Pioneers Insight Method Research Author
How a CEO Learns Not to Be So Anxious | Eight Years of Duozhuayu with 猫助, the Darkest Hours and New Plans
Back to Episodes

How a CEO Learns Not to Be So Anxious | Eight Years of Duozhuayu with 猫助, the Darkest Hours and New Plans

Summary

  • Duozhuayu’s long-term value lies not in used books, but in reorganizing low-residual-value, nonstandard household goods into tradable supply. Its categories have expanded from books and clothing to toys, children’s wear and outdoor gear, with small furniture next; physical stores handle experiences and information displays that are uneconomical to deliver online. The direction is to let young people new to a city, short on cash but unwilling to “settle for a makeshift life,” buy the basics of their material and spiritual lives in one place. 猫柱 sums it up as a “tech-enabled junk collector.”

  • 猫柱’s view of consumer AI is notably restrained: the technology can improve retrieval, but may not create enough paid value to cover inference costs. Duozhuayu tested vector search, clothing image recognition and a “million-book inventory librarian” based on users’ purchase histories. In practice, a human operator—the “person behind the account”—was better at inferring a reader from sparse records, organizing the answer and satisfying the user. Model recommendations were neither interesting enough nor proven to lift repeat visits or sharing.

  • AI has replaced some of 猫柱’s use of Google and Wikipedia, as well as some demand for buying a book “just to look up a few facts,” but high-quality Chinese knowledge and deep books have not lost their scarcity. Chinese content is far less digitized and text-based than people imagine; demand for deep academic books has changed little, while supply keeps shrinking as titles go out of print and surviving copies deteriorate. 猫柱 therefore sees the opportunity shifting from “generating more information” to filtering, preserving and republishing it: “The people who help you block information are becoming more important.”

  • Out-of-print books are a niche business with real demand but extremely difficult unit economics. Some titles have intense demand among fewer than 2,000 people, still not enough to cover Chinese ISBN, reprinting and copyright costs. Connecting authoritative book archives to models faces the same problem: “The cost is high, demand is low, and users’ willingness to pay is low.” Duozhuayu plans a curated space at Beijing’s Langyuan: new publications unlikely to become hits on the first floor, and a reading room of selected out-of-print books for longtime users on the second.

  • Duozhuayu survived the turn in capital markets not through continuous fundraising, but through the cash-flow elasticity of its secondhand business. Its last funding round was in 2018, bringing total financing to roughly RMB100M; annual revenue later exceeded RMB100M and is now above RMB200M. When dollar financing suddenly became more cautious in 2021, the company had to widen purchase-sale spreads that had once been too thin, using operating profit to cover R&D, headquarters and store-opening costs. It sacrificed growth, absorbed user criticism and eventually turned financially profitable.

  • 猫柱 defines founder maturity as acknowledging conflicting interests and designing the company first as an organism that can survive without the founder. Employees, users, shareholders and the organization may appear aligned during rapid growth, but cash-flow and survival pressure expose the conflicts; team members are only accompanying the company for part of the journey—“This is your destiny, not theirs.” Her current ordering is: “The company’s greatest responsibility is to take responsibility for itself,” and “It is a company; it is not a private company.”

  • 猫柱 became less anxious not because she regained certainty, but because she stopped chasing every AI update and grand narrative. The 2022 lockdowns, conflict with her partner, the company’s shutdown and the deaths of family members shattered her old confidence. Her current approach is to treat AI as a factor of production—“When the cost is low, buy the dip; when the cost is high, wait”—and accept that changes beyond anyone’s control do not require personal accountability: “Don’t make yourself the protagonist of everything.”

  • The next increment is more likely to come from AI lowering the cost of creation for small teams than from forcing AI into Duozhuayu. 猫柱 wants to test high-quality Simplified Chinese content, digital publishing and the revival of out-of-print works overseas, using AI for proofreading, translation, typesetting, simple development and article criticism so that “people with ideas can practice them more lightly.” Her broader labor-market view is that highly standardized elite bureaucratic jobs may also be exposed, as AI challenges white-collar workers and the identity narratives of institutional elites.

