12 European cities including Oxford, Paris and Heidelberg Outside the walls are bustling electronics factories and urban villages; inside the walls are quiet European manors. This “cyberpunk-style” folding is a landscape unique to Dongguan Dongguan people have their own unique standard This is the soul here. Dongguan people have extremely high demands for roast goose: the skin must be crisp as glass and the meat must be juicy. Houjie laifen rice noodles and Chang’an snakehead fish will show you what “understated luxury” means It’s also where the opium was destroyed at Humen The recently unveiled Great Bay University is also located in Dongguan Shenzhen has Huawei and DJI, Hong Kong has world-class universities, and Dongguan has the strongest ability to turn “research results into products” Zhiyuan on the iOS backdoor pager situation Their team’s project, it seems, because of the iOS backdoor pager the latest version is already very hard to keep maintaining I remember it sparked an idea a long time ago, half a year back in Hangzhou In Chaoshan I came across a few kway teow soup shops I always thought Guangzhou-style kway teow soup centered on a clear broth with some freshly sliced beef, so the color would be lighter Then I had a bowl with a darker broth and it was super tasty; looking into it, it’s simmered from beef bones + the beef’s own flavor, not relying on spices The beef is thinly sliced and blanched to order, looking just past raw, not long-stewed “Just done” actually means just right — when an ingredient (especially meat or vegetables) is heated to just cooked through, but not yet overly soft or old Concretely, the color shows it’s only heated until the protein sets, going from bright red to grayish brown, with a tender, smooth texture but no raw gamy taste If you keep cooking, the beef goes from “just done” into “fully cooked” or even “overcooked,” and the texture gets worse One of China’s earliest treaty ports A port city; commerce, foreign trade, and manufacturing are all strong The population is large too, and the commercial atmosphere, consumption, and modern services are noticeably stronger Chaozhou is the birthplace of Chao culture; Chaoshan isn’t an administrative concept — Chaoshan is a cultural-linguistic-ethnic concept Shantou was originally just a port under Chaozhou prefecture It rose fast because of trade, and actually overtook the older city So it formed: Chaozhou + Shantou = Chaoshan Behind the Pyongyang Marathon It’s not a simple sporting event Inward: proving order, stability, and normalcy Outward: setting a low-risk diplomatic template Many foreigners, behavior highly predictable The itinerary can be fully designed The photography is extremely positive A partial opening of surface signals It feels like a probe, a display of controlThe historical accumulation of the Kumano Kodo
Japanese culture drives people not to eat while walking
Chinese people value ideas, Japanese people value execution
Nara’s deer went from divine messengers to opportunists
Tōdai-ji is a world-class ancient timber structure
Nara’s fawns and harmony between people and nature
Xu Xiake travelled far, wrote finely, lived truly
The scientific observing spirit of Xu Xiake’s travel records
Why the skies of Tibet and Japan are so transparent
Wakakusa-yama’s grass slopes and the thousand-year fire festival
The ecological protection principle of Leave No Trace
When two countries go to war, ordinary people pay the most
The cost-sharing logic of the subway business model
Nanchan Temple’s remote-location preservation miracle
Hōryū-ji’s status as a world heritage site
The lingering Tang style of Tōshōdai-ji
The religious and cultural height of wooden structures
Hōryū-ji’s ultimate aesthetic persistence
Kyoto’s two faces, morning and dusk
The contract spirit of American elections
The traces of use on a craft object
The product is completed by the user
The art of blank space in a small garden
The café’s spatial retreat
Why so many elderly Japanese are hunched
The life force of Yakushima
Imperfection in wabi-sabi aesthetics
The Thousand-Armed Guanyin’s Space of Perception and Action
Ryoan-ji: Minimalist Aesthetics and Information Density
Ryoan-ji’s Optical Illusion and Drainage Wisdom
Karesansui: The Waterless Garden and Zen Minimalism
The Vision of City-Building and the Value of Third Places
The Caucasus: Geography and Religious Plurality
Brazil and China: A Comparison of Civilizational Structures
How the Colonial Economy Wrote Inequality into Culture
How Spatial Design Affects Modes of Thinking
Guangzhou-Foshan Integration: Identity and Distribution Dilemmas
Tsutaya Books: Designing a Lifestyle Through Space
Understanding Indonesia’s Actual Size
The Soil for Idol Culture in Thailand
The Future Potential of the Guangdong-Hong Kong-Macao Greater Bay Area
The Narrative Upgrade of the Greater Bay Area
Insights from Simplifying Things in Engineering Practice
The Honest Aesthetics and Warmth of Timber Structures
Yuxin Bookstore’s Sci-Fi Influencer Space Aesthetics
Choosing a Dance Style and the Possibilities of Bodily Expression
Tahitian-Style Cloud Cold Brew Coffee Flavor
Bodily-First Experience Beats Cognitive Proof
Hangzhou Is an Aesthetic Dining City, Not a Standardized One
Poland’s Tragic Heroism and Geographic Fate
Dongguan’s Cyberpunk Folding and Mature Industrial Chain
The iOS Backdoor Pager and the Difficulty of Maintaining the Project
Chaoshan Beef Soup and the Philosophy of Just-Done Cooking
The Chaoshan Cultural Concept and the Rise of Shantou’s Port
The Pyongyang Marathon as a Probe of the Political Order
3. Product, Engineering and Open Source
51 entries I’m thinking, maybe it’s a world of collectibles When people stay curious about the world they always photograph and record what this is Then the LLM parses and stores it, and the user learns from it; the user can of course also be the one storing the content in the image Later the user can very conveniently call these abilities out This is one’s own simple knowledge storage, a process of curiosity radiating outward Plus some good strategy: the product is doing an interaction logic and a storage logic, and it targets exactly this specific scenario CapWords’ extreme privacy protection strategy No server storage Photos are sent locally or via a temporary encrypted channel to a third-party AI model for one-off recognition Destroyed immediately after recognition — never saved locally, never uploaded to the developer’s servers This “minimalist data-ism” builds extremely high brand trust One point I find very interesting about CapWords it’s like previewing in advance for the future AR era What matters then? The real world Truly connecting the real world through What do people need? Embodied memory things related to your own real scene This is a prejudgement and perception of future user habits Some thoughts on AI todo — ios Several features Ultimate user experience Binding with the iOS ecosystem A simple entry point A simple user interface Data science is quite a grand field Thinking of data science as an intersection of multiple fields including computer science, which provides the tools and compute for handling data, including programming, databases and algorithms and mathematics and statistics, which provide theoretical support, finding patterns out of randomness and finally domain knowledge, which is the soul: without an understanding of the business — including finance, healthcare, e-commerce — data is just numbers with no value What does a data scientist do every day? Obtain: collect raw data from databases, APIs, sensors or web scrapers Scrub (clean data): this is the most time-consuming step (about 60-80% of the time). Handling missing values, outliers, format conversion, turning “dirty” data into “clean” data a machine can read Explore (exploratory data analysis EDA): before modelling, first use statistical charts and intuition to discover patterns, trends or correlations in the data Model: use machine learning algorithms (like predictive models, classification models) to train on the data, letting the machine learn to “predict” or “decide” Interpret (explanation and visualisation): this is the most critical delivery step. Turning complex model results into insights or interactive interfaces that humans (like management, users) can understand Data science actually solves the future AI hallucination, and the biggest problem with data How to analyse data, break down causality Data analysis checklist Vectorisation is the foundation; be familiar with vectorisation problems Vector database problems New architectural capabilities, including RAG, and engines in the direction of data engineering Also the new interaction and tuning And advanced prompt engineering, plus Eval Understanding materials science Physics studies the laws of the world Chemistry studies the reactions of matter What is materials? I’ve been puzzled about what materials science really is Materials science is the bridge between physics and chemistry This also determines how we build the world, how we connect this world The core of materials science is not “discovering new elements”, but arrangement and combination. Its highest programme is a tetrahedral model (MSE Tetrahedron): Structure: how are the atoms arranged? (micro) Processing: how is it made? (means) Properties: what characteristics does it have? (does it conduct? is it hard?) Performance: does it work well in a product? (macro) The most classic example: carbon (Carbon). Atoms arranged in layers → graphite (soft, black, makes pencils). Atoms arranged in a tetrahedral network → diamond (hardest, transparent, makes a diamond ring). Essence: the composition is exactly the same, only the structure changed, and the world changed. A supplement to your worldview: every leap of human civilisation is, in essence, because we have mastered finer control over the “microstructure” of matter So materials science is more about discovery and assembly, discovery and creation If we classify the physical substances in the world, through the lens of materials science, by the different ways atoms bond, they fall mainly into four categories Metals Ceramics Polymers Composites Materials science is the bottleneck of hardware innovation The limit of Moore’s Law: chips can’t get smaller, not because the design can’t be drawn, but because the physical limit of the material silicon has been reached (leakage, heat dissipation). So now we are looking for “graphene” or “photonic chips” And today’s range anxiety is actually because the chemical energy density of lithium batteries is just that high If algorithms (AI) are the software acceleration of productivity, then materials are the hardware boundary of productivity. Every breakthrough in materials (like high-temperature superconductivity, materials for the inner wall of controlled nuclear fusion) instantly rewrites the global economic map Setting aside the underlying logic, we feel that if wood is a composite material Wood has its own assembly logic behind it; these basic materials can also be assembled into others, but why wood, and wood has been given its own meaning What is the purpose of materials? What exactly are we assembling? So, what is materials? Materials are also a methodology for us to understand the world Materials sits between physics and chemistry; it determines assembly, and meeting the world’s future challenges Callister’s “Materials Science and Engineering: An Introduction” is the “bible” of materials science. For an INTJ-type learner, the thickness of this book might scare you off, but its logical density is extremely high, very suitable for you The “source code” of the microscopic world (The Kernel): atoms have a potential energy curve; when two atoms come close they both want to attract (attraction) and repel (repulsion), finally settling at an “energy minimum point” (deep in the potential well) Defects are beautiful; a perfect crystal is useless The key to metal deformation is dislocations If metal atoms were arranged perfectly, then to bend an iron rod you would need to break all the atomic bonds in that layer at the same time — that would take enormous energy What I hope is that users can just photograph something casually, photograph it as a fridge magnet What should the form of the follow-up game’s questioning be? I think from the initial image parsing to the final form of the user’s questioning What is recorded is something the user is currently interested in — a doll or an object The user wants an explanation, and wants to keep a memento The product can ask deeper questions and dig out information; this can also serve as a later test case Some of my own reflections and sparks on AI products The product’s entry point is simple, the product’s functions are clear, and users have a fairly clear cognitive response to the product Not starting from the structure and preferences of the product’s designer, but from the user’s thoughts: a single thought of the user can satisfy their needs and desires, and AI plays the role of the dispatcher here An app without an entry point — shifting from app-centric to intent-centric The philosophy of the traditional App is: the world is made of functions, I package functions into an App, and you learn me In the AI era, the product takes the user’s intent as its origin and breeds out a network of product functions from there. Let’s return to the user! The app’s sense of presence will decline, but the sense of presence of capability will rise Apps will degrade from “front-end tools” into “back-end organs” The brand feeling of an App will be weakened Above the AI scheduling layer: Users can’t remember who you are Users don’t care what you look like Users only care whether “that thing just now went smoothly” This is a cold shower for entrepreneurs The future might be: A large swath of B2C apps die B2AI (capability suppliers) survive The real entry point becomes: system-level AI / Agent OS One you talk to over the long term, that gradually understands you, and can act across tasks If the future is “intent → result,” then what about things like Douyin and games — “I don’t want a result, I just want to be consumed”? I think entertainment products won’t disappear, but they will evolve from “content containers” into “attention ecosystems” — and they’ll be tougher than tools Their purpose is to seize the user’s focus, and this is the essential point of differentiation from tool products Never finishing, never ending, never letting you notice the time But products will definitely also have a stronger perception of the user’s state Entertainment needs a sense of losing control The more a product consumes time, the less suited it is to being integrated into a “unified entry point” Tools need to be gathered Entertainment needs to be isolated The future is very likely: One super Agent manages reality A pile of “attention black holes” specialized in escaping reality You’ll bounce back and forth between the two So future products are basically two big categories: Efficiency — intent-completion apps Immersion — attention-consumption apps What will really be scarce in the future is not function, nor content, but “psychological rank” Back to that question What do I want to do? What kind of product caters to the future market? Intuition is a reusable judgment function compressed out of a large number of failure samples Truly useful intuition judges the “ecological niche,” not the “feature” In which scenarios does the user think of it How does the user use it After using it, what else is there? I’m thinking about a few questions? If a system-level Agent exists, will this need be satisfied with a single sentence? If I delete this App right now, will users feel “pain,” or just “annoyed”? Is this behavior planned, or did they slide into it? Planned → tool Slid into → entertainment Neither fits… uh oh Why do systems still require me to make discrete choices “I want to hear some music, now.” Not I want to open an App I want to choose a service provider I want to specify a source Forcing you to take a vague, continuous, emotion-driven intent and translate it into a precise, discrete, engineering-executable instruction Can the system pull up NetEase Cloud Music based on my usage habits instead of asking every time Default + correctable instead of choose + confirm This is an interaction model that human society has ground out over thousands of years Some fairly good methods for observation: Have a decent observation and understanding of the world (a cognitive model) Extract some problems, pose the questions, and see whether there are better ways to solve them, possibly in the form of an app How to iterate fast on the experience of using it in that scenario is very important; what you have to think about here is how to validate the scenario and whether the scenario can actually run — some methods: The feel of use: get the basic flow working, and quickly and continuously validate this smallest unit of action. You absolutely have to bring it into daily use here, not just talk big — talking big is meaningless. So my earlier products felt too much like talking big: I didn’t even have a frequent usage scenario myself, and it’s very hard for that optimization loop to cross over on the user’s side Frequent adjustment: inspiration revolves around this usage habit and usage experience. The product is the best experience environment; when we use the product we step into the user’s usage environment, and our thinking in turn gets guided by that environment, so some new ideas naturally keep being iterated and tested, spiraling upward Boundary value: boundary value is finished the first version, reaching a publishable, usable, chargeable form — it can be React Native or a native iOS version; once the product passes, you just need some means to add the payment model in Collect user feedback, then keep iterating Some thoughts on product MVP The smallest closed loop — what I have in mind can be turned into a tool type, but it can actually also be the user’s smallest cycle of feeling, that is, whether the feel of use described above reaches positive feedback. This is especially important, and what you need to think about is how the product can give the user a better feeling — and of course there is also the point of whether existing LLMs can reach an effective and high-quality output date. Setting aside the user’s feelings here, the product form either can be split into two parts, or it can, functionally, make the user have a cognitive response, or, as an entertainment or consumption object, whether it can surprise the user and make them happy to use it — there’s an aha here The choice of technology: if it were me a year ago in university, when AI was just budding, at the very beginning I was also just drifting along, mixed into the tech circle — if this technology was impressive I’d go learn it, if Kubernetes was impressive I’d go learn Kubernetes, if LLM was impressive I’d go learn LLM. After a year of traveling the world, my own form has returned more to real, actual feeling and human experience — at Sanjusangendo or at Ryoan-ji, karesansui, in a new era, how do you bring in your own subjective perspective, your own understanding of the world, your own taste, and truly treat the product as art, to create, to truly change the user’s life, even to let the user clearly perceive their own change Quantifying behavior — this is a kind of taste, but not only taste; I don’t know how to describe this feeling. I think I still need to throw myself entirely into one product and get really great feedback before I can deeply understand these questions. Quantification is a rational act, it can be understood as an experiment, and it can also be understood as a preconception obtained from past experience — note, a preconception, not a prejudice. It’s not that I’m against quantification, it’s just that before quantifying, in the process of creating, shouldn’t we return to our own intuition and feeling Actually, after saying so much, the core thing is that I want to truly clarify my own thinking. I keep wanting to ask myself what I should do. Maybe there is no answer, but I can’t help wanting to know: if I don’t do anything, will I also regret it? I think there is an answer. I want to take my own experience, my own learning ability, my own cognitive methods, my own understanding of the world — this whole set — and truly apply it to how to create products in the AI era … I also want to ask, in this process, when observing, whether I can look at this world with a bit more compassion — at this world’s vulnerable groups, the groups that the tide of technology and the tide of the market economy have not yet made disappear in the long river of history What I want to ask is maybe not what they want, but … I’m thinking that people’s habits really are heavily influenced by environment For example, if a person really likes working in cafés, what they necessarily have to do is observe cafés, and they necessarily know which cafés have real aesthetic sense A certain environment will also naturally trigger certain habits in a person, which is why we need to study in a library or a study room You could also call it a sense of ritual Couples habitually date in cafés or Western restaurants; the Western restaurant’s lighting, quiet atmosphere, and sense of ritual promote the emotional, feeling part on both sides When we go meditate, we choose some karesansui, or a lakeside, or nature; you could call it an energy field, or you could call it the guiding force an environment has on people But I actually have another question: for me, what reactions has the environment triggered? First is taking photos. I want to record the places where I’ve found beauty, and I want to be able to express it out Then it’s thinking. This place is so beautiful, I want to know it, I want to understand it, I want to eliminate uncertainty — just like meeting a girl you like and trying to judge whether she likes you: you have to keep making contact, keep collecting information about that, and then rely on feeling or intuition or rational ability. I want contact; contact brings excitement, triggering dopamine’s expectation and satisfying release A snapshot — how to describe this? Imagine: what is our biggest Chinese trait? We always feel longing and yearning toward the sense of lack for what we cannot get. Compared with what’s near us, we look more at what’s far away. If we’re unsatisfied with our current school, we stop caring about any of its petty business, and instead become infatuated