Vol.68 Is VC Methodology Really “Praying for Rain through Ritual”? — Guest on Mayfly World
Summary
庄明浩把 VC 压缩成“输入一千个项目、输出五个 1”的黑盒，所谓方法论更像人在巨大不确定性中反复执行的“祭祀求雨”。 A method still exists, but it is often entangled with an institution’s people, chemistry, decision-making process and a sudden flash of inspiration, making it difficult to reduce to a replicable SOP. “How do you prove you’ve actually mastered this thing? It’s too hard.”
ByteDance and Pinduoduo were not missed by people with no logic; they were precisely misjudged by the old logic. ByteDance was classified as a “news app” trapped in brutal competition, reliant on ranking manipulation and lacking a business model. Investors failed to see “content, connection and algorithms.” Pinduoduo was labeled a fresh-grocery e-commerce business that “sold fruit by creating groups on WeChat.” Funds’ past losses of tens of millions of dollars in vertical e-commerce directly suppressed positive judgment. “Fresh groceries” should have been the last label, but it was placed first.
The advantage of consistently successful institutions comes more from error correction than from making a more clairvoyant first call. IDG once made a blanket decision to invest in post-90s founders;庄明浩’s rough review found that “six or seven out of seven or eight made it.” After Sequoia saw Matrix Partners had invested in roughly 50 mobile-internet companies, 沈南鹏 asked 张颖 to select 10 for Sequoia to co-invest in—all of them, including Ele.me and Momo. The scarce capability is a “revisit and re-report” mechanism that is willing to lower its head and reexamine old cases.
Early-stage investors see only a random cross-section of a company’s history; founder growth and market shifts can rewrite the ending. The Pop Mart庄明浩 saw during its NEEQ phase was merely an “uninvestable toy concept store,” but 王宁 later changed the company. Livestreaming also worked as a phase-specific thesis and provided strong exits, even though short video ultimately capped its ceiling. “If you always want to see the ending, you won’t be able to do many things.”
AI has brought massive dollar liquidity, a grand narrative about humanity’s future, and professional pressure for VCs who feel that “if you don’t invest, you don’t deserve to be called a VC.” He uses Cursor as the yardstick: the product had only reached version 1.0 and the team numbered just dozens, yet it raised $900M at a $10B valuation. The poker chips seem to have “gained an extra zero”; “making one big bet” today may equal building an entire new fund in the past. But the metaverse is a reminder that some sand remains just sand.
The mismatch between fund duration and evaluation cycles may be more fatal than the investment judgment itself. 汉洋 asks why funds with five-to-eight-year lives require investment managers to deploy capital that same year or lose their bonuses, putting “patience” in quotation marks. Counterexamples include 徐新’s 28-year fund, 程伟’s small evergreen fund and Matrix’s first RMB evergreen fund of roughly RMB200M. 庄明浩 estimates that state capital’s share of China’s RMB primary market rose from more than 50% around 2018–2019 to possibly more than 80% by recording time, while “not causing state-asset losses” is structurally at odds with venture capital’s possibility of going to zero.
When diligencing investors and founders, don’t listen only to a polished logic; ask how they use products, admit misjudgments and decide when to revisit old cases. Some AI investors have not even bought the $20-a-month ChatGPT subscription, yet publicly lecture on “the art of AI investing.” 庄明浩’s own H5-engine project showed that when investors first paint an optimistic picture, founders seeking financing will perform against that expectation, obscuring the other side that actually determines failure. “Anything that is consensus is too flat.”
“The Art of Dragon Slaying” is both a professional metaphor for the previous generation’s internet knowledge losing its object and 庄明浩’s way of preserving the experience of an era. As old platforms stabilize and lose the possibility of becoming the biggest company, know-how in social, livestreaming, games and esports seems to confront a world where “there are no dragons in the new era.” Writing can still preserve the emotions along a firsthand timeline. “Those stale old stories, those emotions sealed inside memory—AI cannot give you that.”