Deep dive

1. Duozhuayu Is Moving from a Used-Book Platform into Offline Life

  • 猫柱 compresses eight years into one sentence—“These eight years have been about making Duozhuayu”—and jokingly calls the company a “tech-enabled junk collector”: it does not transfer luxury assets, but collects the everyday goods ordinary people genuinely leave unused and waste.

  • The category mix has expanded from used books and clothing to toys, children’s wear and outdoor goods, with furniture next. The core question remains the same: can ordinary clothes from Uniqlo, and books once destined to become waste paper, be used again?

  • The first reason for moving offline is cost. Describing and photographing every small piece of furniture online would make information production expensive; browsing a physical store gives consumers an immediate sense of the object as a whole.

  • 猫柱 imagines stores serving people who have just arrived in a city and received their first paycheck, have little money but do not want to “settle for a makeshift life”—the drifters in Shanghai and Beijing who can buy the basics of their material and spiritual lives in one trip.

2. The AI Book-Recommendation Experiment Lost to the Human “Person Behind the Account”

  • Duozhuayu already uses vector models for search and tested Alibaba’s image model in clothing to identify products from pictures. 猫柱 does not deny that retrieval accuracy improved, but stresses that every commercial application ultimately comes back to cost.

  • The original premise for book recommendations was compelling: books are organized from the author’s perspective, while readers usually have only a vague desire to “learn about Tang history” or “Chinese history.” If the platform knew what a user had bought before, it could theoretically summarize the subject at that reader’s level and recommend suitable titles.

  • The team designed its mascot as a “cat librarian” with command of a million books in inventory. When a human operator played the “person behind the account,” they would first browse the user’s page, assess their reading level, find books in the back-end database and adjust the wording. Users were very satisfied with the answers.

  • With a model, the system could retrieve books but struggled to infer “what kind of person this is” from only 1 to 10 purchases, much less explain convincingly why a particular book would help. The quality gap was not yet large enough to support commercialization.

3. Personalization May Not Be a Mass-Market Need; AI Looks More Like a Production Tool

  • The experiment changed 猫柱’s previous view: most consumers may not need highly customized recommendations. They just want to know “this is the best one available,” and for either fiction or history, a relatively clear answer is often enough.

  • In her view, the people willing to pay for ChatGPT Plus and call models frequently are mostly producers: compiling content, analyzing material, writing product documents or giving a product manager who cannot code the equivalent of a “junior engineer.”

  • Thus, even if individual producers are treated as a B2B customer, she still sees AI primarily as “a productivity accelerator,” not a product for amplifying mass-market entertainment consumption. Duozhuayu has long used it internally on the B side; its C-side exploration is “semi-stalled.”

4. AI Has Eaten Knowledge Retrieval, Not Deep Reading

  • Responding to 荣慧’s questions about transaction data, 猫柱 first points out that high-quality Chinese content is far less digitized than people imagine. Many materials have no PDF, and even when they do, they are often just scanned images that have never been converted into text.

  • She cites early GPT as an example: it could even mix the Twelve Beauties of Jinling with characters from online novels. In purely Chinese knowledge domains, model answers are often poor; knowledge from imported books in the English-speaking world is easier for AI to dissolve because the original content is not scarce.

  • AI has “completely replaced the time I used to spend on Google and Wikipedia.” Dictionaries, atlases and knowledge-retrieval books had already been displaced by successive waves of technology, while literary paper books had long faced competition from web fiction. But demand for truly deep, scholarly and well-produced books has not visibly changed.

5. The Scarcity of Out-of-Print Books Is Real, but the Business Is Too Small

  • The problem with out-of-print books is not a lack of readers, but that “demand among fewer than 2,000 people can be extremely strong” and still not support a 100,000-copy business. Chinese book prices are low and ISBN costs high, so publishers often ask, “Why should I print 2,000 copies for you?”