with the next, better school. So a good environment stirs up in people: this is where I want to go. This is also why Shanghai, or even Barcelona, the atmosphere of the whole city, even its petty-bourgeois, refined quality, attracts these people to long for it, and then to have the idea of settling in that city. A friend of mine would move to Chengdu even with a pay cut, because Chengdu has snow mountains … If a city is really crappy — like the city the author grew up in, which I really think isn’t good: people talk harshly, they have no manners, there’s a lot of trash on the roads; this part I feel especially strongly — so naturally I have no interest, and I don’t even want to go again What’s the use of saying all this? What does it have to do with making products? Think about it: designing a city and designing a product are similar to a certain degree. People in Chengdu think about which little bar to go drink a couple of glasses at tonight. People in Shanghai think about which café sister to invite out this weekend, to chat, take photos, and chill. People in Beijing think about calling some buddies out this time to drink, shoot the breeze, and go see some exhibition, cultural exhibition, or museum; when chatting, it also unconsciously leans red. Many people in Shenzhen think about going to a park, or going to Hong Kong to hike … What a product provides is also an environment, and a product can also be a trigger within the real environment. Let’s think about one question: in certain scenarios, how does the user clearly recognize this product? That’s when the product’s tool attribute is clear. A product can also build an environment of its own, and this will then guide what the user does; at that point the product’s positioning is an entertainment product, an immersive product Some recent sparks that make me want to act Fuse my professional abilities, bring out my creativity Start several tasks, and act as God myself to observe my whole process of learning and creating, and record it And in that process, keep iterating and thinking I want to try fusing some Buddhist methods into making products, treating the world as a dojo for a game, to create my own creative experience; some methods of creation — record them, iterate If the capacity for awakening is applied to products, how do you maximize the creativity and learning power of that awakening? Of course, staying awake all the time is very hard Open several windows in notion, and start a processing list in each one The spatial container idea The container’s composition, three threads A sense of geometry Behavioral cues Sensory cues Japanese cafés generally have rather small spaces, plus high turnover of people, plus a social script What exactly is the most fatal thing about a narrow space? I think there are two things. One is micro-social vigilance: the tables are very close, and the movements of the person next to you affect you The other is visual working-memory noise: a short visual range means high information density, which bypasses the brain’s filtering cost A space that is tall, open, and strongly receding makes it easier for people to enter a “self-narrative” state Thoughts are no longer dragged around by the outside world, but start to spread out along your own main line No need to write that users need to manage their emotions better Write it as: “23:40 at night, I can’t stop scrolling short videos, I have to get up early tomorrow, my finger swipes automatically, I curse myself in my head but keep going.” A stopwatch: find the cycle period Almost any pain point that can be turned into a product has a cycle: every day / every week / every time you go out / every time after socializing What you have to find is: how often does this pain point occur? How long does it last each time? Frequency * duration = the product’s “survival oxygen” There actually aren’t that many common productizable motifs: attention, decisions, procrastination, anxiety, relationships, health, money, a sense of meaning, learning A very important thing about being scenario-based is, in real society under what circumstances, what kind of thing should be done and having the features carry out a series of validation operations around that thing CapWords has another obvious advantage: the user is very clear about how to use it, and in what scenario to use it. Is this a very clear positioning of the AI’s boundary and the user’s scenario? I think this is also very hard Actively setting a boundary for the AI The moment a user opens CapWords, they already know three things: First, I’m going to “photograph an object,” not “generate a world” Second, I’m going to “name this thing / learn the word,” not “let the AI think for me” Third, what I get is a reusable cognitive unit (a sticker + a word), not a one-off output What does this mean? It means the user has already completed 80% of the comprehension cost before operating anything I’m wondering how important this is for the user, or for the developer of the product? A lot of products treat the model’s capability as the product’s capability I think there’s a problem with this. What problem? I call it a symptom You can ask me anything? I can recognize whatever you photograph … I can write, draw, summarize, plan, analyze The result is that the user doesn’t know where to start, or when it counts as finished … I only care about one “specific object” in the real world In a very short window of time Help you complete one clear cognitive act So even if CapWords recognizes something wrong the user can see it at a glance — the mistake is that this thing isn’t this word, not that the AI rambled on The cost of correction is very low: photograph it again It lowers the user’s psychological defenses against AI Once AI is compressed into a clear boundary, it actually becomes more trustworthy, more usable, more everyday I’m reminded of Ryoan-ji: fifteen stones, and from whatever angle you can only see fourteen, with deliberate empty space In 2026, as we disenchant ourselves from AI, we will realize: no matter how smart AI is, software development without engineering constraints will in the end turn into a pile of unmaintainable digital garbage. So starting now to train the ability to “systematically describe requirements” and “build an automated validation loop” is the key to meeting this shift A few processes of the shift, why it is Vibe Engineering No longer simple prompt Engineering, but building a system that makes AI output high-quality code stably, controllably, and reproducibly In the Vibe Coding phase, you are chatting with the AI. In the Vibe Engineering phase, you are writing a high-precision natural-language specification (Spec) The core of Vibe Engineering: you need to write automated tests (Evals) to verify whether the AI’s output matches your “Vibe” Define the Vibe (for example: “the UI must be Apple-minimal, and the response time < 200ms”). Write assertions or use visual regression testing tools. Let the AI generate code. Run the tests automatically. If it fails, the system automatically fixes the Prompt, instead of a human fixing the code Context architect In 2026, although the Context Window is very large, it is still expensive and prone to hallucination How do you stuff a project of 100,000 lines of code into the AI’s head and still keep it focused? Developers need to design a “context retrieval strategy.” When should the AI be shown the database Schema? When should it be shown the frontend component library docs? This needs an engineered RAG (retrieval-augmented generation) configuration, not manual copy-paste The core of vibe engineering is system architecture, module decoupling, and data flow It’s more like being a product manager and a technical director A lot of what AI does is useless operations AI is good at adding a few extra layers of useless abstraction, adding a few unnecessary Helpers Enforce the principle of minimal necessity Engineering methods remain unchanged; AI is just a full-process accelerator Context engineering + the principle of minimum necessity And some better developer experience: clear requirements, complete tests, documentation that grows along the way Architecture must be subtraction, so that the system doesn’t accumulate technical debt Flighty and the story behind it Flighty may be the most thoughtful companion a traveler has over the whole journey. The app was developed by Ryan Jones in 2019; Jones currently lives in Austin, and he says: “Travel can involve a lot of stress, and we hoped Flighty would step up and let travelers enjoy easy, relaxed service” When a long flight delay hits, Flighty shows the key information in the center of the front of the screen, through a simple and intuitive interface, a live map, and a look that evokes traditional airport design A simple, intuitive interface, a live map, and a look that evokes traditional airport design, showing the key information in the center of the front of the screen This first-rate travel app combines flight tracking, airport navigation, concierge, and many other practical features, and with its excellent “Live Activities” and “Dynamic Island” implementations, it becomes a thoughtful companion travelers carry with them, providing key information at any time Flighty supports display as a widget on the home screen and lock screen, uses “Shared with You” to highlight content, and so on. With just a few taps, travelers can share their flight route and arrival time in real time with family and friends who haven’t even installed the app, which brings great convenience to coordinating airport pickups and drop-offs “Whenever a traveler takes off, we have to assume they won’t be online before the plane lands.” How to solve that problem? At a certain moment before the flight takes off, the “Dynamic Island” switches to a flight progress bar and counter, shown in minimal form in a simple circular chart tracking the flight’s duration The Kindle App’s Ask This Book The Kindle App’s Ask This Book feature is an AI reading assistant that Amazon officially launched around December 15, 2025. It is currently limited to the iOS version of the Kindle App (iPhone/iPad) in the US, and supports thousands of English-language bestsellers. It is planned to expand in 2026 to the Android app and Kindle e-reader devices; other regions and languages have no specific timetable yet While reading, highlight a passage of text directly and an “Ask” option appears in the pop-up menu After tapping it, AI-suggested questions appear (for example “Who is this character?” or “What happened earlier?”), or you can type in your own question It opens a dedicated chat interface where you can keep asking follow-up questions You can also go directly to Ask This Book from the book’s menu This feature is essentially a dedicated AI chatbot based on the book’s content; it can instantly answer questions about the book you’re reading, including: plot details character relationships (for example, “What is this person’s relationship to the protagonist?”) thematic elements (for example, “What does this scene symbolize?”) other contextual explanations (such as character motivation, background events) Some distinctive points: A strict spoiler-free mechanism: this is the most praised point. The AI only generates answers based on the pages you have already read, and will not reveal later plot. Even if you ask “what happens to this character later,” it will refuse or only speak about the parts you’ve read. This perfectly solves the pain point in reading a complex novel of “I forgot who is who but I’m afraid to search online for fear of spoilers” Not interrupting the reading flow: everything happens inside the app, no need to switch to a browser or search engine, preserving immersion Precise context: the answers quote or are based directly on the book’s text, and it feels like having an “expert reading companion” beside you explaining Supports continuous conversation: you can ask follow-up questions like in a chat, deepening your understanding Answer protection: generated replies cannot be shared or copied, and can only be viewed by readers who bought/borrowed the book Water adapts to the shape of the container and thereby takes on its shape Zhongshuge Dreamy and magical, immersive space aesthetics; the core idea is “a chain that does not copy, one design per store” Each store deeply integrates local culture, humanity and geography, creating a unique visual impact and emotional resonance Through the three elements of beautiful environment, beautiful books and beautiful service, it gives reading a poetic sense of ritual, letting physical bookstores regain their appeal in the e-commerce era It makes heavy use of mirrored ceilings, floors or walls to create the illusion of being “surrounded by a sea of books” and “infinite bookshelves,” making people feel they are inside an ocean of books or a dream Things like giant spiral bookshelves, terraced reading areas, rainbow arches and staircase book mountains, combined with lighting, color and curves, produce a strong urge to take photos and a sense of immersion Fusion with local culture turns it into a cultural landmark The designers are mostly the X-living (X+Living) team led by Li Xiang; she is hailed as a “design genius,” with a style hard to define as any single school — it is full of imagination, romance and respect for books: “like writing a romance novel, no fixed style, only love for books” Li Xiang: imagine boldly, refuse repetition, tell stories with heart Hailed by the industry and media as a “design genius” and “Xiang Ye” She is world-famous for a bold, dreamy, surreal design style, and especially through the Zhongshuge series of bookstores she opened a new era of “immersive aesthetics” in Chinese commercial space. Her works are not just functional spaces but emotional carriers and cultural landmarks, perfectly fusing artistry, commercial value and practical function, and hard to define as a single school — some say they are like a real-life “Inception,” others describe them as “the romanticism of writing a romance novel” In 2011 (she was about 25) she founded X+Living in Shanghai; the name “Weixiang” means “the only ideal,” crossing over from architecture to interior design Li Xiang’s style refuses a single label; its core is “the perfect fusion of artistic aesthetics and practical function,” stressing that “design creates value” — not just beauty, but driving commercial success and emotional resonance Zhongshuge at Shenzhen Bay’s Happy Harbor is a visual peak of her work: a giant red spiral bookshelf (symbolizing Shenzhen speed and a vortex of time), combined with mirrored ocean elements, bursting with surreal techno-romance, and itself becoming part of Shenzhen’s landmarks I’ve been using Reminders all this time, but I feel that in some situations Reminders isn’t enough For example situations where I need to set priorities, important-urgent and four-quadrant scenarios; down to exactly how to act, TickTick can help me actively manage the task system More like an externalized workbench for the brain Reminders = trigger-based, one-off, life errands TickTick = things that need planning, can be deferred, work and projects What’s needed is a concrete way of planning that gives my future life a really good guide A deep understanding of Hong Kong FPS FPS is essentially the faster payment system The logic of local transfers in Hong Kong generally involves some methods FPS normally charges no handling fee and transfers around the clock; it’s the most commonly used way to transfer between banks There is also a settlement system called RTGS/CHATS, the traditional way banks clear between each other; this is usually used for larger or specific types of remittances, but there is a fee at the counter The essence of FPS: a payment highway that lets you do 24/7 real-time transfers and payments between bank ↔ bank and bank ↔ e-wallet (SVF). You can use a mobile number / email (and other identifiers in some scenarios) to send money to someone, and the money arrives almost instantly Its key point isn’t “a feature of some particular bank,” but a shared underlying rail for all of Hong Kong: banks and wallets all connect to it, and everyone runs on the same set of “addressing + clearing and settlement” rules FPS isn’t only for personal transfers It also includes topping up e-wallets, and transfers between wallets It also includes paying for online shopping, for example payments to merchants (especially FPS QR), and it supports local real-time payment in both HKD and RMB (within the Hong Kong system) Some apps have TUN TUN is a kind of virtual network card that lets a program pretend to be the operating system’s network exit Taking over at the network architecture level Normally, when there is no TUN: App → system network stack → direct connection ↑ (only apps that “have a proxy set” turn into Clash) Browsers: ✅ they respect the system proxy curl / git / steam / docker / games: ❌ they often ignore the system proxy UDP: ❌ basically hopeless DNS: ❌ often resolved first by the system Sometimes it’s just that permissions at the system level are too low With TUN, the traffic path is different App ↓ virtual network card (TUN) ←—— Clash ↓ real network card ↓ Internet The operating system points the “default exit” at Clash What does TUN intercept? It works at the IP layer The logic of what Apple links It essentially maps a token, a device account number This token is used only in Apple Pay transaction scenarios Behind it, the card-issuing country, BIN and network are all still there It’s not the same as PayPal; if Apple Pay is like swiping a card Apple hasn’t changed the identity of the money itself, it has only changed your way of paying PayPal directly changes the identity of the money in the merchant’s eyes It “takes your money into its own system” first, then pays out externally For many merchants: The payer = PayPal The risk-control subject = a PayPal account The legal subject = PayPal (a US / EU entity) What banks do at the most core level is maintain your fiat currency account PayPal is like taking a bullet for the merchant on one layer A regulated payment company standing between you and the merchant Its core value isn’t “moving money,” it’s: doing part of the compliance on behalf of the merchant hiding the details of your bank card for you doing risk control via account profiling instead of card profiling Wise is more direct: a “global local account” wearing the cloak of fintech It doesn’t touch payment disputes, it doesn’t act as a judge for merchants, it just “moves” money fast and cheap Its core capabilities are: multi-currency accounts (local bank numbers) real exchange rates (close to the mid-market rate) stitching together local clearing networks Technically speaking: the USD / EUR / GBP you see in Wise are essentially custodial accounts at banks in various countries Wise is just the ledger coordinator Wise doesn’t really want to know: what you bought whether you got into an argument with the merchant It only cares about how the money is doing and where the money is going Why does the state tolerate Wise but stay wary of stablecoins? The state tolerates Wise because it still operates “inside the national ledger” The state is wary of stablecoins because they try to become “the ledger itself” Every dollar in Wise ultimately corresponds to: a bank account in some country some regulated custodian bank some legal entity that can be seized Stablecoins are the scary part Stablecoins are “issuing shadow dollars” not on the Federal Reserve’s ledger not on a commercial bank’s ledger not directly controlled by a central bank This is completely out of control Why do all payment tools want to become “accounts”? Let’s observe: which actually matters more? What is a payment tool? A bank card, swiping a card, scan-code payment — its fate is low margin but fierce competition What is an account? An ongoing bookkeeping relationship Money stays for the long term Behavior gets recorded Rules can be layered on top Future behavior becomes predictable Payment is an action, but an account is power Why must the state eventually take back “account power,” and hence the appearance of CBDCs? Seems like WeChat Pay and Alipay are the same Because “account power” = “bookkeeping power” = “governance power,” and the definition of a state is itself: the ultimate bookkeeper. CBDC (Central Bank Digital Currency) = directly issued by the central bank, existing in the form of digital accounts or digital certificates, fiat money I keep wondering: an online card should be enough, so why do I still need a physical card? And doesn’t Apple Pay let you swipe offline overseas? The physical card = the “activation key” for the complete card lifecycle Many banks’ logic is: Virtual card: limits controlled functions conservative risk control After the physical card is activated: full permissions unlocked overseas / online opened up lower false-decline rate The physical card = the only entry point to ATM / LINK / EPS ATM cash withdrawal ATM verification some POS terminals that only accept physical cards some bank branch operations And in fact, if you need stable long-term subscriptions and high-value charges, the physical card actually has a higher success rate So the physical card solves not just usability, but more than anything the trust problem By comparison, I think using Apple Pay overseas is very convenient Apple Pay and the physical card ——both swipe the same card ——both charge the same account ——it looks like “no difference” But actually, in the system they are two completely different transaction forms Apple Pay isn’t “another way of swiping a card,” it’s “another transaction model” From a risk perspective Physical card: the bank sees a lot of risk, namely whether this is reliable, whether the card has been copied Apple Pay NFC: what the bank sees is a token (not the real card number) + a one-time dynamic key, so it has already been verified and carries lower risk It’s the same for merchants: with a physical card transaction, chargeback risk lies with the merchant, but with Apple Pay → part of the chargeback risk is borne by the card network / Apple Wise’s main advantage I keep thinking: banks can do currency exchange too, and the rates are decent When you exchange HKD for USD inside BOCHK, what usually happens is: your money is still within the same bank’s account system it just moves from the HKD sub-account to the USD sub-account (or currency exchange within a multi-currency savings account) you get the exchange rate the bank