Deep dive
1. The “Art of Dragon Slaying” Came from a Career Spanning Investing, Entrepreneurship and Corporate Strategy
庄明浩 graduated in 2009 and joined Shanda as a management trainee. In 2010, a rotation took him into the investment department, just as the mobile-internet boom was breaking out—back when the industry still commonly called it “wireless internet.” Early projects included Moji Weather, which marked his formal entry into China’s primary market.
In 2011, he joined Matrix Partners because very few people in Shanghai were genuinely focused on mobile internet. His first stint lasted until October 2015. He looked at tools, mobile commerce and social products, then gradually shifted, out of personal interest, toward anime, ACG, esports and games—“new content.”
On November 1, 2015, he joined Panda TV as a founder and stayed until the company shut down in March 2019. Panda was connected to 王思聪’s IG and Banana Games, turning him from an investor into a hands-on participant in livestreaming, teams, clubs and the esports industry.
After three months of wandering, he returned to Matrix in June 2019. In June 2021, he joined Quwan, a Matrix portfolio company whose core product was TT Voice, where he handled strategy and investment. His more candid job title was “the boss’s consigliere.”
2. “There Are No Dragons in the New Era” Describes the Stable State of the Previous Internet Generation
庄明浩 observes that many internet companies that once wanted to become platforms have entered a stable state: the business may be stable and competition may no longer be intense, but “the possibility of becoming the biggest company” has disappeared. This is not operational failure, but it forces founders to keep searching for new variables such as AI and overseas expansion.
Quwan had originally prepared for a US listing, but something happened around June 2021. By the time of recording, the company still had not gone public, giving him a close-up view of how structural change can freeze what once looked like a predictable capital path.
His career keywords have always been entrepreneurship, investing and To C businesses in social, entertainment, livestreaming, games and esports. Precisely because this know-how is no longer naturally effective in the new cycle, the podcast is called “The Art of Dragon Slaying”: “There are no dragons in the new era.”
3. Around 2010, China’s Primary Market Had Not Even Built Its Basic Map
Information was so incomplete that industry participants did not know which funds were genuinely looking at deals and which existed only in name. Many founders did not know the difference between VC, dollar funds, RMB funds and strategic investors. Some even asked investors: “When do I pay this money back to you?”
Mobile internet took off almost in sync with the US, accelerating founders’ access to Silicon Valley information, the operating model of dollar VC and the industry’s rules all at once. 庄明浩 believes that without this wave, China would have taken much longer to establish the rules of venture capital.
When he prepared to leave Shanda in 2011, he did not even know which funds were hiring or actually doing the work. Compared with today, when business students know VC, FA and other roles, that period was “total darkness.”
4. Shanda Bought by the Rankings, Turning Investing into Random Sampling from a Time Slice
After Shanda decided to accelerate into mobile internet, its strategy department pulled the top 20 apps from several Android app markets and iOS rankings. The instruction was essentially: “Invest in everything investable.” The rankings covered roughly one week, at most a week, but were used to hit KPIs.
A two-person mobile-game company in Guangzhou accidentally entered the rankings with a reskinned version of Plants vs. Zombies. Before 庄明浩’s team had even met the founders, they knew they “had to invest.” The two sides negotiated at a KFC beneath a residential building, eventually investing RMB200,000 for 20%. The amount was back-solved from the cost of making the next game for one year; once the money was spent, the company ended too.
He compared 陈天桥 to an emperor, the strategy department to the Eastern Depot and the frontline team to an expeditionary army: “The emperor wanted to micromanage, issuing commands down to the details—details so granular there was nowhere further to go.” This became a major reason he left Shanda, although he retains the counterview that in exceptional periods, an extreme response to the boss’s anxiety is not necessarily entirely wrong.
5. The “Push from Behind” of Mobile Internet Put a Group of Blank-Slate Young People in the Front Rank
The people most firmly committed to mobile internet were Shanda, Innovation Works—founded by 汪华 and 李开复 after he left Google—and Matrix. There were few projects and few people; teams from the different firms would routinely run into one another outside startup offices. 庄明浩 joined Matrix not after a broad comparison, but through a direct match between scarce supply and demand.