  • Duozhuayu has revived out-of-print books and launched reprints several times, and has considered digitization, but copyright relationships are complicated. Physical books also keep deteriorating, while collectors buy them and do not replenish the supply, so quality inventory can only shrink.

  • Another idea was to let a model answer questions only from a book collection or authoritative academic-paper archive, giving users credible knowledge without requiring access to the full book. 猫柱’s commercial judgment remains negative: “The cost is high, demand is low, and users’ willingness to pay is low.”

  • She envies the low-barrier, high-ticket niche publishing markets in Japan and the US: a one-page folded pamphlet can carry an ISBN and sell for $10-$20; rigorous illustrated specialist books can sell for $20-$30. She believes this kind of “high-ticket, small-audience” cultural business is difficult to sustain in China.

6. Once Information Is Infinite, Curation Becomes Scarcer Than Generation

  • 猫柱 contrasts her childhood hunger for information with the present. In the past, she would start visiting bookstalls on the 12th to find a magazine published on the 15th; today, with information effectively infinite, the scarce resource for consumers is roughly 2 hours of attention a day.

  • For out-of-print books, she set screening criteria: an initial pool of more than 20,000 titles was narrowed to perhaps 10,000 after a rough screen, then to several thousand through close selection. Books that rose in price merely because older readers insisted on paper, or that had equivalent information online, were excluded. The survivors had to possess “fair, objective value.”

  • The Langyuan plan in Beijing has a clear curatorial structure: the first floor will introduce new publications unlikely to become Douyin hits but carefully edited and translated, while also seeking to import niche English books through China National Publications Import & Export Corporation. The second floor will be a reading center for selected out-of-print books, open to longtime Duozhuayu users.

  • 猫柱’s key judgment is that “the greatest barrier to information is inertia, not a wall.” Algorithms keep recommending new material in the same field; an editor’s value is to make someone suddenly realize that “there is another world.”

7. Eight Years of Survival Was First a Result of Personality, Not a Growth Myth

  • 猫柱 is not quick to act. She will turn an idea over in her head for a long time, and marketing concepts often go through dozens of drafts. The confidence she projects often comes from having ruled out every other option she could think of.

  • That caution made the company “very healthy but very slow.” She also cares deeply about having a life of her own and does not want employees to look back on a career and remember only “overtime and documents.”

  • She hopes employees retain genuine friendships with colleagues, transferable skills and a clearer sense of what suits them. One sign is that many former employees did not return to conventional jobs but began projects of their own; colleagues have also formed small, enduring circles of mutual appreciation.

8. Users, Employees and the Company Form a Triangle That Cannot Be Held Forever

  • 猫柱 believes the most important thing a company must deliver to employees is not the phrase “we built the company together.” Once employees have delivered service to users, the company must provide salaries and corresponding value in return; it cannot pass off employees’ delivery to users as the company’s delivery to employees.

  • The same logic cannot resolve user-side conflicts. Sellers want broad acceptance criteria; buyers want low prices and authenticity. But niche books cost too much to find their next buyer, so Duozhuayu cannot accept everything. It can only acknowledge the inconvenience and accept user criticism.

  • She calls shareholders, employees and users the “three stakeholders”: users provide money, employees deliver the service, and the organization must retain cash flow to survive and grow. The three objectives are not naturally aligned; the conflicts are simply harder to see in good times.

9. The Secondhand Business Depends on Social Affluence and Time Arbitrage

  • 猫柱 had always liked secondhand goods but waited for social conditions to mature before starting a company. After the first startup boom from 2010 to 2013, and another full cycle of product use, residents of first-tier cities had accumulated enough “good things sitting idle.” Having idle goods itself meant they had moved beyond subsistence.

  • Xianyu proved that C2C demand existed, but a specialist store could solve three problems individual sellers struggled to handle efficiently: setting a fair price, guaranteeing authenticity and cleaning and quality-checking intimate everyday goods, thereby creating trust between strangers.