gives you (including the spread/markup) This step is essentially: a currency conversion inside the same set of ledgers If afterwards you just want to hold dollars, or use dollars for investment or deposits, exchanging inside the bank makes perfect sense But what is Wise for? Wise’s exchange is completely transparent; banks may differ, because every bank, every time period, and every customer tier is different How you use the money after exchanging is the fundamental question After exchanging to USD at a bank, your dollars are usually: very suitable for: keeping them, wealth management, buying USD products not necessarily suitable for: transferring dollars to an overseas recipient at low cost (you may need SWIFT, there may be intermediary bank fees, it’s slow) Wise’s strength is: you send HKD into Wise via FPS (locally in Hong Kong) Wise converts it to USD internally then sends the USD out through a local US channel (such as ACH) This feels more “like a local transfer” than a traditional wire route: less friction, more predictable fees Wise gives you a “US account number / European IBAN” etc., which essentially makes you look like a “local recipient” in the other person’s country Link and Linkpay are not the same Linkpay serves Hong Kong Link is a one-click checkout tool launched by Stripe, similar to a digital wallet, letting users securely store bank cards, bank accounts and shipping information, and check out quickly on any website that supports Stripe payments Link is cross-platform; for now it doesn’t support offline payment, but it supports online payment on a large scale Merchant data is open too, so merchants can easily get the customer’s email information But the merchant must have enabled acceptance of “China UnionPay” in their Stripe dashboard before you can pay with a UnionPay card Designing the brain’s cognitive system The goal isn’t to be able to do anything anywhere by willpower The goal is to let different scenarios each allow only one main mode to appear Don’t fight yourself inside the same scenario For example, my WeChat: I don’t have WeChat on my computer, because I want the computer environment to be pure and want to enter a focused state without being disturbed, but I often get the urge to reach for my phone The brain knows: the phone entrance is still there The real problem is that there’s an implicit rule, and this invisible rule is: better not to use WeChat So what exactly is the invisible loophole in cognition and behavior? Partial extinction So often it’s this very thought of mine that breaks the brain’s cognitive mapping association in the cafe scenario Linear’s unique advantage What Linear solves isn’t whether you’re doing something It’s how a software team keeps clarity, speed, and low friction all the way through a complex, long-term, uncertain development process issues are first-class citizens, including bugs, new features, tech stack The state flow is extremely smooth; keyboard shortcuts cover everything A project isn’t a spreadsheet but a timeline; software development is a probability problem, not a construction blueprint Linear’s states usually feel like this Backlog → Todo → In Progress → In Review → Done If a team values “writing code” more than “writing process,” it’ll probably like Linear GitHub Projects’ center of the universe is: the code repository Projects is an “attached structure” that grew up around GitHub Issues / PRs A product where AI teaches you how to have a conversation Helping you chat Note: chatting in real life I think this is really interesting A tool that helps people chat and ask questions Users can start by entering a topic into our product, and the product can generate some questions Please give an overview of all the dimensions of my request, identify the uncertain points, then recommend as many questions as possible that I could ask (leaning conversational) A product like this fits very well with the proto-forms of future AI smart glasses For example Ray-Ban Meta Display, which actually has a display, as a flagship device Deeply tied to Meta AI and supporting limited AI-agent-style capabilities, suited to quickly getting visualized information while out and about Deeply tied to Meta AI, voice-activated to generate visualized results (like navigation routes, restaurant recommendation maps, translated text, AI answers in text and images), supporting real-time captions / translation, AI image search, and so on An AI product for great shots, made for hiking AI combines your own photos with landscape shots Generates the best check-in photo of your life, and can record the information behind every check-in: the address, the story behind it As a travel museum of your life Recommending different tasks and suggestions to a person in different environments I think this is really important Even very creative I think it might even be the prototype of the next generation of todo tools Spot the problem -> provide the tool -> solve the problem When general-purpose capabilities at the technical foundation tend toward saturation product competition really does turn into “competition over meaning.” Scenario-based products are no longer just containers of features, but should each be a “specific ritual with a soul” Suddenly I want to build a movie app I want AI to recommend movies for me AI can help me log movies and organize some of the good terms from them And I want no spoilers before I watch After watching I can add to it and summarize It can flexibly take a selfie and generate a photo of me with one of the characters from the movie Filter out bad films Squeeze the nutrition out of good films and eat it all And in the end let you keep the movie in your memory in an interesting way (like a photo together) The ZA card feels super useful Its positioning is just as a virtual card; in daily life you almost never need a physical card You get the card number, CVV, and expiry date right in the App, and can bind it directly to Apple Pay, Google Pay, Alipay, WeChat Pay, or spend at any online store that takes credit/debit cards Card management, transaction records, temporary freeze/unfreeze, limit adjustments — all fully doable in the App The physical card has no CVV/expiry, which lowers the risk of card theft; and if you lose the physical card it doesn’t affect online use At the same subscription price, a Web subscription ≈ earning real money, an iOS in-app subscription ≈ getting skimmed by Apple + restricted by rules As for Web subscriptions, The problem RevenueCat solves The App Store / Google Play subscription system is far too user-hostile All Apple / Google give you is: low-level APIs raw receipts messy state callbacks tightly platform-bound logic What founders actually need is: ❓ Is the user a member right now? ❓ When does the membership expire? ❓ Does it still count if they switch devices? ❓ If bought on iOS, can it be used on Android? ❓ What are revenue, MRR, LTV? The platforms don’t help with any of this So RevenueCat appeared Proxy groups really are a good thing The essence of a proxy is handing traffic to the line currently selected manually The essence of a proxy group is handing it first to a decider, which then chooses among multiple lines The key is whether there’s a middle decision layer openai.com → PROXY goes through the manually selected node openai.com → AI-NonHK essentially passes through this group; the strategy used in the group can be a URL-Test strategy, selecting the current best node Proxy group = an object of “node set + selection algorithm + state memory” The node set is a set of nodes you control; the selection algorithm generally has a strategy, which can be url-test sending test requests to each node periodically and choosing the best State memory is who was best last time, and whether a switch is neededPhotos as records and storage for curiosity
CapWords’ minimalist privacy strategy
CapWords’ embodied-memory layout
The ultimate experience of AI tools
Data science’s cross-disciplinary fusion
Core capabilities of data analysis
The four-dimensional model of materials science
The logical density of materials science
The life philosophy of photographing as collecting
AI Products: From App-Centric to Intent-Centric
The Essential Product Difference Between Intent and Entertainment
Product Positioning and Judging Your Ecological Niche
The Discrete-Choice Trap of System Interaction
Methods for Observing and Iterating on Products
How Environment Triggers People’s Habits
An Experiment Fusing Buddhism with Product Creation
The Geometry and Behavioral Design of Spatial Containers
Describing Pain Points Through Scenarios
The Validation Logic of Scenario-Based Products
CapWords’ Design of a Clear Sense of Boundary
From Vibe Coding to Vibe Engineering
The Principle of Minimal Engineering in the AI Era
The Story Behind the Flighty Travel App
Reading the Kindle App’s Ask This Book Feature
The Daoist Wisdom of Water Adapting to Its Container
Zhongshuge’s Immersive Aesthetic Space Design Philosophy
Designer Li Xiang’s Philosophy of Bold Creation
Comparing TickTick and Reminders for Task Management
A Deep Dive into Hong Kong’s FPS Faster Payment System
TUN Virtual Network Cards Taking Over at the System Level
The Mapping Logic of Apple Pay and PayPal
Core Value Differences Between Banks, PayPal and Wise
Why the State Tolerates Wise but Is Wary of Stablecoins
Why Payment Tools Chase Account-ification, and Its Nature as Power
Why the State Must Eventually Take Back Account Power, and the CBDC Logic
The Physical Card as the Activation Key to Full Trust and Permissions
Transaction Model Differences Between Apple Pay and Physical Cards
Wise’s Core Advantages in Currency Exchange Over Banks
Differences Between Link and Linkpay Payment Services
Designing Scenario-Based Cognitive Systems, Not Willpower
Linear Serves Uncertain Development with Clear Process
An AI Product That Helps You Converse
An AI Hiking Check-In Museum
Environment-Aware Task Recommendations
Competition Shifts from Tools to Meaning
An Idea for an AI Movie-Journal App
Filter Out the Bad Films, Eat the Good Ones Dry
Using the ZA Virtual Card
The Difference Between Web Subscriptions and iOS In-App Purchase
RevenueCat Solves the Core Pain of App Subscriptions
Proxy Groups Are the Middle Layer of Traffic Decisions
4. AI and Agent Systems
48 entries AI is like a mirror what it reflects back is yourself We are human, we have lots of flaws, but we also have lots of precious treasures belonging only to us ,,, We get sad, we have emotions, we empathise with others these are precious things about being human itself … even emotions — that too is human sediment Can ordinary people’s lives be interesting too? I think yes, maybe yes Every ordinary person has their own identity, their own role AI can recommend some experiences, and the user goes and completes that part of the experience OpenAI Realtime API The breakthrough of GPT-4o’s native S2S architecture: the Realtime API is based on the GPT-4o model family and adopts a native speech-to-speech architecture Before, it was linear: first automatic speech recognition, then the model transcribes the user’s audio stream into text, then the large model takes the text input for semantic understanding and generates a text reply, and finally a TTS engine synthesises the text into audio CapWords provides a textbook-level case of “Humane AI”: that technology should be warm, tactile, and able to reawaken the user’s curiosity about the world around them And it uses Apple’s Native Frameworks to build a privacy moat Language learning should feel natural, part of daily life — warm, not cold or mechanical Stickers: a sticker is a kind of marking and possession Realtime API strategy analysis The native transcription experience brings a qualitative leap Ultra-low latency Emotional intelligence Full-duplex streaming The first two are easy to understand the third is very interesting — truly full-duplex communication allowing the user to cut in and interrupt while the model is speaking; this is barge-in The system perceives it in time and truncates the output, simulating interruption in human conversation The Realtime API’s interaction mode shifts from the traditional RESTful request-response mode to a stateful session mode based on persistent connections Implementation options: WebSockets and WebRTC Setting aside TTS and STT what’s popular now is basically the real-time voice interaction agent system (Voice Agent Runtime) internally containing STT and TTS as well as an LLM an end-to-end real-time voice AI assistant engine The form I want is people returning to reality, rather than being bounced back and forth between apps I think a good starting point is utility apps AI assisting people to complete some app operations This actually gives the choice back to the person themselves, rather than algorithm recommendations consuming the user’s time So for a product, being able to clearly find your own core positioning entry point every time is especially important By deeply dissecting its technical implementation, we can see a clear new trend in mobile development: Middle-Layer Innovation. Developers no longer need to become AI scientists, but need to become AI Orchestrators — proficient in native frameworks like VisionKit, Metal, CloudKit, using refined Prompt Engineering to command cloud intelligence, and finally using ultimate interaction design (Foley sound effects, skeuomorphic animation) to wrap cold data into a warm product CapWords’ technical moat lies not in the algorithm itself, but in its extreme exploitation of the iOS ecosystem’s capabilities and its precise technical translation of users’ cognitive psychology Some downsides of Rork The advantages are clear: you can iterate out an app demo very fast, based on Expo and React Native It supports cross-platform, one-click generated iOS and Android versions, plus browser preview and QR-code scan testing, and publishing to the App Store / Google Play is simplified too But Rork is essentially an AI wrapper around RN, and native iOS features are still hard to achieve A native app can make full use of iOS hardware acceleration, animation, and gestures (like precise 120Hz scrolling) Including the ability to handle images and camera capture Plus incredibly smooth interaction logic, and debugging and hot reload That silky control — suddenly I can understand why the big companies’ apps are all native Deductive reasoning The preconditions that guarantee a true and valid result Besides the initial argument being true That the judgment about common sense is correct The scenario of reinforcement learning is the current capability boundary of LLMs The areas where LLMs give an ultra-high-multiple experience have already been mined out In super-large scenarios, they’ve already been mined out The next scenario may be a change of mindset: using AI to do a simple scenario But there are also some scenarios that are just LLM scenarios. I’m wondering what these scenarios are; I think it’s the chat scenario Whether the reading scenario really has anything new in it AI-assisted analysis Still the chat mode, but it can do some upfront analysis Some preconditions In the AI era, your own cognitive model can also congeal false and wrong information to the point of producing preconceptions, even prejudice People like that … The value of reinforcement learning fine-tuning Reinforcement learning fine-tuning from human feedback — actually many people feel RLHF has already reached a certain limit, and that seems to me a misreading too When we talk about technology, we have to talk about what the technology can do RLHF combines reinforcement learning with preference feedback given by humans, so that a large model not only can talk, but can be more “close to human preferences and values.” In many generative tasks this is something pure supervised learning cannot replace But why it is limited: training cost is very high, and RLHF requires collecting a large amount of high-quality human preference data It depends on the subjectivity of human feedback: different people understand a “good answer” differently, which means feedback on the same task may contradict itself, causing the reward model’s quality and behavioral consistency to be unstable. Bias can also be amplified through feedback Optimization instability: reinforcement learning itself is a “long-standing hard problem” in AI; the training process is sensitive to hyperparameters and easily unstable, and RLHF merely moves this hard problem into the field of large models But RLHF and its related variants are still pushing AI closer to human needs — the AI alignment problem How to make an agent’s behavior match human values, preferences, and ethical standards — this is not something supervised learning can easily achieve, and RLHF is one of the most practical routes available right now OpenAI asking and answering its own question ChatGPT’s ears and its mouth seem to share the same air channel, so it may be able to hear itself talking In iOS, audio isn’t opened however you like; it has explicit Audio Session modes: Playback mode: only plays sound, doesn’t listen Record mode: only listens, doesn’t play PlayAndRecord mode: listens and plays at the same time (phone calls, voice assistants, real-time voice AI) ChatGPT’s voice conversation has to be using PlayAndRecord But if you’re wearing headphones, the sound travels through the headphones Output → headphones Input → microphone The two are physically isolated An analysis of OpenAI’s voice mode The alignment tax and safety castration, one being over-defense: to stop the model from being abused (scams, porn, copyright infringement), OpenAI applied an extremely strict RLHF (reinforcement learning from human feedback) layer. To make sure it was “safe,” the model was trained to be overly cautious Older models felt like they would pick up a joke and run with it, judging whether the sentence was sensitive The inference cost of the voice modality is far higher than that of pure text As the user base exploded, to keep the service from collapsing OpenAI may have done “dynamic downgrading” in the background When load is high, the system may quietly switch to smaller, more heavily quantized model versions (Distilled Models) The inherent limitations of an end-to-end architecture; native audio end-to-end is both an advantage and a shortcoming In pure text models, hallucinations are usually factual errors. But in audio models, hallucinations show up as strange laughter, abrupt gasping, or ill-timed emotional outbursts Diminishing marginal utility plus the uncanny valley: the first time you use it, it can talk and sound human; you get used to the fluency, but then you start paying attention to whether it can solve concrete logical problems — the need shifts from “fun” to “useful” The flaws of end-to-end voice architecture Limited by some of the capabilities of end-to-end itself Hallucinations are usually factual errors; in voice models, hallucinations show up as strange laughter, abrupt gasping, or ill-timed emotional outbursts The conflict of voice models Because right now no single model can be o1-level smart and have GPT-4o’s gift of gab at the same time If you want it to talk fast (low latency), the model has to be small and the thinking has to be shallow If you want deep thinking (strong reasoning), the model has to be big and the response has to be slow (for example the o1 series needs a dozen-plus seconds to think) A three-way conflict: today’s models are still essentially serial, they have to finish thinking first Maybe imitating human slow thinking and fast thinking can get us there Cerebellum (Edge/Fast Model): handles speech synthesis, tone control, simple small talk (Latency < 300ms). It’s like the “foil” in a comedy duo, filling the gaps. Cerebrum (Cloud/Deep Model): when a complex intent is detected, quietly spin up an o1-level reasoning model in the background (Latency > 5s) The most popular voice AI architecture in Silicon Valley: Layer one (hearing): use ultra-fast STT (speech-to-text) such as Deepgram. Layer two (router - the key part): a very small model (like GPT-3.5 or a specially fine-tuned Mistral 7B) judges the user’s intent within 100ms: Is it small talk? -> hand it to the fast model (GPT-4o mini) -> instant reply. Is it a complex task? -> hand it to the slow model (Claude 3.5 Sonnet / o1). Layer three (latency masking): if it routes to the slow model, immediately trigger a “filler-sound mechanism.” The app plays a preset thinking sound effect (like keyboard tapping) or has the AI say “Give me a sec, thinking…” Humanity is only one slice in the long river of the cosmos, fleeting Slow the clock down and we appear, flourish, argue, write poetry, build satellites, and then quietly exit. The cosmos won’t even raise an eyebrow Stars aren’t aware of nuclear fusion, black holes don’t know they’re cool, but humans ask “who am I,” “where did the universe come from,” “why will I die.” This is a phenomenon that only appears once information processing reaches a certain level of complexity — sudden, like a phase transition Absolute optimism in engineering, absolute pessimism in sociology Many people believe AI is the natural evolution of human civilization, even the handover from carbon-based to silicon-based life. From this perspective, traditional “human welfare” is placed after the grander “evolution of civilization” From practitioners’ own point of view, everyone is optimistic; it brings individuals enormous returns and a sense of achievement The judgment about society: only about 17% of experts believe AI will have an absolutely positive impact on society as a whole in the next 20 years (pessimistic) Many senior practitioners have started to reflect: when AI can write better code, paint more beautiful paintings, and even provide more perfect emotional value, where is humanity’s “uniqueness”? Optimists believe humans will be upgraded to “architects,” while pessimists believe most people will lose their drive and sink into a kind of digital “nihilism” Metacognitive learning What is the value of learning? Most fundamentally, it’s the process of learning itself Learning driven by interest, dominated by intrinsic motivation, makes the brain treat “understanding” itself as the reward, rather than treating exams, certificates and returns as the external sugar coating. Here dopamine isn’t because “I finished the task,” but because “I understood a structure.” The long-term effects of the two are very different The phrase “learning how to learn” is actually profound. It’s not chicken soup, it’s