He relays a generational pattern: technology that existed in childhood is taken for granted; technology encountered between ages 15 and 35 is believed capable of changing the world; technology that arrives after 35 is more likely to feel as though it “should be punished by heaven.” New graduates have no path dependence and are therefore most sensitive to the wave’s “push from behind.”
庄明浩 believes the cohort of post-85 investors still active in China’s primary market should all be grateful for that wave. It pushed inexperienced people to the front and made them more willing than senior colleagues to embrace new possibilities head-on.
6. Declaring the End of Early Mobile Internet in 2014 Was a Premature Conclusion after Organizational Excitement Faded
The evidence was not absurd. Smart-device penetration may have reached a new order of magnitude, while Baidu, Alibaba and Tencent—the “regular armies”—had fully entered games, social and e-commerce. Early-stage funds appeared to have fewer and fewer opportunities. 王兴 did not speak of the “second half of the internet” until 2016, so early investors should arguably have sensed the inflection point sooner.
But the conclusion came too early. Pinduoduo launched only in 2016; ByteDance had been founded, but Douyin came later. The companies that ultimately became the largest were either still in the cradle in 2014–2015 or outside the field of fire entirely.
The deeper cause was human and organizational inertia. After watching one direction for years, excitement naturally decays. Existing portfolios also cause new companies to be mapped onto old cases; once an old case performs badly, a new project inherits the negative judgment. People who have received positive feedback find it especially hard to recalibrate.
7. ByteDance’s Real Question Was Hidden by the Wrong Classification as a “News App”
According to 庄明浩, Toutiao met almost every major investor in its first two rounds, but only one fund, SIG, fully captured the ByteDance upside. The earliest angel investment was closer to a bet on the person. 张一鸣 was not known for infectious charisma, further weakening his score under traditional “look at the founder” criteria.
At the time, Sina, Sohu and Tencent were fighting over news apps with pre-installs, paid acquisition and content teams. ByteDance also carried negative labels—ranking manipulation, piracy and disrespect for creators. 庄明浩 recalls that a 财经 interview even opened by asking: “People say you’re a thief. Are you?”
Investors watched the App Store every day and knew which rankings came from manipulation. Add an unclear business model and valuation disputes, and a purist rejection was easy to form. Yet ByteDance’s early Series A and B decks consistently spoke of “content, connection and algorithms.” The market simply lacked historical samples from the internet era to anticipate “algorithms eating everything.”
8. Pinduoduo Was Really Leveraging the WeChat Ecosystem, but Was Treated as the Hardest Kind of Fresh-Grocery E-Commerce
Early Pinduoduo looked like nothing more than “selling fruit by creating groups inside WeChat.” For funds that had already lost heavily in vertical e-commerce in maternal and infant products, cosmetics and fresh groceries, that classification was enough to trigger rejection. Fresh groceries also brought logistics, delivery windows and returns.
At the time, a fund could invest tens of millions of dollars in a single e-commerce project—an enormous sum for the period. The losses became embedded in institutional memory. Past experience was not illogical; the problem was treating an old scar as the new company’s main variable.
庄明浩’s retrospective is that Pinduoduo grew by tapping the enormous traffic of the WeChat ecosystem. “Fresh groceries” was merely its earliest and least important label. It should have appeared last in the label sequence, but investors moved it to the front.
9. Investment Logic Is Trained on Old Experience, while Startups Keep Rewriting Their Assumptions
汉洋 asks: if investors make their living from logic, why did almost nobody derive ByteDance or Pinduoduo, which now seem relatively straightforward? 庄明浩’s answer is that logic can only be built on prior experience; it cannot naturally extrapolate into a future with no sample.
When investors ask questions, they often already have a standard answer in mind. The question, answer and judgment system sit in the brain like an SOP. Analysis is often just checking whether a new company can fit into one of the boxes left by past winners.