  • C2C requires one logistics payment; C2B2C requires two. Duozhuayu had to prove that the efficiency gained from centralized pricing, inspection and cleaning saved users enough time for consumers to give up the spread they could otherwise capture themselves.

  • That is why Duozhuayu users place a high value on their time. They may be capable of selling an item themselves; they simply have to decide whether researching market prices, setting a price, cleaning and authenticating the item is worth the extra RMB10.

10. Startup Opportunities Come When an Old Cost Suddenly Falls

  • Duozhuayu also caught a technology shift at the moment it was founded: contracts and settlement for reverse logistics from C to B were becoming standardized, high-resolution phone cameras made scanning, photographing and direct uploading routine, and bandwidth was a secondary factor.

  • 猫柱 compares Xianyu with early Taobao. Individual sellers once had to buy a separate camera, transfer images to a computer and then upload them; today, the phone camera has eliminated that entire block of operating cost. Technology changed not the demand, but the transaction steps people previously found too troublesome.

  • Her advice to founders is therefore concrete: look for models people considered in the past but could not get running because one step was too expensive, then ask whether AI genuinely lowers that cost. The collapse in the batch cost of translation, and the ability of people without programmers to build independent products or front-end demos, are clear examples.

  • But Duozhuayu already has a long business logic and a large codebase, so simply layering new code on top with a model is difficult. New technology is a factor of production, not magic that automatically cuts through organizational complexity.

11. The Value of Funding Is Allies and Information, Not Just Cash

  • Duozhuayu’s early funding was mostly in dollars and did not come with especially harsh exit terms. If a shareholder no longer wanted to continue the journey, 猫柱 was willing to find a way to buy the stake, because she herself still believed in Duozhuayu’s equity.

  • The company raised roughly RMB100M in total; its last round was in 2018. Revenue later exceeded RMB100M a year and is now above RMB200M. She describes secondhand goods as a kind of “something-from-nothing” business, which is why the company did not raise that much money.

  • 猫柱 wanted to “raise the least money possible” in every round because money eventually runs out. More important were the investors’ new perspectives, strategic partners and business information unavailable online; competitive information can never be fully transparent.

  • Early on, a friend introduced her to the investor 柯羿. After roughly 15 minutes of conversation, she received funding. 柯羿 often reminded her that she needed “to become a mature entrepreneur.” She had never founded a company before and, of course, was not mature.

12. Maturity Means Accepting That Entrepreneurship Ultimately Belongs to the Founder

  • Another sentence that stayed with 猫柱 came from 香敏: “Entrepreneurship is always your own business.” Companions come along with the company and founder for part of the journey; they do not treat the same organization as their own destiny.

  • In her early years, she expected the founding team to stay together forever, everyone to be friends, and employees to receive generous benefits and protection indefinitely. Only after the pandemic nearly exhausted the company’s cash did she admit that such promises might exceed the organization’s capacity.

  • Her conclusion is not to dismiss friendship, but to separate responsibilities: the company should “help people succeed” when it comes to employees, but first “get the work done” when it comes to the organization. Employees can choose to leave; the founder must answer why the company exists and how it will survive.

  • 猫柱 still values relationships among colleagues, but no longer tries to stuff her personal need for belonging into the company’s objectives. She acknowledges that she was lonely as a child and longed for companions, but that emotional need and organizational survival cannot always be served by the same structure.

13. The Company Must Be Designed as an Organism That Can Replace Its Founder

  • 猫柱 has long imagined the company as “an organism that can take in and expel resources and operate by itself”: it obtains resources from the outside, produces value that improves its own survival and continues to iterate after the founder exits.

  • She draws the boundary sharply: “The company’s greatest responsibility is to take responsibility for itself.” The company does not exist for the founder, a particular employee or a single shareholder. “It is a company; it is not a private company,” and the founder herself should be replaceable.