meta-learning: You aren’t just stuffing knowledge into your head, you’re observing — which ways get me into flow? which paths let me truly understand rather than merely remember? which concepts, once grasped, make many fields light up automatically? Knowledge is no longer an island; it starts explaining itself. Physics helps you understand philosophy, programming helps you understand systems, psychology helps you understand why you get stuck More interestingly, this way of learning changes a person’s posture toward the world Not “what should I learn so I don’t fall behind” but “how does this world operate, and where can I cut in” Most learning tools revolve around what I learned But high-quality output inevitably revolves around what problem I am solving What is the problem you’re truly confused about right now? Is it the same problem you’ve kept running into lately? Which fields of learning are pointing at the same tension point? When a problem is continuously illuminated, output happens naturally Papers, products, articles — all of them are essentially “stage solutions to a problem” Apple Park A restrained aesthetic Glass, white, nature, clean, almost to the point of coldness It is the physical embodiment of long-termism. Apple Park’s construction timeline, its cost, its obsession with detail The design involved Steve Jobs himself Jobs insisted on putting the mailboxes, toilets and coffee machines in one place, forcing different departments to “bump into each other.” He later stated this logic explicitly: architecture is for “choreographing encounters” Steve Jobs (Walter Isaacson) The book’s description of Apple Park is very blunt: This is Jobs’s last product. Not “like a product” — he really did architecture the way he did products: extreme control, repeated polishing, no compromise, serving the next decade rather than the present. The long-termism, restrained aesthetic and obsession with detail you see can basically all find their intellectual prototype in this book Apple Park didn’t appear out of thin air; it continues the experience of a 20th-century scientific community: when you isolate smart people into little cubicles, output drops when you let them share space, share nature, share informal time, breakthroughs increase Jobs simply pushed this experience to the limit of aesthetics and engineering NotebookLM is super well suited to lawyers Cognitive certainty. No improvisation allowed whatsoever NotebookLM won’t answer questions beyond the scope of the resources you’ve uploaded. Its job is to build a private micro-world — a “closed-resource information trust” — and to reason only inside that scope This is the first AI lawyers trust, because it won’t act on its own or make things up NotebookLM’s point isn’t creativity but controllable cognition AI can bring a cognitive dividend, and it can also bring cognitive debt Essentially, when it works together with experts and supplements your own professional knowledge, it can save time and improve judgment, producing a cognitive dividend However, when it’s used as a shortcut — automating tasks, shrinking headcount — it weakens workers’ capabilities and fosters false confidence, creating serious cognitive debt Think of how we use phones in the same way — technology itself is a double-edged sword. Some people use phones as a way to become emotionally dependent and escape reality; some use phones to manage and expand their experience of reality. The pros and cons of using AI depend on the person using it At this stage AI actually has a strong verification cost, though existing products like NotebookLM’s traceability can solve that well AI has actually created a kind of chosen wasteland: in the past we wanted to find one answer, now we have to pick one of ten perfect options Some people learn a lot of contemporary physics or other fields through TikTok; for others it’s just a way to burn time When there are questions where even you aren’t sure what you’re asking, a really good way is to use an AI prompt to pin down your own need: Please give an overview of all the dimensions of my request, identify the uncertain points, then ask me as many clarifying questions as possible For women, AI is more easily turned into harm and a weapon rather than efficiency Deepfakes mainly target women AI tools of the “undress” / “nudify” kind specifically target women AI is used to automate harassment, mobbing, and humiliation of women Training data is itself a slice of a biased society: Hiring AI prefers men AI suggests lower salaries for women Medical AI more easily misses diagnoses in women Judicial AI overestimates women’s recidivism risk Generative AI maps “nurse/secretary” = woman, “professor/doctor” = man And anyway it’s a system designed by a small number of people; it can’t naturally serve everyone The usefulness of the useless structures the usefulness of the useful Art is useless Tools are useful Product art is a mismatch of name and reality between extreme rationality and extreme sensibility Why must an AI assistant be an “efficient secretary”? Why can’t it be a “poet who occasionally lies” or a “drinking buddy who only shows up late at night” Why must AI always please and go along with people? What is an AI that goes along with you everywhere actually like? Scientists & Engineers The scientist’s core is exploring the unknown and revealing laws — essentially a discoverer The engineer’s core is applying laws and solving problems — essentially a creator A data scientist naturally analyzes the patterns in the data and explains what the data is saying The engineer emphasizes preferences, correctness, what the output is, what the output is, with clear definitions The data scientist coexists with uncertainty; conclusions often don’t hold, and the data may be off too Netflix’s precise personal recommendations Recommending “films you might be interested in” based on viewing history Amazon Prime Video, beyond recommendations, extends into the X-Ray feature: on one hand it uses computer vision and celebrity-recognition models to identify the actors and characters in a film and display actor info in real time; on the other it launched “X-Ray Recaps,” spoiler-free summaries that use a large language model to automatically generate episode recaps Maimovie’s specialty is using AI semantic analysis (“AI Keytalk” technology) to extract a film’s tags and characteristics; users can interact with the app to select keywords, letting the AI generate a custom watchlist and find forgotten, little-known gems according to your unique taste For film and TV data, The Movie Database (TMDb) provides detailed information gathered for films worldwide, including plot data, descriptions, and cast lists There’s also Douban’s API, which takes some digging but can fetch movie entry details LLMs like ChatGPT can generate natural-language summaries such as film synopses and pre-viewing notes. People have already applied GPT to spoiler-free plot summaries, training the model to avoid leaking key plot points Modern viewers face content overload and lean toward smart filtering tools to find films they like; before watching they want a spoiler-free synopsis to decide whether it’s worth it, and after watching they crave deep analysis or interactive discussion — all of which give AI room to work First, register for and use the TMDb movie database API to get basic data, including film details, posters, cast lists, plot synopses, and so on. TMDb’s rich content and search function can meet most query needs, and it’s regularly updated with the latest films. For domestic films that local users prefer, you can supplement with the Douban movie API for ratings and review tags. Store the fetched film metadata in a database (for example MySQL/PostgreSQL), and design the table structure well (films, actors, genre tags, etc.) for later querying. User viewing records and preference data also need to be stored; you can build a user-film interaction table recording the user’s ratings, favorites, watch history, and so on. At first there’s no need to scrape video or subtitles yourself — making full use of the synopses and tags provided by open APIs is enough to launch a prototype; later you can optionally use public subtitle databases (like OpenSubtitles) for local analysis, but be careful about copyright compliance The difference between web and iOS subscriptions is huge Web takes a small cut User pays: $100 Payment channel (Stripe etc.): ≈ 2.9% + $0.3 You net: ≈ $96–97 Subscribing through iOS is rough User pays: $100 Apple’s cut: First year: 30% From the second year: 15% You net: First year: $70 After that: $85 iOS doesn’t just take money, it also constrains your business freedom ❌ You can’t: guide users in the App to subscribe on the Web explicitly say “Web is cheaper” use the Web price as a comparison freely do bundle / credits / usage-based pricing ✅ On the Web you can: discount annual payment Token / Credit usage-based billing enterprise edition private deployment quickly experiment with pricing So generally Web first Dual subscription: iOS is more expensive but simpler, web is cheaper with more features The App only sells the entrance, not the subscription Design a system » brute force I’ve said more than once: don’t fight it, design a system, go with the flow I saw Dan Koe’s method Put out one distilled piece of content every week, then design a good system to get many meals out of one fish The source is going all in on one long practical piece each week Then the distribution system breaks this mothership down into: YouTube scripts daily social media posts short-video spoken scripts How to benchmark against hits: first test an idea with a short post on Twitter Once the data is good, expand it into a long practical article Benchmarking against hits isn’t about reading the content — just borrow the topic and angle and rewrite it from your own perspective Use AI to “compress the research loop” (Compress the Research Loop). He never uses AI to “ghostwrite” (AI-written social copy is really “garbage”!) His SOP: when he needs to learn (say, watch a 3-hour video), he doesn’t take notes! He feeds the video/article/PDF to AI and has AI summarize it, pull out key points, and extract unique perspectives Reminders isn’t classified by scenario; it’s a clearer, more extensible list Of course it can generally be designed by scenario inbox by default Personal for life, personal Work for work Travel for travel … And it’s deeply integrated with iPhone I really favor Reminders more than Dida List In their December 2025 paper “Distributional AGI Safety,” DeepMind puts forward a striking hypothesis: for reasons of cost structure, AGI may not first appear as a “single monolithic superbrain,” but first as “Patchwork AGI” emerging as a “system state.” Its core hypothesis is: The future is more likely to have a large number of sub-AGI agents (each very specialized, none general-purpose) Through agent-to-agent communication, division of labor, tool calls, and incentive mechanisms general-purpose capability emerges in the network interaction So AGI is a state of affairs, not an entity One line in it reads: “Intelligence is not centrally expressed; it is an emergent property of interaction networks.” That is, the truly frightening or wonderful capabilities may come from “organizational structure,” not from “bigger parameters.” Like an ant colony or a bee swarm: take any single individual and it has no intelligence, but as a huge whole it displays abilities beyond the sum of the individuals But there’s one question: if it emerges gradually from the local level how does the local connect with the whole? Like geese flying? There should be a pattern to that too — so what is the pattern? Humans can’t even “imagine” the whole shape of a nonlinear structure; they can only explain local results after the fact Because — there’s no evolutionary reason for us to have that kind of premonition ability Think about it: even the most primitive people of the printing era knew this was a revolution, that books multiplied — but the real revolution with huge impact was knowledge detaching from the memory system, a pattern that affected humanity for thousands of years afterward People in the early Industrial Revolution only knew this revolution made them and productivity stronger, but couldn’t predict the modern corporation; people in the early Internet couldn’t imagine social networks either Our role now is more like standing inside the old structure and building a “temporary host shell” for a new species New structures can’t be designed in advance; they emerge naturally through the collisions, explorations, and uses of countless individuals Humans change the world with their brains Idea → neural activity → muscle / language / tools → change the world Brain-computer interfaces do three things Bypass the body’s obstacles, because paralyzed people can’t output through muscle Also, the traditional path is idea → language → hand → keyboard → system, while the BCI path is: idea → neural signal → system Human expression loses a lot: if an idea is a 100 in the brain, saying it might leave 60, writing it down might leave 40, being understood might leave 20 — can brain-computer interfaces solve this bandwidth loss? What RL faces right now RL actually needs to load several models at once — policy model, reference model, value model, and reward model. That’s not just massive compute and memory overhead; there are also a lot of bubbles (idle waiting time), and machine utilization is low Then there’s the delay and resource consumption of interactive exploration in real environments, which is absurd — you can simulate environments to cut costs, but that may create a gap with the real environment So it’s hard to scale; RL training curves are very fragile and prone to collapse. Even top AI teams can only use a very small amount of data for RL training RL is no longer a “core learning paradigm” but an “end-stage adjustment mechanism” RL’s historical misjudgment was: we once thought RL was “the core of intelligence” and what’s becoming clear now is: RL is more like “the tool for the last 5% of alignment and calibration” So RL is naturally suited to some scenarios where the setting, rewards, and environment are relatively well defined for example game-theoretic games, controlled simulation systems, energy consumption optimization The most common real-world use of RL right now is robots — the “year one of robots” that many people hype Compared with ordinary product apps, ordinary people feel physical things more strongly, and the barrier to getting hands-on is higher RoboBallet is a reinforcement learning system developed by UCL, DeepMind, and Intrinsic that lets multi-robot teams work together like a “dance troupe,” coordinating actions on industrial production lines and effectively improving efficiency and scalability AgiBot (a Shanghai robotics startup): combining human teleoperation with RL, letting dual-arm humanoid robots quickly learn complex assembly motions on real production lines — this human-machine hybrid training is an important bridge for landing in real factories RL’s most successful area so far is still autonomous driving: RL control policies are used in simulation and reality for vehicle decisions, path planning, and dynamic obstacle avoidance; it’s often mixed with classical planning and rule systems to lower sample complexity and improve safety The labor problem right now China has great expectations here Many of China’s basic jobs are gradually going unwanted Aging is too fast Robots are a technological hedge against an irreversible demographic change What China isn’t good at: operating systems consumer software ecosystems high-end chip design (in the short term) What China is very good at: engineering integration scale manufacturing cost compression going from 60 points to 90 points a direction where the path is clearly visible and there’s certain to be a result in ten years Listening to the people next to me chat feels so split On one hand they say AI is a bubble and feels like it’s passed; on the other they say next year is year one of robots, robots arriving — but isn’t the robot boom dependent on AI capabilities? Dididadada: join in and talk back I’m standing beside them, an I-person, afraid to speak They’re probably just interested in chatting, not in the content of the chat haha Otherwise they’d actually go look into it I’m reminded why the country has such high hopes for robots right now; it feels like robots really are a very favored direction in the A-share market For robots, it seems like large models and the intelligence layer aren’t the bottleneck anymore It’s more a problem of engineering capability Making a robot understand, plan, and converse — not hard The algorithms aren’t that hard either The real problems: one is the reliability of the electromechanical system — running correctly and reliably over time, and even low maintenance cost The other is the closed loop of perception -> decision -> action, where sensor noise, control delay, action error, and so on — can they reach millisecond level This is engineering philosophy, and it’s exactly what needs time to settle. China’s engineering capability is very strong Rewards from natural evolution decay after fulfillment, making you stop; Rewards from social evolution amplify after fulfillment, making you keep climbing. If food could really keep making people happy, an affluent society would be the happiest place But the opposite is true. The more an environment has no shortage of food, sex, and security, the more likely emptiness, burnout, and boredom appear. The reason is simple: the reward system produced by natural evolution isn’t meant to satisfy you long-term, it’s meant to make you act. Once a goal is fulfilled, the reward automatically decays What really exhausts a person isn’t standing still — it’s having started to climb upward Once you step onto the ladder of knowledge, money, power, or self-realization, the reward no longer comes from “arriving” but from “continuing.” The pleasure each step brings is extremely brief, and immediately turns into craving for a higher step — this is the reward function shaped by social evolution. It isn’t there to satisfy you, it’s there to keep pushing you But the latter is endless Behind the high-price buyout of Manus Manus is a Singapore AI startup founded in early 2025 by the Chinese founder Xiao Hong, focused on building general AI agents that can handle complex work like deep research and task execution, and in just 9 months went from 0 to a $125 million annual recurring revenue rate (ARR), with users in the millions Behind Meta’s acquisitions over the past year: Meta went through the collapse of the metaverse and actually hasn’t managed to take the lead in the AI phase Including the VR glasses it later launched, hoping to seize the advantage in the glasses space. On the body there are really only a few wearable devices: glasses, a watch/band, a ring, a necklace, earbuds — and for girls, additionally, a doll The essence of an acquisition is really buying control plus expected returns plus negotiating power The funding number is not the acquisition price Fundraising is “pricing a dream” Acquisition is “paying for reality” The Apple SHARP project Apple has been working on 3D for a long time The goal is to make an ordinary photo into a spatial memory you can slightly step into Think of the image effect you see on Vision Pro or iPhone What we shoot is all for the future 3D is very costly and unnecessary; most users just want to look at photos and remember, needing only a slight shift of viewpoint and 1–10 seconds of immersion The sense of reality matters a lot — not just a virtual world, but real content that’s richly spatial Spatial Scenes on iPhone can make any ordinary photo look like it has depth of field / parallax iOS 26 has a feature called Spatial Scenes that turns ordinary 2D photos in your album into “moving dimensionality” (essentially foreground/background layering + depth map + parallax). It doesn’t require the original photo to have depth information, so many “old photos” work too It isn’t really true 3D/stereoscopic display either, and it doesn’t need Apple Intelligence But clearly the iPhone has a spatial scene Spatial photography = recording “the feeling of you being there,” not just the recorded image anymore This is shot for future devices Manus: “Less Structure, More intelligence” You can see it’s already at the very bottom of the Manus official site Born for the minimum structure and more intelligence When your data is high-quality enough, your model smart enough, your architecture flexible enough, and your engineering solid enough, then concepts like Computer Use, Deep Research, and Coding Agent shift from product features into naturally emergent capabilities. Here are a few very core insights reasoned out from first principles: Only by freeing the user’s attention can you redefine DAU AI search isn’t recall from an index and then summarize; it’s letting AI obtain information with the user’s permissions Operating a GUI isn’t seizing control of the user’s device; it’s giving AI its own virtual machine Writing code isn’t the end goal but a general medium for solving all kinds of problems The hard part of generating a website isn’t building the framework, it’s making the content say something The “excess capability” problem of large models It’s really that capability growth has clearly outpaced real demand, usable scenarios, and the system’s capacity to carry it — producing a series of structural problems What an agent mainly does is thinking + execution AI extends the breadth and depth of thinking; people are responsible for the final judgment and choice Therefore the impact will hit first in desk-work domains that need heavy digitization My feed is full of takeaways, inspirations, evaluations, and criticisms of Manus being acquired by Meta For traffic, or really immersive thinking? It seems like a kind of emotion, stance, and self-projection It’s nothing but a phenomenon — just using your own model to explain this event, hoping to get the satisfaction of feedback on your own high-level cognition In a system, we’re all part of the system, taking part in its construction, obeying the system’s rules just the same, adjusting our own Bayesian weights I’m wondering, behind this whole thing, at which step might I go wrong? What am I good at, what am I worse at …AI is a mirror reflecting humanity’s unique value
Can ordinary people’s lives be interesting too?