Founders face a different information structure: “I have a pocketful of things to say, but I can only tell you two.” As a project develops, its details, connections and assumptions multiply. The business model, product form and operating cadence in the initial BP may all be rewritten by reality.
Success therefore contains luck. 庄明浩 still believes a small number of funds can remain successful, but most investors cannot reproduce the era’s dividends. Industry peers look down on one another precisely because everyone suspects that someone else’s success came from outside logic.
10. VC Methodology Resembles “Praying for Rain through Ritual”: the Ritual Has Value, but Cannot Prove Causality
庄明浩’s core metaphor is that early-stage investing resembles ancient rituals for summoning rain. In conditions of enormous uncertainty, people need actions that can be fixed and repeated. If the outcome partially materializes, those actions are retained and gradually become ritual.
These rituals may be called strategy, taste, preference or genes. They are not entirely useless, but it is difficult to prove whether the ritual produced the result or the result was likely to happen anyway. That is why investment methodology cannot be verified like an engineering rule.
He does not conclude that everything is luck. Methods for sustained success still exist, but may be tightly coupled with team composition, organizational atmosphere and a boss’s momentary flight of fancy. “How do you prove you’ve mastered this thing? I think it’s hard—too hard.”
11. The Most Replicable Advantage May Be an Institution’s Willingness to Make a Blanket Decision at an Inflection Point
庄明浩 cites IDG’s “invest in post-90s founders” strategy. 孙雨辰, the CEO of Insta360; 刘靖康 of Bilibili; 齐俊元, founder of Tim Busion; and the founder of FaceU all came from that cohort. He roughly reviewed the group and found that “six or seven out of seven or eight made it,” with both major and modest successes.
What mattered was not only whom they selected, but that they dared to formulate and execute what looked like a blanket strategy at the time. By 2025, very few institutions were willing to publicly announce a systematic bet on post-00s founders, let alone post-10s founders. That itself shows how difficult it is to convert an era signal into organizational action.
汉洋 distinguishes IDG’s investment in Circle from its investment in Insta360. After deep research, a stablecoin company could still be identified as a business with a high probability of success; Insta360 in its early days looked more like a bet on a person and an unknown path. 庄明浩 agrees that the latter better tests how an institution faces the unexplainable.
12. VC Preserves Only “Kings”; Failed Projects Do Not Even Leave Behind “Defeated Foes”
庄明浩 says bluntly: “VC only looks at success, not failure.” It is not even fully a matter of the victor becoming king and the loser becoming a bandit, because once success arrives, failed cases disappear from the narrative. The public world retains only the “kings” told and retold.
When training interns and analysts as early as 2012–2013, he described VC decision-making as a black box: roughly 1,000 projects enter each year, and five “1s” come out. The middle can only be made explainable to oneself; it cannot be scientifically validated.
After working for a while, an investor may need to choose only five from a small pool of familiar or endorsed projects. 庄明浩 immediately asks: “But what happens to the other 950 projects?” One of them might contain another exceptionally good company.
If asked to choose a GP as an LP, his honest answer would be: “Too hard.” A fund of funds can examine DPI, IRR, historical investments and decision processes, but an individual can barely see through the black box.
13. Investor Seniority Improves Efficiency, but Makes the Most Important Samples Disappear
After several years in investing, someone may no longer need to choose five from 1,000 unfamiliar projects, instead selecting five from personal referrals and already-endorsed deals. Efficiency improves, but the outliers hidden among the remaining projects disappear from view forever.
Newcomers can therefore help an institution keep scanning unfamiliar samples. 庄明浩 sees 沈南鹏’s persistence as rare: even today, he still looks at projects every day like an investment manager rather than compressing his inputs into five answers screened by others.
Investors must build their own judgment function or they cannot pull the trigger. But once the function is built, its parameters become coupled and extremely difficult to revise. A small number of people can keep changing industries, effectively switching functions frequently; most are eliminated when a major or minor wave ends.
Cross-industry commonalities ultimately reduce to soft descriptions such as “look at the person,” initiative, excitement and network-building, none of which is absolute. Too little excitement makes investing impossible; too much can scatter attention. If you are genuinely extremely excited about something, the most rational choice may be to do it yourself.