  • The company as a vehicle can be used to make money, make friends, carry a group’s collaboration or build an asset that compounds in value over time. There is no single correct answer, but the first motive must be clear: are you helping the work succeed or helping people succeed, helping others or helping yourself?

  • She uses nineteenth-century Scottish distilleries as an example: a distillery can close, reopen and change owners; the owners may not even know how to make whisky, yet the asset can continue appreciating. An organization’s quality may conflict with the personal aims of an owner who holds 100% of the shares, which instead shows that the asset has an independent life.

14. Good Times Conceal Conflicts of Interest—and the Founder’s Ignorance

  • 猫柱 believes that during rapid growth, accomplishing the work, developing people, maintaining relationships and delivering shareholder returns can all happen at once. Once blame arrives, stakeholder interests reveal their divergence, regardless of how close the relationships were before.

  • In Harvard courses, she saw people analyze corporate management through motives, interests and key drivers. East Asian narratives are more likely to explain organizational failure as “this person was good and that person was bad,” while avoiding the pain and disorder inside failed companies.

  • But classic American cases cannot simply be copied either. Every company ultimately has to answer its own question: does it primarily seek money, friendship, a company built to last or an asset that can sustainably appreciate?

15. Duozhuayu’s Early Run of Luck Delayed the Founder’s Learning

  • Duozhuayu’s first demo went viral immediately, with almost no product pivot. Fundraising also went so smoothly that 猫柱 never wrote a pitch deck; it was more like she was choosing investors. She does not see this as extraordinary talent, but as a case of the right candidate appearing just as the secondhand market was taking off.

  • Pop-up stores, expansion into clothing and offline experiments before the pandemic all received strong feedback. After the market recovered in the second half of 2020, opening stores and adding categories remained smooth.

  • The good times led her to say something that now looks “too arrogant”: “If you face a dilemma where it’s either A or B, it means you don’t have enough ability to find C.” Reality later showed that some constraints have no C for an individual to invent.

16. The 2021 Capital-Market Turn Forced a New Profit Model

  • Duozhuayu’s offline stores had relatively short payback periods, so the team wanted to raise money to accelerate store openings instead of waiting for one store to pay back before opening the next. The Sanlitun store, however, became a poor investment because of high rent and its prolonged inability to operate in 2022.

  • After the Didi incident in 2021, dollar investors became abruptly more cautious. The company was not short of cash and individual businesses were profitable, but once headquarters R&D and new investments were included, overall cash flow remained negative. The old expansion model could not continue.

  • Financing had previously subsidized R&D and store openings, allowing Duozhuayu to keep purchase-sale spreads very low. Once the company had to fund itself, it needed wider spreads and operating profit to cover all costs. Growth slowed, and changes in acquisition and selling prices brought user criticism for a time.

  • 猫柱 is glad she chose a business with “a lot of elasticity”: prices, categories and the pace of investment could all be adjusted. As revenue grew, spreads could gradually narrow again, but the interim adjustment could not be avoided.

17. The 2022 Shutdown Turned Operating Problems into a Survival Problem

  • After the capital-market turn, 猫柱 was handling the business, fundraising and cash flow at the same time, while also entering into conflict with her co-founder. She does not elaborate, but admits she no longer had the capacity to manage external changes and team relationships simultaneously.

  • Just as the finances had been adjusted into the black, mass lockdowns shut warehouses, factories and stores, and logistics could not ship. The company had no revenue. Her daily work was no longer creation and growth, but “logistics is cut off here, and that warehouse has been locked down again.”

  • The factory had several hundred workers, many outsourced, but Duozhuayu continued paying them when work could not resume because low-income workers were more vulnerable during the pandemic. The company kept spending every month while unable to restore normal cash flow.

  • When friends who ran physical businesses took stock together, many companies had only around 6 months of cash left. 猫柱 did not give up on Duozhuayu. What truly shook her was the question: “Can I still do anything in this society?”