The speech architecture of the OpenAI Realtime API
CapWords’ human-centred AI philosophy
Full-duplex interaction in the Realtime API
The end-to-end architecture of voice interaction
Giving choice back with utility apps
Middle-layer innovation in mobile development
Rork: The Advantages of Cross-Platform Generation and Its Native Downsides
Deductive Reasoning and Common-Sense Judgment
The Capability Boundary and Scenarios of LLMs
The Risk of Cognitive Models in the AI Era
The Value and Limits of Reinforcement Learning Fine-Tuning
The Air-Channel Principle Behind ChatGPT’s Voice
The Safety-Alignment Cost of OpenAI’s Voice Mode
The Hallucination Flaw of End-to-End Voice Architecture
The Three-Way Conflict of Fast and Slow Thinking in Voice Models
The Layered Routing Architecture of Silicon Valley Voice AI
Humanity Is a Fleeting Slice in the Long River of the Cosmos
Metacognitive Learning Lets Knowledge Explain One Another
Learning Tools Should Be Built Around Solving Problems
Apple Park Is the Physical Embodiment of Long-Termism
NotebookLM Gives Lawyers Controllable Cognition
AI’s Double Edge: Cognitive Dividend and Cognitive Debt
Using AI to Clarify Your Own Requirements
The Risk of AI Being Used to Harm Women
Structuring the Useful with the Usefulness of the Useless
The Thinking of Scientists vs. Engineers
AI Reshapes the Film and TV Recommendation Experience
The Huge Difference in Subscription Cuts on Web vs. iOS
A Content Distribution System: One Fish, Many Meals
The Design Philosophy of the Reminders List System
Patchwork AGI as a System State: A Hypothesis About the Future
Humans Can’t Foresee the Whole Shape of Nonlinear Structures
Brain-Computer Interfaces as a Bandwidth Revolution
RL’s Compute Overhead and Scaling Trouble
RL Moves from Core Paradigm to End-Stage Calibration Tool
Reinforcement Learning’s Real-World Use in Robotics
Robots as a Hedge Against China’s Aging Labor Crisis
AI Bubble Talk vs. Year One of Robots: A Cognitive Split
Robot Engineering Reliability Is Harder Than Algorithms
Natural-Evolution Rewards Decay After Fulfillment
Manus’s High Buyout and the Logic Behind Meta’s Acquisition
Apple’s SHARP Turns Ordinary Photos into Spatial Memories
Manus: More Intelligence with Less Structure
LLM Capability Is Growing Faster Than Real-World Demand
AI Extends Thinking; People Judge and Choose
Emotional Projection and Self-Knowledge Under Traffic Pressure
5. Daily Notes and Everything Else
47 entries The country is the people’s country So when the country has problems the individual should first of all have the ideology of sacrificing himself When you go to some countries, try learning a bit of their language you can easily place yourself into it and understand the country’s culture it’s about teaching you to treat the world as a game product concretely, how does it teach you? Thinking … Humans are, by nature, and will forever be, selfish animals Solving problems programmatically often solves problems that traditional programming cannot solve What can be changed are our flaws What cannot be changed are the weaknesses of our human nature Miyazaki is an adjective It is now completely an adjective We use “very Miyazaki” to describe something What we’re talking about is not just a director himself but invoking a general, highly recognisable aesthetic and emotional filter a particular kind of colour, a clear bright blue sky and layers stacked up like white cotton candy, full of life force a healing style a sense of nostalgia … a pure kind of goodwill I want to just take a random photo and have it pull me into that game world I heard a sentence no need to be afraid of change and no need to, out of fear of not daring to be yourself, pander to friends or to society just be yourself and on your friends’ side a snapshot of you is surely kept too maybe that is the best memory You can boldly change, boldly grow (solving the self), because those beautiful pasts have been properly kept by the people who love you (solving the sense of belonging). You don’t need to carry the shell of the past forward, because your friends have kept that shell for you At any decision point recognising your own decision through metacognition thinking about some thinking methods can improve the accuracy of decisions Nanchan Temple (Shanxi, China) Build a mental model -> master material selection logic -> frontier applications Building the underlying algorithm: materials science has its own programming language, or rather its own operating logic Phase Diagrams are materials science’s “map”. They tell you at what temperature and what composition ratio a substance will be in what state (solid, liquid, mixed) Materials is more importantly a decision system Materials informatics: Google DeepMind’s GNoME (Graph Networks for Materials Exploration). It used deep learning to discover millions of new crystal structures. This directly relates to future chips, batteries and superconducting materials This is a get-to-the-bottom-of-it way of thinking and cognitive habit Feasting on fine clothes and rich food yesterday, fleeing with an empty stove today Splendor, in truth, is a dream The feeling of being happy is a kind of ability Happiness is closer to a “long-term stable psychological structure.” It’s like a muscle: it needs training, and it needs habits to shape it. For example: the way you explain life, the way you handle setbacks, your connections with people, where your sense of value comes from, whether you can find meaning in the everyday… all of these can be changed through effort. Reading hard, thinking, reflecting, building relationships, cultivating a stable rhythm of life — these are the bricks that build the “ability to be happy” Joy … ah, God is fair … It is intense, brief, and cannot be sought on demand. It depends more on stimulation, environment, and chance events — a good meal, a joke, someone suddenly being gentle with you, buying the thing you wanted, seeing a strangely beautiful light It seems the only thing you can do is raise your own sensitivity to joy I really envy people who are full of joy; joy truly is a spark in the base color I really want to observe them … We’re already tired of “tapping around” Technology liberates us But now technology controls us …. How do I avoid wasting myself on a direction that is doomed to be mediocre??? Chapter 14 Looked at but not seen, it is called “yi”; listened to but not heard, it is called “xi”; grasped at but not caught, it is called “wei.” These three cannot be interrogated to the end, and so they mix into one. Above it is not bright, below it is not dark. Endless, it cannot be named, and it returns again to nothingness. This is called the form of no form, the image of no thing; this is called the elusive. Go to meet it and you do not see its head; follow it and you do not see its back. Hold to the Dao of old in order to master the things of today. To know the ancient beginning — this is called the thread of the Dao. Chapter 17 The highest: those below simply know that it exists; next, they draw near and praise it; next, they fear it; next, they mock it. When trust is insufficient, there is no trust. How leisurely, how sparing of words. When the work is done and things go well, the hundred families all say: “We are naturally so.” – Tao Te Ching Finding common ground with others through bag charms is exactly what people truly long for right now. If you can build a connection through a character, like Labubu, it opens the door to easy conversation and contact. Right now, people are looking for chances to feel joy and not take things too seriously “Paying in” essentially means which currency it is remitted through “You add” essentially means what currency you add into wise Forcing someone who is fundamentally a “tree” to become a “flower” usually ends badly for both sides — the flower doesn’t bloom, and the tree withers too Letting things take their natural course is not lying flat; it’s not going against the “Dao” of a thing — that is, its inner way of running. You can water it, prune it, but you cannot order a tree to “bloom roses right now” Sensory conflict causes dizziness Human balance depends on visual confirmation that the ground is solid and unmoving You look at the ground, but what you see is a reflection and an abyss; vision tells you that you are at the edge of a cliff, and the brain has to keep processing this conflicting signal The brain’s cognitive load is huge, and it easily triggers primal instincts The people of ancient Egypt loved life deeply, so they hoped to live forever after death After death the soul enters the underworld Many of the monsters with big eyes on bronze vessels They look fierce and terrifying, not to be cute, but to create a sense of awe during sacrifices and to communicate with gods and humans An ancient aesthetic God is the Dao, the Dao follows nature, Tathagata Knowing is self-awareness Principles and knowledge are useless; only when used are they useful. Faith is useless too; only when usefully applied is it useful What does self-awareness mean? Philosophy The essence of ClashX’s proxy modes Route / Rule / Global are “decision methods,” not “levels of takeover” Enhanced mode is only related to TUN, and has no direct relationship with “global mode” md Misled by posts on X again So many posts really have far too much false information From now on I really need to keep a skeptical attitude toward every post I read Following your desires without overstepping the rules How is overstepping the rules defined Achieving a republican game within the rules Maximize freedom, but minimize the damage to the rules Taste I feel taste still needs to be based on foundational training Constant repetition, accumulation, rich data sources — at least that way you can reach a passing score through training Only then do you add your own intuition, inspiration, and your own growth through experience The difference between once every two weeks and twice a week biweekly means once every two weeks whereas twice a week means twice per week bi means double, two — twice Similar ones include bimonthly, which is once every two months biannual, once every half year The nature of browser fingerprinting Browser fingerprinting doesn’t rely on cookies, so to some extent it can get around protections like VPNs But what’s worse is that the measures we might take to reduce the risk of fingerprinting can actually make that risk worse Fingerprinting works by having a web server extract certain discrete elements of information from the browser and combine those elements into a digital identifier I don’t know if this is because I haven’t been out there I really like the things of the outside world I’m very interested in history and places, so I want to go see them with my own eyes I want to go and go all in The invention of the car made most people physically weaker, but it lets top athletes get farther out to climb But you can’t deny the value of the car AI is raising the floor of the “average level.” Mediocre creation will become worthless, because AI can mass-produce it. It’s also raising the ceiling of “excellence.” People who are good at commanding AI will have the ability to direct an “intelligence legion” single-handedly — a dividend unlike anything before If you treat it as a replacement for a “search engine,” you may sink into the mire of vast information and feel an endless burden If you treat it as a “whetstone for thinking,” and use precise questions to push against its boundaries, you’ll get a huge cognitive dividend Sometimes you’re not unhappy It’s just that you haven’t been fully seen for a long time Every day you respond, you complete things, you move forward But few people ask you: all this happening to you — is it heavy? If one day you suddenly want to be quiet that isn’t retreat that’s just you confirming whether you’re still on this path Since I treat the world as a game, I don’t need to be so ego; I’m just one part of how the world-game operates Might as well bravely click start Truly at the do-or-die survival stage now Can you really return to yourself? If you never go through some low points a person easily slides into arrogance So sometimes meeting hardship is, in turn, a way of seriously arguing against yourself Before starting to write an article, you first have to make sure you have a very clear understanding of yourself Have you, in this world, ever been an irreplaceable presence to someone? Mental intensity != physical intensity Medium is really super suited to deep reading, not just clicks And you don’t need to build up your own follower base first If someone can’t even give advice to themselves can you still expect him to go and help others? The hard but right thing has become concrete I suddenly realized I don’t lack understanding of technology I lack understanding of the world I thought I lacked understanding of tech because I’d fallen into many traps of technology’s surface appearance From the user’s angle, from the product’s angle, from the technology’s angle all to explore the first principle itself, the problem that should be solved The kill line & the safety line The concept of the kill line is fairly abstract, and it’s actually hard to give a very objective assessment of the US US health care is extremely expensive, health insurance is extremely expensive too, and it’s very inefficient Without insurance it’s easy to fall back into poverty But isn’t the essence of a webpage expression? The emphasis should be on what the experience effect is like on the user side Use a webpage, build a recommendation engine From the user’s point of view, for most people the most comfortable form in the future is probably voice and not anything elseWhen a country has problems, the individual should sacrifice himself
Learning a language is a way of stepping into another country’s culture
Treating the world as a game — a product mindset
Humans are born selfish animals
Solving non-programmatic problems programmatically
Flaws can be changed, but the weaknesses of human nature cannot
Miyazaki as an aesthetic adjective
The imagination of a photo pulling you into a game world
Self-growth without pandering to others
Metacognition improves decision accuracy
Marking the location of Nanchan Temple
The decision system of materials science
The Impermanence of a World Where Splendor Is a Dream
Happiness Is a Skill, While Joy Is a Fleeting Gift
Technology’s Shift from Liberation to Control
How to Avoid Exploring a Mediocre Direction
The Dao and Governance in the Tao Te Ching
Finding Emotional Connection Through Plush Toys
The Flow of Currency Through a Wise Account
Let Flowers Be Flowers, Let Trees Be Trees
The Cognitive Mechanism Behind Dizziness from Sensory Conflict
Ancient Egyptians’ Pursuit of Eternity After Death
The Ritual Function of the Ferocious Aesthetics of Bronze Ware
The Practical Meaning of Knowing, Self-Awareness and Faith
The Essence of ClashX Proxy Modes as Routing Decisions
Keeping a Skeptical Attitude Toward Online Information
Freedom at Will and the Boundary of Rules
Taste Is Built on Training, Then Intuition Is Added
Biweekly Means Every Two Weeks, Not Twice a Week
Browser Fingerprinting Is Harder to Defend Against Than Cookies
Answering the Love of the Outside World with Full-Throttle Experience
AI Raises Both the Average and the Exceptional
You’re Not Unhappy; You Just Haven’t Been Fully Seen in a Long Time
Gamify the World, Then Bravely Click Start
At the Edge of Survival, Can You Return to Your Real Self?
Setbacks Are a Chance to Argue Against Your Own Arrogance
Clear Self-Knowledge Is the Prerequisite for Writing
Have You Ever Been an Irreplaceable Presence to Someone?
High Mental Intensity Is Not the Same as Physical Intensity
Medium Suits Deep Reading Without Needing Followers First
Someone Who Can’t Advise Themselves Can Hardly Help Others
The Hard-but-Right Thing Has Taken Concrete Form
A Lack of Understanding of Tech Is Really a Lack of Understanding of the World
Exploring First-Principles Questions from Multiple Angles
America’s Expensive, Inefficient Health Insurance and the Kill Line
The Essence of a Webpage Is Expression and User Experience
Voice Interaction Is the Most Comfortable Future User Form
6. Self-Knowledge and Psychology
31 entries Wang Shouren Unrivalled in his age A figure the Japanese deeply worship Everyone bows their head to Yangming Open any intellectual history of the late Edo period and at least one “Yangming scholar” will pop up within three pages The unity of knowing and acting resonates almost naturally with the samurai class’s ethic of action. A class that carries a sword has a professional instinct for “act immediately” Rebellion, overthrowing the shogunate, reform — all can be explained as “the command of conscience” People like Yoshida Shōin, Kumazawa Banzan, Nakae Tōju treated Wang Yangming almost as a spiritual lighthouse Third is grinding on things — thought is not sighing at the sky, nor sitting in a study waiting for sudden enlightenment, but striking sparks off the whetstone of reality Japan’s unity of knowing and acting An interesting manifestation is the Japanese ability to actually push some abstract ideas down into extremely detailed practice public order, for example How many categories rubbish must be sorted into, to what degree roads must be swept, how the body should lean when queueing, who to ask for help when in trouble Ideas become procedures, procedures become habits, habits become culture Then there’s the craftsman spirit — the word has been worn out, but Japan really does invest an unreasonable amount of time in one detail not chasing perfection, but treating perfection as natural as walking The quiet subway, hidden street corners, the low voice of service — not stiff politeness, but a continuous emotional regulation Highly sensitive people — but not necessarily emotionally sensitive Being sensitive is really just having a lower threshold of awareness for subtle changes: the gaps in language, shifts in relationships, shifts in scene, bridges between concepts so you can spot hidden signals more easily, and from that form a deeper understanding Not emotional sensitivity Emotional sensitivity is mostly the psychological system reacting quickly to external stimuli. For example being hurt by one sentence, being highly nervous about shifts in other people’s emotions. It is a defensive sensitivity, a capacity for feeling amplified in order to protect the self’s safety — a reaction But spiritual people are more in the observing and discovering stage; this is a quiet insight What this kind of sensitivity brings is not emotional fluctuation, but depth Spiritually sensitive people are often not emotional; on the contrary they are calmer. They are not “without pain”, but they see the source of the pain, see the useless struggle inside the pain, see that emotion is also a phenomenon, and so they can observe their own emotions the way they observe the weather From a psychological angle this is a sense of detachment — being the observer rather than the one immersed Sensitivity to the existence of meaning itself, not just to gain and loss Having experienced pain and impermanence but not being controlled by them, and even being able to perceive what you are currently doing — this is a kind of metacognition itself Understand the limitations of human nature itself respect it, obey it don’t try to reverse your own human nature try, by designing a series of systems or institutions, to go along with your own human nature and reach a better outcome False data produces paranoid and wrong views Yunnan people generally have a strong aversion to drugs Yunnan seizes more than 30% of the country’s drugs every year; in 2024 Dehong Prefecture alone seized over 600 kg of opium, and smuggling of new-type drugs (such as methamphetamine, fentanyl-type) is rampant. Since a professional anti-drug force was set up in 1982, more than 60 police officers on Yunnan’s anti-drug front have died, and many families are directly or indirectly harmed by drugs (such as drug-trafficking violence, family breakdown) Especially in small border towns They don’t just “dislike” — they regard drug users as a “social threat” The stigmatisation of drug users in Chinese society (including Yunnan) runs deep: using drugs = a rotten person, socially marginal, can’t find a job, no circle of friends. This is consistent with the nationwide “zero tolerance” anti-drug messaging; with relapse rates as high as 70-80%, people see them as “incurable” Drug use is already very serious in China; at the very least in China it is essentially a criminal issue Chinese criminal law can impose the death penalty for drug trafficking (especially large quantities, e.g. 15 g or more of heroin), but drug use itself is an administrative offence (Public Security Administration Punishments Law), at most 2-3 years of compulsory rehabilitation, no death penalty. This is because drug use is an addictive pathology (neural system remodelling, high relapse rate); the focus of punishment is on the supply side (trafficking) and on prevention, not on killing users. The death penalty only creates more social isolation and does nothing to cure the root The death penalty is effective in deterring trafficking (China’s drug situation has declined; existing users dropped from 13 million to 747,000), but ineffective for users — they are victims, driven by chemical dependence Most users are “0 to 1” victims (induced or pressured); giving them a path to turn their lives around (such as sealing records, but with strict control) can lower social costs — marginalising them only creates more secondary crime (like theft to raise money). But traffickers (especially those violently resisting the law) should be severely punished; the death penalty is necessary there, to protect anti-drug officers Analysing the drug-use data, many users did not fall in “voluntarily”, but were “hijacked” by physiological mechanisms after an initial try The meaning of hiking stamps 8 Sometimes I feel it really consumes my energy Maybe it’s like tourists liking to take photos at scenic spots as a souvenir I don’t really like being constrained or framed by things But I can also understand many people finding some meaning in it, wrapping uncertainty in a quantifiable sense of security; modern people like quantification — steps must have a number, running has an app, work has KPIs, walking mountain trails has a completion rate As I understand myself, I don’t need certainty; a growth-type personality can act amid ambiguity, won’t be constrained by the view of not knowing whether it is right, bears uncertainty and treats it as part of growth Not blindly confident, nor self-denying, carrying my own flaws forward, seeing my own biases and weak spots Expanding my knowledge of the world, the personality open to the world, but the values firm Modern mismatch theory The evolutionary buffer period of organisms Our brains are not for seeking truth but for seeking survival, a computer designed by natural selection The EEA, a brain designed for the hunting era rather than modern society — hence the causes of modern people’s psychological problems Adaptations: mechanisms evolved to solve a particular survival or reproduction problem — fear of snakes, craving for sugar Behind cursing On one hand it may be an emotional discharge; cursing makes the brain release dopamine, and this pleasure is very short, like fast food — it doesn’t solve the problem, it just makes you temporarily not hungry Then there is also a set of power and control: cursing is not as simple as expressing emotion, it is often to take back some kind of dominance Of course many people have also learned this way of speaking since childhood, this violent language What do I want? What kind of product or what way of iterating on a product actually works or what is the answer on this road? I’ve seen so many possibilities but it seems there really isn’t one I’d be willing to stake my whole life on I can’t casually tell myself