14. AI Has Put “Sand, Money and Humanity’s Future” in Front of VC at the Same Time
汉洋 describes this cycle as massive dollar liquidity combined with “humanity’s future.” 庄明浩 reduces VC’s historical myth to “making chips out of sand.” When sand and money appear again, it becomes difficult for an institution to explain why it is not betting.
The awkward part is that years later, some sand is still just sand. The previous metaverse cycle also saw companies valued at hundreds of millions of dollars despite having almost nothing, later proven to be “complete nonsense.” Yet an investor who refused to invest at the time could have suffered damage to bonuses and organizational standing.
庄明浩 is willing to read the views of Sam Altman, 王小川, 李开复 and 汪华 because their frameworks, scales and deductions are worth absorbing. But he also admits that the person with the most logical articulation does not necessarily deliver the best investment performance. If the founders he consistently favors all turn out badly, the problem lies in his preference.
汉洋 asks whether one can bet against one’s preferences. 庄明浩 answers: “Everyone says to raise the muzzle by one centimeter, but the muzzle won’t move.” After DeepSeek ultimately became the protagonist rather than celebrity founders such as 王小川 and 李开复, who seemed to have “collected all seven Dragon Balls,” he began to understand why Confucian, Daoist and other classical-theory explanations are so seductive. He still admits he cannot find a method that can actually be executed.
15. Investment Taste Is Not Acquired by Reading Articles, but Beaten into Shape by Experience and Real Money
庄明浩 believes investment taste comes mainly from personal experience—put bluntly, it is “beaten into you by money.” You must commit a meaningful amount of capital and bear enough sunk costs for a preference to take shape. Whether the resulting taste is good or bad is another matter.
He favors business writing with a long timeline in which major and minor events intertwine with the author’s emotions. One English-language long read tracked Peter Thiel, founder of Founders Fund and one of PayPal’s founders, from university through PayPal and its merger with the company originally founded by Musk. What attracted him was not a single exclusive detail, but how a person gradually takes shape over time.
When writing about Tencent Music’s acquisition of Ximalaya, he stayed up all night organizing a mind map out of an “unfounded urge to express himself.” A friend suggested letting AI make a PPT. He refused: “Those stale old stories, those emotions sealed inside memory—AI cannot give you that.”
16. Early-Stage Investing Often Requires Pursuing a Phase-Specific Outcome, Not Pretending to See the Final State
Livestreaming was 庄明浩’s largest bet, because he entered the industry with his own “body.” Short video ultimately capped livestreaming’s ceiling, and livestreaming ceased to be everything. But it once grew large enough in the short term and provided investors with strong exits, so the phase-specific judgment was valid.
His lesson is that investing, business analysis and strategy cannot share the same time scale: “If you always want to see the ending, you won’t be able to do many things.” An effective primary-market judgment may simply identify a monetizable window rather than predict what the company will ultimately become.
17. Pop Mart Shows That Investors May Have Misjudged Not the Company, but the Change the Company Had Yet to Complete
When 庄明浩 encountered Pop Mart, it was still on the NEEQ. Its financials and operations were available for inspection, yet it looked like a digital-goods collection store at an airport—an “uninvestable toy concept store.” Rejecting it on that cross-section was not without evidence.
He believes the part most worth adding to the record is neither Pop Mart’s earliest startup phase nor its post-Hong Kong-listing surge, but the intervening period when nobody believed in it, operations were poor, and the company then transformed. 王宁 struggled at Yishang, later moved to Sanbai, which was not especially successful either, and only then gradually became today’s Pop Mart.
汉洋 asks whether many misjudgments can be understood this way. 庄明浩 answers: “The key is not whether you were right or wrong, but what did 王宁, as founder, change?” A company’s final state includes the founder’s growth, and that part did not yet exist when the financing meeting took place.