18. When the Trampoline Broke, She Learned to Fall for the First Time

  • 猫柱 uses a trampoline to describe the experience of the era: all her past experience had been about working hard to jump upward, and she had never studied what lay below. When a large hole suddenly opened in the trampoline, everyone began falling and finally saw that an abyss had always been there.

  • She grew up in an era when rules were gradually becoming clearer and commercialization and professionalization were advancing. She was used to believing that following the rules could create improvement. The pandemic made the rules people relied on to organize their lives and businesses suddenly “not work.”

  • The heavier blow came from family. Her maternal grandparents, who had raised her, died one after another. She even had to rush to take a COVID test so she could bring the result back to Shenyang for the funerals, and she never saw either of them one last time. The company, family and society all seemed to collapse at once.

  • This led her to ask what was necessary. Staying alive is not the only purpose; people also seek meaning and happiness. Her summary of that period was, “Everything I used to believe in died,” made harder by her habit of blaming herself first for everything.

19. Traveling the World Repaired Her Market View, but Did Not Restore the Old Narrative

  • At the end of 2022, unable to tolerate her circumstances any longer, 猫柱 began traveling around the world. She found that every country had its own problems, and found a “shared community of fate” in other developing countries. Her overall view of the Chinese market became less pessimistic instead.

  • After she returned, Duozhuayu operated smoothly through 2023, 2024 and last year. The business was broken into stable processes that could continue moving forward without the boss spending all day in meetings. She had finally built the “machine” she once imagined.

  • But once the machine existed, she could not find a new personal narrative. She was no longer certain that she had to influence society or reshape how a group of people lived. The company had matured, and she had become an entrepreneur who was at least no longer naive, yet she sighed: “The mature world, the adult world, is really boring.”

20. What She Lost Was Not the Dream, but Blind Confidence

  • 猫柱 distinguishes between vision and belief: “What I lost was confidence, not the dream.” She still has things she wants to do, but no longer believes she can naturally rally a group, bend reality and turn a grand goal into fact.

  • She even believes the foundation of charisma is often “blind confidence without any basis.” A product may be ordinary, but if its creator shows intense passion and conviction, onlookers may begin to think it might genuinely have value.

  • The shocks of 2022 shattered her old self-structure. She describes herself as still “gathering the fragments of my soul scattered across the universe,” forming a new self that will not return to its old shape.

  • The “ordinary person narrative” has also relieved her of excessive responsibility. She is now willing to admit that she is “just a person who isn’t particularly capable, running a small company,” and no longer makes herself endlessly responsible for changes across society.

21. Pain Cannot Be Avoided, but It Can Become Deeper Experience

  • The host invokes Drew Houston’s experience as a hero narrative: after Dropbox went viral and faced attacks from major tech companies, he became depressed and hid on a small island in Hawaii. He later admitted that weaknesses in his character had affected the company, and realized that “you can’t escape pain; you can only move closer to it and feel it.”

  • 猫柱 accepts this third perspective. She has always understood life as a pursuit of broader and deeper experience, so joy, grief and hardship all become part of that experience. An easier road is not necessarily better; it is simply difficult to see that way while in pain.

  • She now states that “the first pot of gold is basically luck hitting you.” One success does not prove that you understand the rules. Someone may be made a winner by 20 years of good luck, but she refuses to substitute results for an honest assessment of whether she has actually grown.

  • From 2022 through early 2024, she barely laughed. After recovering her footing, she came to like herself more. The old romance and blind charisma have diminished, but her core no longer swings with success or failure: “At the same time, there’s nothing left that can destroy me.”

22. A Sense of History Returns Personal Setbacks to Their Cycle

  • 猫柱 acknowledges that “happiness is always relative,” while success narratives turn life into a series of checkboxes. When all the expected boxes have been checked, a midlife crisis often comes from not knowing what should go in the next one.

  • Reading history changed her frame of reference. Being born in China as a member of the post-1980 generation, entering a 985 university and moving from an agrarian society into the ranks of the global elite in a single step was itself an extraordinary stroke of luck. Why should an 80-year life proceed smoothly, without illness or disaster?