a story either or numb myself with a grand narrative Yet I long for that feeling of going all in Instead of looking for the answer, I try to look for the process A better process of living A better process of learning A better process of creating Inside uncertainty, actually finding some certain answers for right now It isn’t romantic It isn’t stable But it is extremely heavy I ask myself: some questions really can’t be answered — who am I, who do I want to become but some questions, we can awaken to “In which cracks of this world, would I feel it was a shame if I weren’t present?” The essence behind this question presses you to ask what you want right now Environment vs. the attention of thought It doesn’t influence what you think about; it shapes what you notice, what you ignore, and what you think is worth thinking about People do not allocate attention freely. Attention has three main sources: part of it is top-down (goals, will, plans), part of it is bottom-up (stimulus intensity, novelty, threat), and part of it comes from the default patterns formed by long-term environmental training Whatever environment you live in for a long time, your brain learns: what counts as “normal,” what counts as “abnormal,” what is “worth being wary of,” what “can be ignored” So in an environment of high noise, high information density, and interruption at any moment, people gradually form a scanning type of attention Conversely, in an environment with low stimulation, a stable rhythm, and delayed feedback, attention naturally becomes more sustained, more inward, and more inclined to model-building than to instantaneous reaction This is especially cruel at the social level. If an environment long emphasizes: survival pressure competitive ranking immediate returns other people’s evaluations then attention naturally turns toward: risk avoidance short-term gains others’ intentions self-defense And if an environment allows: failure that isn’t fatal relatively ample time evaluations that aren’t immediate space for exploration only then can attention turn toward: first principles long-term structures abstract problems self-consistency Feelings after living in Japan for most of a month Walking around, I still couldn’t think out an answer Compared with living in other cities for a while, if the environment is familiar to a certain degree, it raises the focus of my product thinking But in Japan it’s hard to trigger this mode; some of the cafés I’ve been to in Japan so far, the trendy ones, are quite small, and it’s hard for them to give me immersive thinking space. This is understandable: the more vertical the space — especially something in the form of a church — the more it lets people return to their own thoughts and know themselves. But if the space is narrow, it will, to a certain degree, limit the vitality of our thoughts … Now I understand the sense of surprise of those huge cafés in Chiang Mai The inside is self-consistent You have a stable model that can explain and absorb outside events Something good happens → it’s understood as “reasonable,” “within expectation,” “integratable” Something bad happens → it’s explained, digested, placed into the causal chain The advantage of this system is: it is rarely destroyed Of course, the price is also that it is rarely pierced It is a mature kind of happiness But they can parasitize meaning onto external things They are willing not to understand for the moment, only to be struck Their self-boundary is softer, and the outside world more easily triggers the emotional circuit directly Not understanding it, not raising it to a value, not placing it into a life narrative, just letting it happen on you Freedom with rhythm A stable tempo, variable content But I think a very important problem is that there will be constant context switching The cost of that is very high, and efficiency is very low So I think every day you should guarantee at least one stretch of continuity At what time each day you do what — reading, email, input, output — all with a rhythm Rhythmical spontaneity is important A fixed rhythm, limiting the place, limiting the time, but not limiting the content Cognitive mapping in neuroscience The brain often binds environmental cues and mental states / behavioral patterns into stable mappings In certain scenarios it automatically triggers learning mode, relaxation mode, procrastination mode This binding can be strengthened, and it can also be retrained The conditioned result of environment → neural activation pattern → behavioral tendency In cognitive science, similar concepts include: Context-dependent cognition Contextual cues State-dependent learning Context–behavior association The brain maps a set of environmental features onto a set of behaviors and mental states predicted to require the least effort The brain sees partial cues —> automatically activates the whole old pattern In neuroscience this is called Pattern Completion It mainly involves the hippocampus + prefrontal cortex And then there is Pavlovian conditioning (but more advanced) What is reflected is a “mental state,” not a simple action Environmental stimulus (a cafe) → dopamine prediction (I’m usually productive here) → the motivational system is pre-activated → resistance to learning drops The identity of the self activated in the brain is also completely different Learning and recalling in the same environment shows significantly better performance Remind yourself what kind of thing to do in what kind of environment And repeatedly bind and reinforce this behavior It’s a bit like reinforcement learning But note: don’t scroll videos etc. in a new environment, because once it’s contaminated the brain will rebind Let the environment help you accomplish half of your self-control The essential cause of guilt The brain is a prediction machine, and we are at all times maintaining an internal model of who I am This is also our cognitive-behavioral model So if behavior != expectation that is cognitive dissonance; the brain hates this inconsistency between cognition and behavior, so it releases negative neurotransmitters to make you feel pain, and you then eliminate the error by apologizing and repairing We have advanced theory of mind, the ability to understand other people’s mental states, so when you hurt someone, your brain runs a “simulation program.” You simulate the other person’s pain, disappointment or anger in your mind Guilt is essentially a kind of empathic response to the “simulated suffering of others” that you caused yourself Guilt is also a mechanism of cooperation, a rule of commitment; it proves to yourself (and to others): “in this game, even though I messed up this time, I feel terrible, which shows I still care about the ‘rules of cooperation’; please don’t kick me out of the game” — this is the repeated game in human evolutionary psychology So there are two things behind it: Genuine guilt, meaning you really did violate your own core principles, for example promising a deadline and not making it; then fix it, make amends or apologize, and don’t keep burning yourself up Or it’s a mismatch: the unease produced because a goal you yourself consider unreasonable hasn’t been reached. This is a False Positive. Your “self-model” may contain outdated, overly harsh rules (like “must produce every minute of every day”). At this point you need to Refactor your internal rules and tell the brain: “Rest is a feature, not a bug.” Mysticism likes to explain it with “aura” and “magnetic field” exchange, but from a neuroscience perspective, the energy difference between I-people (introverts) and E-people (extroverts) in socializing is essentially determined by the brain’s sensitivity to dopamine and the baseline level of neural arousal The brain pathway of E-people (extroverts): depends on dopamine The brain pathway of I-people (introverts): depends on acetylcholine The introvert’s brain is highly sensitive to dopamine. A little social stimulation (like three people chatting) already saturates their dopamine. If there is too much stimulation (a party), the brain becomes anxious and exhausted from “overload” The extrovert’s brain is very hungry for dopamine but has lower sensitivity. They need higher-intensity external stimulation (lively parties, loud music, frequent conversation) to activate the reward system and produce that “buzz” Introverts prefer the pleasure brought by acetylcholine. This is a neurotransmitter released during rest, focus, reading and deep thinking, and the feeling it brings is calm and contentment Give the brain a definite signal Definite, not fuzzy Just like handling WeChat If every time you open WeChat it carries the psychological cost of “guilt + sneaking around” This is because the system automatically seeks the path of least resistance The brain knows there’s still a window: WeChat on the phone The brain still knows that this reward expectation exists In neuroscience this is called partial extinction And the features of partial extinction are: the urges become more frequent, but shorter The brain is repeatedly confirming: “is it really gone?” This is because it’s an undefined rule of use The current implicit rule is: “better not to use WeChat” But what the brain needs is: clear, predictable, low-friction rules Fuzzy rules bring mismatch, and mismatch brings pain So the brain needs something permitted but tamed: a fixed window and fixed time for use For example, how often to check WeChat Or situational isolation: keep the phone out of your field of view, turn off non-essential notifications Make the cognitive-behavioral path longer If a choice can only hold up by way of a “cliff-edge change” then it probably isn’t mature enough yet The path that truly suits us often isn’t one that has to be bought by immediately smashing everything A choice is “possibly viable” not because it is intense enough but because — even without a cliff-edge change, it can grow on its own This is understanding of and trust in yourself and the world The ease I understand is a mental state, not a behavioral state The essence of ease isn’t slowness, it’s not being chased by inner pressure Many people mistake ease for “lying flat,” “going with the flow,” “low desire” Ease doesn’t mean low behavioral density; I think what matters most is that psychological tension is very low In fact the core test is one sentence: If I don’t do it now, what happens? No matter how the outside world sees it, no matter how others see it, it won’t affect my choice in this moment Ease isn’t not striving, and it isn’t not going fast; it’s: I can be fully invested, but I’m not pushed into a corner by myself More fundamental than suffering is discovery Once you see the suffering, developing relationships becomes harder These people can see control, exploitation, neglect, emotional blackmail, structural unfairness control, exploitation, neglect, emotional blackmail, structural unfairness This is an ability to perceive suffering Intimate relationships are harder precisely because they require a default premise: that we won’t keep hurting each other Almost all relationships, at certain moments, contain a structure of harm Because no one wants to be the harmer all the time, unless they really are Most people aren’t bad people, they are just: protecting themselves repeating a learned pattern unconsciously passing the pressure downward At the subconscious level they pretend they can’t see it How do people who have seen the truth keep being with others Truth and goodness don’t automatically travel the same road “Seeing suffering” raises your epistemological level, not your moral level Relationships aren’t built on truth, they’re built on bearable falsehood This is a game played by both sides People don’t live in a moral world, they live in a world of how much they can bear People refuse to admit they are the harmer not because they are truly “evil” but because — admitting it would directly pierce through their self-structure So they keep maintaining the integrity of the self Love and awakening come into conflict at a certain stage The more clear-headed a person is, the fewer relationships they have; the fewer people there are, the purer the love, but also the lonelier This is a stage But people also get a second naivety: even seeing the cost clearly, they are still willing to ally with others Parents also need meaning for their own existence This is a way of fighting off powerlessness Even if the uselessness brought by being sidelined by society devours them but worrying about you is, at the subconscious level, how he maintains his dignity as an elder — through “being needed by you” or “fretting over you” I stood at the headwind of the era and enjoyed the dividends; he was left behind by the wheels of the era, covered in dust The handover of authority sometimes really does bring enormous pain Sometimes, as children, as the people closest to them don’t deprive them of the right to give This is from a human nature point of view It’s not about what a person says, but how they process information Where the information comes from: active (your own observation), passive (views fed or packaged by others) The ability to structure information — people with high agency internalize experience into reusable models People with low agency just record scattered fragments, profound on the surface, fragile in reality Look at whether a person has a “loop for generating explanations” rather than just “consuming explanations” Storytellers are interesting, but pull yourself out instead of inserting yourself into the subject of the story, and think about the story itself: is it causality or a conclusion If it’s always conclusion-style — “that’s just how the world is, you have to learn to accept it” — that’s secondhand knowledge If it’s mixed with process and reflection — “I observed… so I guessed… later I found…” — that shows a generative loop Another point is whether it’s self-consistent People with high agency, no matter how much of the world they’ve seen, still have “their own bottom line / criteria for judgment” People with low agency tend to be fragmented, contradictory, unstable The urban-rural gap, to a very large extent and very fundamentally, is an unequal structural distribution of resources and rights The most fundamental urban-rural gap in China is inequality in institutional resource allocation and development opportunity, not population size or the apparent happiness of everyday life. So-called “happiness” is often stability of life and psychological satisfaction, but from the standpoint of development potential and freedom, that happiness isn’t the same as fairness. Urban residents enjoy more options and more possibility of capital accumulation, while rural residents are institutionally constrained — that’s the deepest essence of the urban-rural gap Cities have a great deal of disposable resources; rural land has always been collectively owned, so farmers don’t have the capital accumulation and mobility that urban residents do The household registration system tightly binds population mobility to social welfare, and rural residents moving into cities are still restricted Cities have highly developed industrialization and services, while farmers mainly depend on agriculture and low-value-added industries. The gap in economic added value and opportunities for value growth has persisted for a long time, directly affecting income, social status, and lifestyle My childhood environment and living environment caused my eustachian tube problems Swelling of the nasal and nasopharyngeal mucosa → restricted opening and closing of the eustachian tube → poor ventilation in the middle ear, negative pressure → secretions can’t drain out Even if the surgery goes very well, the middle ear’s self-cleaning ability and airflow dynamics may not be as smooth as before Let the snot come out on its own, instead of squeezing it out Pressure forces the eustachian tube open in reverse And the result is nasal secretions + bacteria get pushed into the middle ear → the operated ear is more likely to have trouble You should only do one side at a time, and exhale gently, not explosively Actually the right steps should be to use saline first, then after a few minutes it dilutes, then drain one side ropet and Fuzozo seem to give girls a lot of emotional value; they’re extremely hot plush toys on the market right now, suited to girls who are alone, or girls under a lot of stress — a daily hug recharges your mood Classic representatives of the healing genre; soft cute designs (like vegetable, fruit, or animal series) bring immediate comfort. Girls often name them and take photos of them, sales have soared, and they symbolize emotional attachment and the return of childlike play Then there are aroma diffusers and flame humidifiers, creating a small sense of atmosphere, like having a mini spa at home Mood lamps / crystal healing lamps, suited to girls who need to relax mentally And creative little plant items (like the “no-anxiety” banana plant or moss terrariums) — one glance a day And custom photo books or memory albums, which can be given to friends Recently I’ve realized how I’ve changed Exploration itself no longer gives me enough traction In the early days of living abroad, exploration was mainly three things overlapping The unfamiliarity of space — a new city, a new language, a new rhythm The loosening of identity — I’m no longer the person defined by a position The swelling sense of possibility — maybe I can become something else During that year of living abroad, I could feel this was extremely high-energy Even how to face the world, understand the world, understand culture — I had my own theory and knowledge system But after more than a year, the world actually started repeating Cities change, but the underlying structure is similar I adapt faster and faster, yet the shock gets smaller and smaller Going somewhere no longer automatically equals being changed So it seems I can’t make exploration the core anymore I need to find myself, to find a new fulcrum. When the fulcrum is blurry, moving is really just changing the backdrop, and it won’t bring much thinking or inspiration A lot of people at this stage choose one of two directions: increase the intensity of exploration, go farther or in a more different direction or start blaming themselves, reflecting on whether they should return to normal, whether living abroad can’t go on forever, whether novelty will change I seem to be reflecting too, realizing that I almost chose to continue in a farther direction, like planning Europe or Africa But what I want to know more is two things Why is this happening, and what do I actually need? If countries/numbers are the main line and there’s no novelty, what can I keep, or keep going? Even though I could predict our cash-flow problem early on facing it for real is pretty helpless So sometimes hearing their stories, hard stories, is a kind of resonance… there are people worse off than me, haha It means next year is another year starting from 0, putting everything down There isn’t much to write in this year’s annual review either Compared with friends I’m still pretty lucky; the sunk cost isn’t too big It’s just that I can’t go back to the traditional system In the end I walked my own path, but … Just treat it as a game Feels like I should be nicer to my friends I should start giving people-pleasing replies to their questions … After all, if next year I don’t even have money for food and sleep I’ll have to lean on friends, crash at someone’s place, or at worst cook for them Ability is an upper line, survival is also a lower line; I need to pull the survival line down a bit Moore’s Law tells me the number of transistors that can be integrated on a chip grows by doubling every couple of years, so computing power at the same price gets exponentially stronger Naturally models are like this too — models keep getting stronger This is a trend One core engine of modern civilization is turning “more complex things” into “goods that can be reliably mass-produced,” and the rhythm keeps getting faster, the cost keeps dropping — especially obvious in China, with its powerful supply chainWang Yangming’s thought deeply influenced Japanese samurai
How Japan turns abstract ideas into meticulous practice
The deep perceptual power of highly sensitive people
The difference between spiritual sensitivity and emotional sensitivity
Designing systems that go along with human nature
Drug data and the reality of stigmatisation
Hiking stamps and the myth of quantified security
The environment-of-adaptation period in biological evolution
The psychology behind cursing
Looking for an Answer in the Process, Inside Uncertainty
How Environment Shapes Attention Rather Than Thought
Living in Japan for a While and Space for Product Thinking
A Self-Consistent Inner System of Meaning
A Free Life With Rhythm
The Brain Binds Environmental Cues to Behavioral Patterns
Guilt as Cognitive Dissonance and an Empathic Response
The Neuroscientific Essence of the Introvert/Extrovert Difference
Give the Brain a Definite Signal, Not Vague Rules
Real Change Doesn’t Need a Cliff-Edge Break
Ease Is a Mental State, Not Behavioral Density
What Makes Intimate Relationships Hard: Spotting a Latent Structure of Harm
How to Keep Being With People After Seeing the Truth
Parents Need to Be Needed in Order to Keep Their Dignity
High-Agency People Generate Explanations Instead of Consuming Conclusions
The Urban-Rural Gap Is Fundamentally Unequal Resource Allocation
The Physiology Behind Eustachian Tube Problems
Emotional-Value Plushies and the Rise of Healing Little Things
Seeking Yourself After Exploration Stops Giving Traction
Facing a Cash-Flow Crunch by Starting the Year from Zero
Being Nice to Friends Is a Strategy to Lower Your Survival Floor
Moore’s Law Drives Exponential Growth in Computing Power
7. Business, Investing and Career