18. Investors See a Random Cross-Section; Strong Institutions Build a Mechanism for Regular Reexamination
An investor’s intersection with a company often occurs at a single moment, and that moment’s weight in the company’s full history is random. It is fundamentally no different from Shanda grabbing one week of app rankings, except that a company’s “human cross-section” is even more partial and illusory than a ranking.
庄明浩 especially admires Sequoia’s discipline. After deciding at the first meeting that a project was unsuitable, it would still proactively bring the deal back for review at fixed intervals, forming a “revisit, re-report, revisit, re-report” mechanism. This challenges organizational processes, personal pride and existing conclusions at the same time.
When Matrix had already invested in roughly 50 mobile-internet companies, 沈南鹏 went directly to 张颖 and asked him to select 10: “Give them to me. Invest in all of them.” The group included Ele.me and Momo. The difficulty was not just providing capital, but willingly reentering a battlefield others had already explored.
19. Matrix’s Relative Success Also Came from Separating Financial Investment, Strategic Investment and Incubation
Shanda, Innovation Works and Matrix all saw mobile internet early, but used different mechanisms. Shanda made strategic investments and often sought a large stake or control. It had its own strategic agenda and cannot be equated with independent financial investment.
Many of Innovation Works’ early projects were incubated: Dianxin, Wandoujia and Zhihu, among others. 李开复 and 汪华 often first assembled a group, found people, developed a direction through collision and provided seed capital.
Matrix was more committed to market-based financial investment: taking minority stakes and trusting entrepreneurs to decide the company’s course. 庄明浩 believes the three models are not matters of moral superiority; their objectives are determined by their positions.
The work itself is also different. Primary-market investing is a black box of “1,000 in, five 1s out.” Business analysis organizes history, reality, the future and competitors in pursuit of “complete rationality.” Strategy makes trade-offs after analysis, deciding what can and cannot be done.
20. AI Liquidity Has Added a Zero to the Chips on the VC Poker Table
庄明浩 predicts that the 2025 annual AI review will devote several pages to a “year of M&A,” because abundant liquidity has moved from financing into acquisitions, lifting transaction sizes by a full notch.
He uses Cursor as the clearest yardstick: the software had only reached version 1.0 and the team numbered dozens, yet it raised $900M at a $10B valuation. A vintage of an early-stage fund in the past might have had a total size of only $200M. “The poker chips gained an extra zero,” so the game naturally changed.
New opportunities and new liquidity have made VC more suitable for young people again than it was several years ago. But 庄明浩 says he has already “gone past the point of return”: checks that large make him panic, and he cannot deploy them. VC always carries the instinct to “make one big bet”; today, one bet may equal an entire fund in the past.
21. Genuine Long-Term Capital Is Rare; Annual Deployment Pressure Turns “Patience” into a Mismatch
汉洋 points out that a fund may last five to eight years while requiring investment managers to invest in the hottest direction that year or lose their bonuses. That is pursuing short-term results in a long-term career. 庄明浩 responds that in China’s primary market, “patience” must be placed in quotation marks.
His scarce counterexamples include 徐新’s 28-year fund, which allowed LPs to exit at several points; 程伟’s smaller and therefore more discretionary evergreen fund; and Matrix’s first evergreen RMB fund of roughly RMB200M, perhaps half of it supplied by several partners.
That Matrix fund invested in All-in-One Memo, which after more than a decade reached an A-share listing and a peak market capitalization of roughly RMB50B. 360 was a similar case. These examples say more about how capital duration creates outcomes than many of the hottest AI projects of the day.
This does not mean long-term capital is guaranteed to succeed. It means investment managers only truly have the option to skip the metaverse and wait for a tools company to mature when their capital is not demanding an answer within the year.
22. Dollar Retrenchment and State-Capital Expansion Have Put a “Cannot Lose” Constraint on Venture Capital
After the US-China environment changed, China’s dollar funds shrank substantially. Only the very top institutions could continue raising capital. Other dollars, where they still existed, came increasingly from the previous generation of entrepreneurs, the Middle East, Singapore or Hong Kong Web3 capital rather than traditional US pensions and university endowments.