  • She likes the 20th century. A relative born in 1903 could still tell stories about the Qing dynasty and “the Japanese invaders entering the village.” China went from kang beds and collecting cow dung to internet life in only a few decades. Readers see history as grand and sweeping; people living through it experience pain every time.

  • She recommends Towards the Republic, especially its sense of fate embodied by Li Hongzhang: knowing that the Qing government was a paper house, he could only keep patching it as a “plasterer.” “Each generation has its own fate.” Age, beliefs and historical timing explain choices better than judgments about good and bad people.

23. Not Chasing Every AI Update Is Her New Management Discipline

  • 猫柱 admits she was anxious at first. When GPT first appeared in 2022, she heard Google product managers in Silicon Valley call it “pretty stupid.” By 2023, its capabilities were changing so quickly that just as she understood one thing, another batch of new developments appeared.

  • Her final choice was to stop chasing and return to what she understands: social structure, supply and demand, costs and the combination of resources. “When the cost is low, buy the dip; when the cost is high, wait.” If nobody can keep up with everything, the market cannot operate at a level beyond the average ability of all participants.

  • AI has become her “close friend” in writing. She still gathers the material and writes the first draft herself, then asks a model to identify doubts and places that are unclear. She does not follow every suggestion, but in this respect a model can make criticism easier for her to absorb than criticism from human friends.

  • Her new plan is to return to content, possibly producing high-quality Simplified Chinese content overseas under a new name, while experimenting with digital publishing and reviving out-of-print books. AI can assist with translation, proofreading and typesetting; the code is not complicated and can be outsourced as needed, bringing one person closer to a multimedia-magazine-style “super individual.”

24. AI May First Shake the Identity Structure of Elite Bureaucrats

  • Most people 猫柱 met in Harvard’s leadership course were conventional elites from government, NGOs, major consulting firms, investment banks and giant corporations. They had spent their lives advancing through fierce competition and checking boxes one level at a time. Entrepreneurs were rare; Duozhuayu already counted as one of the larger startups in the class.

  • Their main interests were US foreign policy and AI. 猫柱 felt that Kennedy School discussions leaned left, with many guests having worked for Biden or Obama. Criticizing Trump was close to politically correct in elite circles, but that did not explain why he was elected or how different countries should respond next.

  • Many participants used AI competently in daily life but were deeply worried about governance and unemployment. Some did not even know in what respects DeepSeek was good; they only asked “which one is better,” ignoring differences in capability, cost and speed. People actually building or investing in AI startups were more neutral or optimistic.

  • 猫柱 uses Max Weber’s bureaucracy to understand the fear. Organizations strip away human agency by standardizing authority, responsibility, information aggregation and command distribution; they have always tried to train people into machines. If AI can better perform the work that humans previously handled, the people most exposed may be white-collar workers with highly abstract, precisely divided jobs—and even the elite class.

25. Once the Fear of Survival Is Overcome, There Is Room to Find the “Drive of Love”

  • 猫柱 recommends calculating a minimum living floor first: the cost at which one can still survive, and how to live relatively happily above that floor. She likes farming and gardening herself, and at the worst point would be willing to return to an agricultural life. Knowing the floor keeps survival from becoming an infinite fear.

  • Many elites work far removed from food, clothing, housing and transportation. They know only how to sustain life through money and purchases, so once their income path is threatened, their sense of security disappears. Returning to work connected with actual survival can also restore an understanding of human agency.

  • A sense of probability provides a second buffer. Elites capable of winning through multiple rounds of competition are already winners in that competition and need not be excessively afraid. If not even one person can find work that allows them to survive, human society is facing a very serious problem; there is no need to interpret a systemic failure as personal failure.

  • Only then can fear give way to happiness: ask what you genuinely want instead of merely fearing a fall from the top of the pyramid. 猫柱 summarizes the two fundamental drives as “the drive of love and the drive of death”: first overcome the fear of death, then look for the drive of love.