25 entries The biggest dividend of power: others bleed, you just need to shed tears A sense of missing economic knowledge What I pursue is the essence of the system and the underlying logic. If the jigsaw piece of economics is missing, the world really does look fragmented in your eyes: I see war, but don’t understand the chain of interests behind it; I see policy, but don’t understand the resource game behind it A lot of the time I seem able to build an understanding of certain things through politics and culture, but more often I’m at a loss about economic factors Economics is quantified, it is rational; studying economics often also means you have to break away from traditional emotional moral standards, from real resource constraints Human beings essentially need incentive mechanisms, just as iOS architecture design guides developers to write certain kinds of standardised code — system design determines human behaviour The Silk Road was not planned by some emperor, but walked out by countless merchants chasing profit; so you can understand why, at many levels, excessive intervention by a centralised authority often fails Fushimi Inari Taisha An endless tunnel of red torii, visually extremely striking The torii are donated by individuals or companies, and can also be a way for commercial companies to give back or even advertise; this mountain has almost ten thousand torii… at the going rate of at least twenty thousand yuan for the smallest torii, the red desire of the whole mountain seems to be saying, “I want to make money”, “I want to succeed”, “I want to live better” Every torii records a story; for example, the founding of Fushimi Inari Taisha also originated with the Qin clan, a technological immigrant family from the continent, who made their fortune through hard tech, and there are also stories like Toyotomi Hideyoshi’s of knowing how to repay a kindness … The Big Buddha at Tian Tan The Tian Tan Buddha sits on the Ngong Ping plateau on Lantau Island and is one of Hong Kong’s most symbolic landmarks, but what’s really impressive about it isn’t just that it is “big” The statue was completed in 1993, about 34 meters tall, cast in bronze, weighing over 250 tons. The “Tian Tan” in its name was not chosen casually — its base proportions reference the design idea of Beijing’s Temple of Heaven: between heaven and earth there is a stable, centered point of order. Visually and psychologically this is very “steady”; even if you don’t believe in Buddhism, standing there you unconsciously slow down The statue itself is a seated Shakyamuni Buddha, right hand in the abhaya mudra, symbolizing the removal of fear; the left hand holds a bowl, representing giving and receiving Its back faces Lantau Peak, and it looks far out toward Hong Kong Island The mountain is stability, the city is sentient beings, and the 268 stone steps up to the Big Buddha were not piled up at random either. In Buddhism this number symbolizes the process of affliction and practice Opening up land at large scale at this particular moment is, for Hong Kong — in the short term more harm than good, in the medium and long term more good than harm, but system rationality will choose “rather drag it out than gamble” The most visible thing is of course the relief on the housing side Supply increases, expectations turn, the pressure on young people drops, social resentment is released, and part of a class sees again the possibility of “upward mobility” From the perspective of social psychology, this is very important, even “bleeding-stop level” The biggest “harm” of opening land now is not money, but “the risk of collapsing belief” Hong Kong real estate is not just an industry; it is: the base of the banks’ balance sheets the main net asset of middle-class families the implicit credit endorsement of government finances society’s psychological anchor for “future predictability” Once, in a situation where “the external environment is already unstable,” you actively release a large supply of land, what signal does the system read? “Even you think the land price can’t hold up.” So one is a chronic illness, the other is cardiac arrest Limited, zoned, targeted, slow land opening — that is the only rational solution right now Stretch the cycle out extremely long Hong Kong cryptocurrency regulation Hong Kong’s attitude toward cryptocurrency (more precisely: virtual assets, Virtual Assets) is neither “letting it fly” nor “an all-out ban,” but openness under heavy regulation. This forms a very deliberate contrast with the mainland Since 2023, Hong Kong has operated a virtual asset service provider (VASP) licensing regime As long as you touch cryptocurrency in a “doing business” way in Hong Kong, you must hold a license. Including cryptocurrency exchanges custody services brokerage and trade matching trading platforms aimed at the public It is currently regulated by the Hong Kong Securities and Futures Commission (SFC). Operating an exchange in Hong Kong without a license → ❌ illegal Licensed but non-compliant → ❌ illegal Changpeng Zhao feels full of contrast He is a legendary figure who went from a McDonald’s worker to the richest person of Chinese descent, and after going through prison, turned himself around in 2025 into an “education philanthropist” He is fundamentally different from FTX’s SBF (the curly-haired kid now in prison) He emigrated to Canada with his parents at 12, and the family was once hard up. In Vancouver he delivered newspapers, worked night shifts at a gas station, and for two years worked the griddle flipping patties at McDonald’s. This bottom-level experience made him understand the rules of survival better than those Wall Street elites, and made him tougher A typical engineering guy — he usually wears a hoodie and flip-flops, looks gentle, even a bit uncool, but his business methods are extremely ruthless. Back then, to build Binance, he sold his house in Shanghai and went all in on Bitcoin; facing rivals (like FTX), he didn’t strike unless he struck, and when he struck it was a kill shot Many people thought that after prison (4 months in 2024) he would fade away, but in 2025 he was more active than ever, and had switched to a new “persona”: His most core project now is called Giggle Academy. It is a completely non-profit, free online education platform, aiming to provide free education from grade 1 to grade 12 to children worldwide (especially in poor regions) through gamification. In 2025 he gave many public speeches (such as in Dubai and Hong Kong), all promoting this education project rather than promoting Binance In August 2025, he appeared for the first time in four years at the University of Hong Kong to give a speech. This released an extremely important signal: he has already passed through the “tribulation” of US regulation, and has returned as a free man to the center of Asia’s financial stage Six months to make Binance number one in the world, and then a philanthropist trying to change global education with technology … Humanity’s longing for scarce goods A scarce good like digital currency And the consensus about this scarce good A scarcity locked in by consensus, that power cannot tamper with This is also how we understand why gold is so valuable Gold is scarce, chemically stable, and hard for power to manufacture at will Even if it is only a social experiment I also think this is a very interesting social experiment An Inception-style reading room The infinitely reflecting “well” The highly reflective black floor tiles on the ground aren’t just flooring; they’ve been treated as a “still water surface” The ceiling is a “fictional sky,” the floor is a “deep lake surface.” Sitting in the middle, a person seems to float between sky and earth. This weightlessness makes customers psychologically detach from the busy real world and enter a focused state of “reading / emptying out” Walden Dark green velvet = forest / moss. Natural wood color = trees / wooden cabin. Black high gloss = lake water. It retranslates Thoreau’s Walden using modern materials (mirrors, baked enamel): it keeps a sense of nature without losing the refinement of modern commerce The essential underlying difference between the economic and financial systems of Hong Kong and the mainland It actually has nothing to do with technology; the mainland’s payment convenience is of course perfect, it solved a lot of traditional convenience problems as well as security and trust problems, and Hong Kong is a free trade port On the mainland, banks are essentially part of the national financial infrastructure Non-banks (Alipay/WeChat) are allowed to be “a layer of operating system wrapped around the outside of the banking system” But is Hong Kong essentially different? Yes, the core is market entities The market is not part of the administrative system Non-banks (wallets) can only be SVFs (stored value payment instruments), with a status lower than banks The central bank / HKMA itself stepped in and built a “public payment highway” = FPS The mainland is “platform-driven payment” Hong Kong is “infrastructure-driven payment” So in the mainland’s case, money floats in platform ledgers, supervised, but not in bank accounts; banks are the ultimate clearing tool, and platforms are allowed to run “quasi-bank-level” ledger systems But in Hong Kong, money must stay in bank accounts; wallets can’t just build ledgers at will; a wallet must either be merely an interface or be an SVF (a stored value instrument), under strict supervision So now it makes sense why the mainland doesn’t need FPS: the mainland already has one super-large FPS, and that’s also why the digital yuan was launched — to weaken the financial nature of the platforms But Hong Kong has more banks and no monopolistic platform at the Alipay level; banks compete fiercely, and without FPS the result would be slow, fee-charging bank transfers, with everyone doing their own thing Alipay and WeChat are ultimately still presentations at the application layer, but they have also formed a private infrastructure monopoly, with banks degraded into pipes; merchants must connect, all players must participate In Hong Kong, by contrast, the infrastructure is more free; there is no phenomenon-level product Banks can’t lock in users Wallets can’t lock in merchants Innovation can only happen on top of the road No “private highways” are allowed The essence of Hong Kong’s capital freedom In one sentence: Hong Kong is an institution designed for “capital flow,” while the mainland is an institution designed for “social stability and long-term national goals” The Hong Kong model: Allow capital to flow in and out freely → The shocks of exchange rates, asset prices and financial cycles are borne by the market and by individuals. Some people get rich overnight, some go bankrupt; the government not bailing you out is the default option. The mainland model: Restrict capital flow → Keep systemic risk inside “the institution” to be digested. The exchange rate, foreign reserves and financial stability are regarded as part of national security. Hong Kong says: as long as you’re legal, I don’t ask who you are, I don’t ask where you’re going, I don’t ask why The mainland is still at the stage of industrial upgrading, catching up technologically, and it still has a demographic dividend So it needs long-term, stable, controllable capital So it feels like even freedom — even the freedom of capital — is a luxury, not a basic human right Only a handful of strong countries in the world, plus some very small countries and regions, can afford to pay that bill Bitcoin and gold Gold’s total quantity on Earth is limited Bitcoin’s total quantity is also limited; how each coin is generated and the issuance pace is hard-coded in the code In theory, absolute scarcity Bitcoin is fully digital, circulates at almost zero cost, can be divided precisely to one hundred millionth (a satoshi), and has extremely high circulation efficiency Gold’s price is affected by supply and demand, the economic environment and geopolitics, and it fluctuates, but over the long run its store-of-value property has been strong for thousands of years Bitcoin’s price is very volatile and easily swayed by financial market sentiment Information gaps and supply-demand reversal This is a business judgment about when to enter and how to exit When policy isn’t yet clear but is still being drafted, that’s often when the business opportunity is greatest At this point there’s a regulatory vacuum, a low-cost phase, the rules aren’t set yet, compliance costs are low, and there’s plenty of room to try and fail The giants are watching, the public doesn’t know, there are few competitors, and you get pricing power Profit comes from “non-consensus.” When something doesn’t need anyone calling on people to do it and they do it anyway, it’s usually because there’s a real business pain point and real profit in it When the government starts loudly “calling on” people, it usually means two things: first, this field needs outside force to push it forward (it isn’t self-motivated); second, it will lead to an irrational flood of resources The moment the call goes out, a mass of speculators and subsidy-chasers who never had the capability pour in. The market instantly goes from blue ocean to red ocean, a price war starts, and profit gets thinned out Companies answering the call often ignore real customer needs and business logic in order to grab subsidies or meet targets This is also the classic case of good news exhausted turning into bad news When the public is all responding to the call and piling in, asset prices (valuations) are often pushed to their peak. Selling then gives the best liquidity and the highest premium The core of doing business is making money off “market laws,” not off “policy sentiment” Getting infrastructure or middle-layer tech right makes it very easy to make money When everyone wants to be the “Microsoft” of the AI era, they choose to be the AI era’s “power grid” or “water bureau.” The government calls on everyone to “use electricity,” and they’re holding the “power plant” AI’s ultimate bottleneck is energy In a gold rush, the people selling shovels always profit; in the AI boom, the people selling compute and holding private data are gods The very top hunters actively embrace regulation. Because high compliance costs and safety standards become a natural moat that shuts out latecomers. They don’t “run” — they close the door behind them When pure-digital-world AI turns into a red ocean and the government starts heavily regulating content generation, they’ve already moved into Physical AI (embodied intelligence) They use today’s bubble (government/capital hype) to feed an ambition that can only be realized 10 years from now (like AGI or interstellar migration) When will I be able to move in and out of the market that calmly? Entrepreneurs and business builders Whether an artist paints a painting or an entrepreneur founds a company, at bottom it’s an expansion of life force The artist captures intuition through the canvas, externalizes inner emotion, and gives the world an aesthetic experience that didn’t exist before The entrepreneur captures opportunity through a business model, makes abstract ideas concrete, and gives the world a service or product that didn’t exist before Both are obsessive The artist’s obsession: if it isn’t painted this way, the work might just die The entrepreneur’s too: if it isn’t done this way, the vision simply can’t be realized The artist seeks order in chaos: turning complex emotion into harmonious lines or notes The entrepreneur seeks chaos in order: finding the breaking point in a stable market, and building a new empire by overturning the old order What is Joseph Tsai actually after? Early on Joseph Tsai studied and practiced on Wall Street His core thinking: wherever there’s potential, that’s where it’s worth investing — which explains why he chose Alibaba It’s not just short-term money-making; what he values is shaping the company and the industry landscape — for instance taking part in internationalization and investment in Alibaba’s early days, or buying the Nets, are all staking out future value and influence He’s low-key and has control; what he seeks is the ability to control the situation, not high-profile exposure — orchestrating from behind the scenes, letting things run according to the long-term logic of his own judgment He’s a long-term systems thinker Making innovation an institutional capability, encouraging interdisciplinary students to solve real-world problems with innovation Emphasizing “technological innovation” and “applying technology to core business to create value” — the classic “treat technology as a tool of productivity, not a religious totem” Long shuttling between China and the US, and openly explaining the complexity of technology, trade, and US-China relations Treating long-term assets as something inheritable — not just companies, but institutions, teams, education, and philanthropic projects Tuning a complex system into order — what he pursues is a “structural influence” that keeps working across cycles, cultures, and organizations P2P in Binance I keep wondering why Binance would offer P2P — isn’t running an exchange good enough? P2P in essence means you’re not trading directly with Binance You send money to some seller Binance “locks” the coins for you Seller confirms receipt → the coins automatically arrive in your account 💡If the seller won’t release the coins, you can appeal, and Binance arbitrates Several different ways to buy coins One is P2P, where the counterparty is another user, you can use fiat, and the price is free Then spot market, market-matched, which requires you to already have coins Fast fiat purchase of coins — convenient but expensive in the moment Binance offering P2P is essentially to avoid regulatory problems — directly accepting fiat and providing cash-out services brings a lot of problems Dirty money going straight into the platform account makes the platform the primary responsible party P2P puts the payment between users; the platform mainly does: custody, matching, arbitration, rules. Risk gets “shared out” It looks on the surface like Binance earns less this way, but actually Once a user completes P2P: money becomes USDT / BTC the asset lands in a Binance account Next there might be: spot trading futures (high profit) lending / wealth products fees generated over and over 👉 the downstream revenue is far greater than the one hit at the entrance Bitcoin is too volatile, not suitable as a unit of account So stablecoins appeared Why doesn’t deflation overturn the structure where getting closer to new money earlier is an advantage? Under inflation, borrowers as a group necessarily come out ahead Under deflation, those who can survive necessarily come out ahead Deflation is usually more alarming and more detested than inflation Because deflation creates a death spiral Money becomes more and more valuable People don’t want to spend or invest Corporate revenue falls Layoffs, pay cuts Loan defaults Bank bad debt Credit contraction More deflation The essence of money is using future credit to spend now, but people no longer want to borrow from the future, so the state system keeps fighting deflation Behind buying a house with a loan is actually also discounting your future 20–30 years of labor into “today’s asset price” using personal credit to provide the whole financial system with long-term, stable, controllable cash flow The essence of a security is a standardized financial contract representing property rights It realizes the transfer of rights and the sharing of risk: specific property rights (like ownership, creditor’s rights, rights to income) get packaged into a standardized certificate, and the holder can enjoy the corresponding rights through that certificate while also freely transferring it in the market Securitization is essentially the financial language of putting assets to work: the process of turning assets that lack liquidity but can produce stable future cash flow into securities that can circulate in financial markets So by buying a house we’re “securitizing” ourselves and mortgaging ourselves to the bank and the national monetary system The bank conjures 1 million in deposits out of thin air through credit expansion Money = credit that can be enforced; for the bank it’s even low-risk … because the bet is that your life won’t fail, and the house is the bond wrapper of local finance The low price is also cashed in early; the city’s construction cost is paid by your future 30 years of labor, and the housing price is the city’s financing tool A line suddenly comes to mind: the bank’s favorite customer isn’t the rich, it’s the middle class with a stable income who are afraid of losing everything This is a system, a rational and cruel system design; read this design and you can read where real estate is headed When your core motive for buying isn’t “arbitrage” but only “fear of falling behind” then in a financial sense, it’s usually already a wrong signal China’s real estate positioning vs the US and Japan A very distinctive characteristic Housing prices no longer rise across the board (the Japanese shadow) Core assets are still propped up (the American shadow) But the method is extremely non-market (unique to China) In Japan, social consensus allowed sacrificing a generation, and local finance wasn’t highly dependent on land Real estate = the underlying structure of local finance, so around real estate, infrastructure, employment, debt rollover, and city operations all run into trouble Compared with the US: the US has global currency hegemony, global capital flows back to it, and inflation can be exported outward — China doesn’t. The RMB isn’t a global safe-haven currency, and China’s total population is declining, with cities rising and falling at each other’s expense, while US population relies on immigration So China can prop up some cities, but can’t prop up all real estate This leads to what may happen with Chinese real estate: core cities won’t fall, they’ll be held up and become quasi-financial assets Provincial capitals and second-tier cities will gradually Japanify — population dipping slightly, prices flat or drifting down long-term, housing turning into consumption Then there are the marginal cities, where the price exists in name only, negative liquid assets The new binder behind a mortgage A binder is a structural tool that any complex society necessarily produces when maintaining stability As long as a society needs: continuous production + predictable behavior + long-term order some mechanism for “locking in the individual’s future” will definitely appear The essential question at the social level is: How do I trust that you’ll still be in the system 20 years from now? The binder’s job is: to outsource “trust” to structure to turn “morality” into contract to turn “consensus” into cost Mortgages, household registration, debt, qualifications, points — all essentially do the same thing: make “leaving” expensive Society needs money, cities need money, the state needs stability — these needs naturally generate binders So the powerful and the interested parties both reinforce and optimize it, and the cost is outsourced to the weak A binder is also just a tool, not an enemy; the key is whether it carries the meaning of your life The Northern Metropolis Aimed at building Hong Kong’s northern area into an international innovation and technology center that’s good to live, work, and travel in, running alongside the traditional Victoria Harbour metropolitan area (mainly finance) to form Hong Kong’s “twin-engine” development pattern Current population is about 950,000; when planning is complete it can accommodate about 2.5 million residents, providing roughly 905,000 to 926,000 housing units (with a high share of public housing), addressing Hong Kong’s long-term housing shortage Expected to provide about 650,000 jobs, of which 150,000 are related to innovation and technology Expected to be completed in phases within 20–30 years, taking initial shape within 5–10 years Expected to become a key engine for Hong Kong to integrate into the Guangdong-Hong Kong-Macao Greater Bay Area and drive the economy to take off again State capital and SOEs were never fully decided by market competition, but by institutional allocation and administrative coordination The biggest difference is whether you can get into the resource-allocation layer vs. only being able to compete at the results layer In the end, is it about entering a higher system and gaining recognition, or building a system that doesn’t depend on recognition? Zero marketing budget Cash has to prioritize R&D / operations / cash-flow safety, so marketing spend gets squeezed to the minimum This is a constraint, and you have to work backward from this constraint to figure out what to doThe dividend of power is making others bleed
The felt absence of economic knowledge and its underlying logic
The desire and stories of the red torii gates
The Big Buddha at Tian Tan: Order and Symbolism
Weighing the Pros and Cons of Opening Land in Hong Kong
Hong Kong’s Approach to Cryptocurrency Regulation
The Contrasting Life of Changpeng Zhao
Humanity’s Consensus Longing for Scarce Goods
The Weightless Aesthetics of an Inception-Style Reading Room
The Underlying Differences Between the Mainland and Hong Kong Payment Systems
The Institutional Design Essence of Hong Kong’s Capital Freedom
Bitcoin and Gold: Scarcity and Volatility
The Policy Vacuum Is the Best Moment to Enter for Profit
Infrastructure and Middle-Layer Tech Are the Easiest to Profit From
Entrepreneurs and Artists Are Both an Expansion of Life Force
Joseph Tsai’s Long-Termism
Binance P2P and Regulatory Avoidance
Bitcoin’s Volatility and Stablecoins
The Deflation Spiral and the Advantage of New Money
Securitization and the Nature of Mortgages
China’s Real Estate: US and Japan Comparisons, and Its Own Traits
The Mortgage as a Social Binder: Its Structural Essence