The RMB market became increasingly dependent on state capital. 庄明浩 estimates that beginning around 2018–2019, state capital’s share of China’s RMB primary market had probably exceeded 50%; by recording time, it “might already be in the eighties.” This includes central and local governments, SOEs, ministries and related university capital.
Venture capital is an outlier in the financial system, with outcomes ranging from total loss to returns of several thousand times. State-capital governance, by contrast, is highly alert to “state-asset losses.” One demands enormous volatility; the other demands no losses. The mismatch is structural.
He offers no simple institutional solution, only says: “What exists is reasonable; the market will tell you.” That preserves the boundary of the judgment: the contradiction is obvious, but participants cannot change the capital structure with a slogan.
23. You Can Bet without Using the Product, but Valuable Interviews Must Enter the Decision Black Box
汉洋 observes that many AI investors do not use the products; some have not even bought the $20-a-month ChatGPT subscription, yet can still discuss their own art of AI investing. 庄明浩 acknowledges that the phenomenon is real and that older-generation investors may still succeed by having people around them fill in the product experience.
This unexplainability has not made him dismissive of all investors. It has instead made him respect people who continue to produce results: “I used to try to learn how they did it. Now I realize I can’t.” He now pays almost no attention to generic investment opinions online.
He praises a conversation hosted by 刘晟, editor-in-chief of Dark Tide. 曹曦, XVC’s 胡博予 and Black Ant’s 何愚 received the questions in advance but still had to answer unannounced follow-ups on the anti-consensus behind successful investments at the time and which counter-consensus founder they had met in the past two months who could be tested again in 3 years.
The stumbling and pauses were precisely where the value lay, because the guests were genuinely thinking, organizing and constructing their answers. 庄明浩 recommends asking investors in reverse: “Why didn’t you invest in this? What was your non-consensus view relative to everyone else?” Because “anything that is consensus is too flat.”
24. Investors’ Positive Presets Induce Founders to Perform and Hide the Counterevidence
庄明浩 once invested in an H5 engine company that provided middleware for webpages, mini-games and interactive content. The founder had extensive experience and resources, the project fit the institution’s taste, and the team even believed that if it failed to develop independently, a large company would probably acquire it.
In reviewing the failure, he realized that investors should not reveal their positive stance to founders too early. Once a founder senses the investor’s preference, they will “follow your lead,” presenting only the parts that fit the investor’s expectations. Other aspects of their personality, capabilities or operations disappear.
The project ultimately vanished entirely, perhaps precisely because of the dimensions that were never shown. 庄明浩 does not blame the founder for being dishonest: anyone trying to raise money would express themselves along the picture the investor had illuminated. The investor also bears responsibility for failing to seek counterevidence.
25. Continuous Writing Preserves a Timeline That AI and Later Generations Cannot Reconstruct
From a highly upvoted Zhihu answer on Didi in 2014 to writing about Tencent Music’s acquisition of Ximalaya a decade later, 庄明浩 treats expression as his comfort zone and long-term SOP. He keeps finding a new balance between external stimuli pushing him to take one more step and his natural laziness pulling him back.
He does not want to package midstream industry analysis as a product just to sell courses, and feels that a blog is sufficient. He is unsure “who would read” a book. 汉洋 counters that these pieces are first written for himself: if firsthand participants do not write, later researchers may collect the material, but they will not have the emotions from the same timeline.
The podcast gives him a comfortable boundary. The content can be extremely information-dense and may even require illustrations, without pandering to listeners who only want to relax. Turning it into a Bilibili video would place it in the visual competition of the knowledge section. No medium is simply better or worse; the result is an accumulation of small habits, atmospheres and comforts.
At the end of the episode, the architecture he wants to preserve is not an internet landmark, but the compound, office and central rockery of the forest farm in his hometown beneath Changbai Mountain. Memory returns with the sweetness of the local water, the oily aroma of freshly roasted pinecones and the springy pine nuts, echoing his understanding of writing: what is truly scarce is not information, but “having your timeline laid alongside it.”