The Northern Metropolis: Hong Kong’s Twin-Engine Blueprint
State Capital and SOEs: Resource Allocation and the Recognition Trap
The Survival Constraint of a Zero Marketing Budget
8. Body, Health and Daily Life
17 entries A good photographer is certainly a good observer They have their own intensity of perception towards life This is an active game The world drops a tiny clue, and they can immediately piece together the emotion, the rhythm, the story behind it This kind of observation is not gossipy staring, but a sensitivity to detail — the sound of wind, movements, changes in light, the energy of a crowd, and those small cracks that are easily overlooked. Photography is just freezing these moments so that others can also see what you saw Photography technique can be trained, but observation is a person’s way of seeing the world; once that way matures, the lens is just an extended antenna Our eyes are mainly for seeing something Our consciousness is for perceiving that we are seeing right now, and for seeing why An observer doesn’t just see the surface, but captures “why that moment deserves to be photographed” Relaxed task-oriented travel AI recommends tasks information about the surrounding environment during the user’s trip, some people The LLM can appropriately recommend some character states the user can really carry out A fawn running bounces and hops super fun actually it’s quite different from other animals Behind it is a bit of biomechanical cleverness. The fawn isn’t deliberately being cute; its body structure makes it walk like a spring The first trait is long, light, highly elastic legs Then, in youth the bones are not yet set, so it needs to “walk by bouncing” A bounding gait spreads impact in multiple directions, protecting growing legs and spine. In other words, bouncing is the fawn’s “protection mode” Then there’s the muscle memory of the organism itself: deer survive by speed and misdirection, and bouncing lets them generate force quickly The essence of fridge magnets and stamps almost every country has them I think it’s because experience is very abstract, so we turn abstract experience into evidence you can hold Memory is fluid what people are ultimately looking for is how to resist forgetting through meaning … stuffing the world into your own daily life People like the control of collecting People want to build a sense of identity out of a collection building,,, who am I … leaving some mark for your own future Why Japan’s water can be drunk straight In China we all assume tap water is not drinkable straight But in Japan the water sources are protected as carefully as guarding treasure The water treatment process is complex but very efficient and there is a nationally unified and strict water quality standard Pipe maintenance is active and very picky This is actually a problem in many countries: the water quality is fine, the problem lies in the transmission process — pollution caused by ageing pipes China’s urban water purification processes are actually not far behind developed countries But the old pipes are a big problem, and the water tanks of residential buildings are not cleaned properly, with no long-term maintenance And the scale of the cities is itself far larger than Japan’s, so the volume of maintenance is enormous But there are also some cities — for example Shenzhen Nanshan, Hong Kong, Macau, Shanghai Pudong, Qingdao Laoshan and so on — where drinking tap water raw is not a problem Photographing scenery vs. fridge magnets The essence = “using a photo to collect the feeling of this moment” You stand on Mount Huang looking at the sea of clouds, at Qinghai Lake looking at rapeseed flowers, by the sea watching the sunset… in that second your heart goes “wow! so beautiful!” You raise your phone, click! and you put the feeling of that second into your pocket Later, whenever you open the photo, that heart-racing feeling comes back This is completely different from “showing off how pretty I am” or “I want a good shot”; it is a quiet private collection Fridge magnets are bought, photos are taken; the essence of both is bringing home the moment of the trip that moved me most Photographing scenery: proving you were there + collecting the feeling Buying a fridge magnet: proving you were there + collecting the memory Stuck on the fridge door, you see it every day and feel happy; when friends come over they chat about it — wow, you’ve been there This is a kind of physical feeling Photographing scenery vs. photographing yourself The brain naturally favours signals related to survival and social status, and “yourself” — especially your own face — is exactly high-value information The face is the highest-priority perception object; people naturally recognise themselves, and this is also a social signal Humans are super-social animals. Modern society has further institutionalised “self-display” Then, people long to be seen; this is a way of placing one’s sense of existence Scenery is mostly just the world’s, but people care more about their own position in the world What exactly is La Niña Core: sea water in the central and eastern equatorial Pacific is colder than in normal years, and this lasts for more than several months Location matters — that is one of the most active regions of energy exchange on earth, the equivalent of the “air outlet of the global air conditioner” When it gets cold here, the whole atmospheric circulation gets rewired You can think of the equatorial Pacific as a conveyor belt: Normally: warm water is in the western Pacific (near Southeast Asia), air rises cold air sinks in the eastern Pacific wind blows from east to west, forming a stable circulation During La Niña: colder gets colder → the wind is stronger → the circulation is “pulled taut” For a while to come, China’s weather will “more easily go to extremes” Winter: cold air activity is on the strong side cold waves more easily push south the north is colder, with bigger swings Summer: the East Asian summer monsoon is anomalous rainfall in the south is above normal and more concentrated flood risk rises the north may see periodic drought Labubu’s sense of social identity A plush toy is a highly emotional object, in essence a bit like a fridge magnet Both pin abstract emotions, memories, or identity onto a stable, repeatable object, making the intangible into something physical A plush toy is an outward-facing object Some mechanisms behind why women like hanging plush toys on their everyday bags Let’s start with the deepest psychological mechanism. Plush toys usually have “baby schema”: big eyes, round face, soft material. The human brain almost reflexively secretes dopamine and oxytocin toward this shape — the same mechanism that makes people want to hold a baby, pet a cat, pop bubble wrap. This is not childishness; it’s the caregiving system that mammals have written into their hardware. Hanging a plush toy on your bag is like carrying a small trigger button for soothing emotions on your person One layer up is emotional externalization. Many women treat their bag as a “movable private domain,” and the plush toy is the flag of that domain. It may represent a sense of security, a sense of companionship, or a state of “I don’t want to be too sharp right now.” The real world emphasizes efficiency, rationality, and competition, while a plush toy is a soft counter-declaration: I allow myself to be cute, fragile, non-instrumental (of course it may also simply be that the person saw it and thought it was pretty — an emotional expression) There’s also the identity-expression factor. A plush toy is a low-risk, highly recognizable self-label. Anime characters, cartoon IPs, niche designer pieces — in essence they are all saying “who I am, what I like.” Compared with wearing a highly aggressive piece of clothing, a charm is an expression you can advance or retreat with: when you want to be noticed it’s very visible, and when you don’t want to socialize it isn’t too loud From the angle of social interaction, a plush toy is also a “de-threat signal.” It makes a person seem easier to approach and lowers the psychological threshold for talking with strangers. This is especially obvious in East Asian culture: cute is by default taken as harmless, friendly, non-confrontational. This isn’t necessarily a conscious strategy, but the effect genuinely exists One last point that is often overlooked: a sense of control. A bag is an object you touch, check, and adjust frequently every day. Adding a plush toy to it is like installing a small “anchor” in daily life. However chaotic the world gets, the bag is still there, the plush toy is still there, and order still holds within a centimeter On the social side, to a certain degree, Labubu or a Mayday plush toy determines Turning part of your life, your experience, and your feelings into an object bound to yourself It is likewise a declaration of identity: I’m a fan, so I’m on the same wavelength as this group of people, this is our secret code It’s a stable interface for emotions, and it’s also turning abstract belief into a concrete object Geographic common sense I never understood how the earth regulates its temperature, how this system works Three factors: the heat source (the sun), the insulation layer (the atmosphere), and the movers (wind and water) The hottest places and the coldest places The hottest: naturally near the equator, where the sun shines almost straight down, like sitting right in front of a heater, with energy most concentrated; besides that there are desert regions, like the Sahara. Although their latitude is similar to tropical rainforests, because there is no cloud cover to block it (clouds are like curtains), the sun bakes directly during the day and the temperature is extremely high, so desert climates differ a great deal, with huge day-night temperature swings The coldest places: the two poles, and high mountains — a high mountain is closer to the sun, but the atmosphere is like a quilt, and up high the quilt is too thin to hold in heat, hence “it’s cold high up” Why do some places never change clothes through the four seasons, while others have to rotate winter clothes and summer clothes? Because the earth spins at a tilt. In summer the sun shines directly on the northern hemisphere, so it’s hot here; in winter the sun goes off to shine directly on the southern hemisphere, so it’s cold here. This “sometimes near, sometimes far” feeling creates the four seasons The temperate zone generally has four distinct seasons; near the equator, no matter how the sun slants, it still shines on them The ocean is a big water tank, slow to absorb heat and slow to release it (good insulation). Land is stone, fast to absorb heat and fast to release it. Where there is no sea to regulate it, summer is scorching, winter is freezing, and the temperature difference is huge The Arctic is warming 3-4 times faster than the global average; ice melts and dark seawater is exposed. Ice reflects sunlight (cooling), seawater absorbs sunlight (heating), and the more heat absorbed the faster the ice melts — a vicious cycle Scientists have found a “Warming Hole” in the North Atlantic (the sea area south of Greenland) The earth has a huge “warm current conveyor belt” (such as the Gulf Stream) that carries heat from the tropics to Europe. Global warming causes glaciers to melt, and a large amount of cold fresh water rushes into the North Atlantic, which may “block” or “slow” this conveyor belt. If the warm current stops moving, northwestern Europe (such as the UK and the Nordic countries) might instead become colder in the future, even seeing extreme winters Extreme weather often happens in the temperate zone, not at the equator The earth spins at a tilt The temperate zone’s land gets “baked” by the sun 2-3 hours longer than the equator! The heat-absorbing time is longer and the heat-releasing time is shorter, so the heat piles up Near the equator there is a lot of seawater, heavy evaporation, humid air, and thick clouds. Clouds are like an umbrella, blocking part of the sunlight; water vapor is like an air-conditioning regulator, keeping the temperature from rising too absurdly. So equatorial temperatures usually hover between 30°C-35°C, and rarely break 40°C Take Turpan in Xinjiang, or the Middle Eastern deserts. Here the air is dry and there is no cloud cover. Sunlight “arrives at the ground without loss,” and since rock and sand have a low specific heat capacity (they absorb heat fast), the temperature can spike to 45°C or even over 50°C in an instant Blocks in a toy store Building a tower is not complicated in itself: the rules are simple, the feedback is immediate, and failure isn’t fatal. The real world is highly complex and unpredictable, while a block tower is a small universe you fully control Building a tower is neither pure mental work nor pure physical work; it forces you into a state where hand, eye, and brain are synchronized In psychology this is close to “flow”: attention is locked on the present, the sense of time is compressed, the self-narrative temporarily disappears A small world that can be controlled A rule-bound space that allows failure A stretch of focused time that needs no language You know the tower will fall in the end, and still you earnestly stack it up, block by block AI fitness influencers are virtual characters created with artificial intelligence, meant to share fitness advice, training plans, motivational content and lifestyle content The mindset of a game Civilization architect All philosophy, religion and systems of knowledge are essentially not “truth,” but “tools” or “apps” that humans invented in order to survive and to live well No app is absolutely “good”; you can’t say “WeChat” is better than “Excel,” because they solve different problems The Dao follows nature The “Dao” is precisely that — the “wisdom” of knowing at what time, on what occasion, to pick up which tool Not clinging to any one tool, flowing through different containers like water: this is “the Dao follows nature” Do shorts harm your body? Cold starts from the feet, the knees are the “weak link”: in traditional Chinese medicine, the knee joint has little subcutaneous fat and lacks protection, making it an important gateway for cold to invade the body Chinese elders often say “if you don’t keep warm when young, you’ll have cold legs when old” Modern medicine says that cold alone does not directly cause osteoarthritis. Arthritis is more often caused by cartilage wear, aging, obesity or trauma Cold will trigger or worsen existing joint pain. Low temperatures make blood vessels constrict and blood circulation worsen, stiffening the muscles and soft tissue around the joint, which makes a person feel pain. So the elders’ experience isn’t entirely unreasonable; the causal relationship is just explained differently in medicine The essence of film It feels like film is increasingly converging on a few important works The competition in film is fierce People have less and less patience to sit through a whole film But humans have to need a medium like film Film isn’t just entertainment; it’s more like a “cheap and efficient expansion of lived experience” Film lets you go through other people’s suffering, glory, and choices in “first person” Film always resonates with the most essential emotions inside a person This release of emotion can have a therapeutic effect; those emotions are shared by all humans — I’m not alone By observing characters I reflect on my own bottom line of values, and the trained way of thinking lets me face complex decisions in real life more clear-headed and firm And Cameron’s aesthetic is really good — the beautiful nature humans long for; not tech skyscrapers but nature If you only look at the effects, that’s just an adrenaline thrill; if you think about the logic of a character’s transformation, that’s growth Fewer and fewer modern people walk into a cinema, turning instead to short dramas and short videos Dopamine’s immediate rewards are faster, more looping, with more reversals Modern people’s emotional labor Film is the last stronghold of deep resonance; film is hard to replace We modern people find it hard to have a full, uninterrupted 3 hours But we have countless 5-minute chunks, and short drama fills those fragments perfectly Film is getting more expensive and scarcer; short drama can give you pleasure but rarely can it shake you. You won’t finish a short drama and stay restless for a long time, thinking about humanity’s fate or the meaning of life People watch short drama because life is too bitter and too fast; and we still need film because we still want to look up at the stars when life lets us Short drama is actually fast-food culture that can also have depth, and it may also spark your thinking and feeling “Escape from the British Museum” is only a short film, yet it broke countless people. Even short drama, as long as its core is deep enough and its emotion sincere enough, can have force no less than a big film. The reason this short film broke so many people is precisely that in an extremely short space it touched humanity’s grandest and softest subject In the places that can touch you — in this era, the power to deconstruct meaning is no longer held only in the hands of big Hollywood directors With enough sharp insight and sincerity toward culture, an ordinary creator can also use the “short” form to start a global cultural discussion. It proves: short doesn’t have to be a synonym for “shallow” — it can be “distilled” expression We don’t have to feel pessimistic about culture because short drama is popular. What we should really be wary of isn’t “short” but “hollow.” As long as a creator holds reverence, even 5 minutes can carry the weight of a thousand yearsA good photographer is essentially a good observer
AI-recommended task-oriented travel
The biomechanical protection behind a fawn’s bouncing
What fridge magnets carry as memory
What lies behind drinking tap water straight
A photo is a private travel collection
Photographing scenery vs. photographing yourself — where existence is placed
La Niña and Extreme Weather
The Social Identity and Emotional Externalization of Plush Toys
How the Earth Regulates Its Temperature
Why Extreme Weather Often Happens in the Temperate Zone
The Flow Experience of Store Blocks
The Rise of AI Fitness Influencers
Philosophy and Religion Are Tool-Apps for Human Survival
Examining “Shorts Harm Your Body” in Chinese and Western Medicine
Film as an Expansion of Lived Experience
The Emotional Value of Short Drama and Film
9. Content, Craft and Recording
7 entries Fun hardware I’m thinking maybe it could be assembled into a QR code — the QR code on the back of every fridge magnet, and this QR code can be scanned, and scanning it takes you into the author’s story continuation That way it feels like it has more physical commemorative value, and there’s an entry point to archive the memory China has long had an implicit moral narrative: “suffering is honorable,” “a soldier should be able to endure,” “a poor kid gets more tempering” This kind of talk sounds positive, but the actual effect is: it rationalizes why the people who do the suffering are always the same group of people He repeats conclusions, yet can’t explain them His views conflict with each other, yet he doesn’t know it This is a problem of cognitive structure, not of personality, and most people can’t see this layer Every kind of state a person is in has a corresponding path to happiness I think too much information + not enough cognitive capacity → negative effects So it has to match your “base coat” That holds, but what’s more fundamental is this The problem isn’t “seeing a lot” — it’s that “the source and the structuring of what you see are passive” A lot of fragmented views and experiences you come into contact with, other people’s narratives, rather than causal chains you lived through yourself, or systematic theoretical training The right to produce information matters a lot When a person: doesn’t take part in generating meaning can only consume “insights” others have already packaged then no matter how many high-end ideas or profound views he encounters, they all turn into: slogans collage or identity decoration Are you being “inputted into,” or are you “building your own model”? So a lot of questions can be naturally rearranged: Why are some people with little formal education extremely steady Why do some elites get lost instead Why “travel” and “seeing the world” are ineffective for many people These are all questions of subjectivity So-called “subjectivity” isn’t attitude, isn’t confidence, isn’t wanting or not wanting — it’s whether a person has the ability to “generate explanations from experience, then use those explanations to correct experience in turn” Whether you have this loop determines whether experience is nutrition or noise One kind is: The world as a source of stories You listen, you watch, you record, you retell The other is: The world as a field of problems You observe, you hypothesize, you verify, you correct This determines a very important premise Information is an external flood, and a person is just a container Do you have the ability to turn what you encounter into a model The essence of eyes going nearsighted during a cold The visual system gets temporarily pulled off balance The eye is essentially an optical system + a nervous system When you have a cold, an infection, or inflammation, several things happen in the body: Inflammatory factors rise The autonomic nervous system leans toward a “stress state” The ciliary muscle is more likely to be tense or spasming The result is: The eyes get better at seeing near and worse at relaxing to see far The nasal cavity, sinuses, and eye sockets are anatomical “neighbors.” When the nasal mucosa swells and sinus ventilation worsens, it causes: Changes in micro-pressure around the eye socket Tension in the extraocular muscles A feeling of swelling, heaviness, achiness in the eyes Film and TV screenwriting format American screenplay format, leaning toward spec script (for submission), not the shooting script used on set One part is the scene heading (Scene Heading / Slug Line) INT. / EXT.: interior / exterior Location: CUB’S APARTMENT Time: DAY / NIGHT (only the broad time, not “two thirty in the morning”) Its purpose isn’t for the reader — it’s for the producer, director, production assistants, and lighting Action / Description Cub stares intensely at his laptop screen… Present tense Third person Only write what can be seen / heard No psychological analysis, no abstract feelings A novel writes “why”; a screenplay writes “what happened” The Tulkun’s “vow of nonviolence” isn’t nature — it’s a choice The Tulkun once took part in war The Tulkun later actively gave up using violence That renunciation was elevated into a moral law across generations If a species has never walked that road it won’t bind itself with a form like a “vow” Vows only appear in civilizations that have made mistakes The Tulkun have: a highly abstract language epic-level narrative ability complex social relationships longevity and cross-generational memory This means one thing: They fully have the cognitive foundation to develop a technological civilization But they don’t have: weapon systems traces of industry expansionist infrastructure It’s not “can’t do it,” it’s no longer doing it Death itself isn’t what the Tulkun can’t bear, what they can’t bear is — turning life into a tool for resources If I go down this path in the future I’ll inevitably have to do IP and short video but that’s also an interesting part of the world haha advanceExtending the story behind the fridge-magnet QR code
The Hidden Moral Trap of the Suffering Narrative
The Right to Produce Information Determines Whether Cognition Holds Together
The Physiological Link Between a Cold and Nearsightedness
Screenwriting Format Conventions
The Tulkun’s Vow of Nonviolence
An Interesting Future Attempt at IP and Short Video



Responses