Venture Capital & Investing

Reading guide

Wu Jun examines legendary KPCB investor John Doerr’s approach: invest in the future and the impact a venture can have on the world; then consider the founder’s DNA, seek companies in stealth, and emphasize network effects after investing. Doerr often asks entrepreneurs three questions: how will the world be different two years after receiving the investment? Why aren’t Microsoft or Google doing this? If a large company does it, what are your barriers to entry? Using Google’s shelved and restarted facial-recognition project and its complex-question answering system, the article explains the approach Doerr values: tackling old, previously unsolved problems from a new dimension. He does not judge people solely by outcomes. Despite failures with Segway and a new-energy fund, he maintains that money should be a tool not merely for making more money, but for changing the world.

Reading guide

In Build, Tony Fadell explains how Nest approached marketing. The central tool is the customer journey: list every way people discover the product, identify which parts of its story they encounter at each touchpoint, and build a messaging activation matrix that turns the right facts in the right places into advertising and content. Marketing must be honest—even a harmless lie can destroy trust, and every claim needs supporting data. It should also participate from the first day of product development. The “Why We Made It” page evolves alongside the product, giving marketing a voice in product decisions. Despite limited resources, Nest invested heavily in high-quality imagery; ten years later, Google Nest was still using those assets.

Reading guide

Wang Shuo outlines five features of cycles: everything important has them; they overpower people and institutions; each cycle rhymes without repeating exactly; live long enough and you will experience everything—ergodicity; and cycles are layered and nested, with individual fortunes often shaped by higher-level cycles. He distinguishes three types: positive-feedback bubbles in assets without income, reflexive loops in which share prices and profits reinforce each other, and credit-driven economic cycles that eventually reach a Minsky moment. Responses differ by horizon. In the long run, survival is victory: avoid maximum leverage in good times and endure without giving up in bad ones. In the short run, do not imagine you can reliably catch turning points; use a simple, sensible yardstick, such as Shiller’s valuation measure, to gauge whether levels are high or low. Be skeptical of any theory claiming precise turning-point predictions, including the LPPL model.

Reading guide

In this dialogue, Zhang Lei advocates “doing the difficult but right thing.” He says three things stand opposite to value: zero-sum games, a gaming mindset, and trend chasing. Understand what value is not before discussing what it is. Lu Qi shares why he founded MiraclePlus at 57: China has talent, technology, markets, and capital, giving it the potential to become the world’s largest innovation ecosystem. Early-stage founders need to survive and accelerate product-market fit; in an uncertain age, faster iteration is the only way to improve the odds of success. He also discusses cycles—raising capital in a tight market can be more valuable—sudden disruptions such as the pandemic, and choosing partners whose motivations fit, abilities complement one another, and values align. Bill Gates was the person who impressed him most that year.

Reading guide

The first principle of early-stage fundraising is to secure the first term sheet quickly. Once you have it, you gain the initiative. Three guidelines follow. First, plan investor meetings as you would job interviews: do not meet your favorite first; practice with investors whose processes are simple and decisions quick, but do not leave your preferred investor until last either. Paul Graham suggests an expected-value calculation: probability of investment multiplied by the weight the investment carries. Second, only written terms count. Treat anything an investor says as a no unless the terms are written down for signature; investors have strong incentives to delay. Third, create a concentrated competitive process. Schedule fundraising within the same period. Once the first offer arrives, harsh terms may disappear, and firms previously too busy to meet may suddenly make every partner available.

Reading guide

In an age of frequent black swan events, there is no master key. Zhang Lei nevertheless sees long-termism as a compass through cycles: commit time and conviction to things that create lasting value, learn the most effective ways of thinking, follow first principles, and keep seeking truth. For individuals, it offers clarity against short-term temptations and noise. For businesses, it provides perspective, rejecting zero-sum games and continually rebuilding a dynamic moat through innovation. For society, it is a passion that brings resources to education, science, and the humanities, sustaining a positive cycle. The article concludes that investment has one true standard: whether it creates genuine value that contributes to society’s overall prosperity. Flowing water does not compete to arrive first; it seeks to flow without end. Be a friend of time.

Reading guide

Dedao Headlines explains Richard Koo’s theory of balance sheet recessions through his book 大衰退 (Great Recession, descriptive title). After Japan’s bubble burst, falling asset prices left businesses and households owing more than their assets were worth. Everyone repaid debt rather than borrowing: innovation stalled, economic activity froze, and confidence collapsed so completely that even zero interest rates could not induce borrowing. Toyota, with its rule against debt, was a rare exception. Individually rational decisions produced a macroeconomic catastrophe—Samuelson’s “fallacy of composition.” Koo’s prescription was for government to increase deficits, borrow, and spend, buying businesses time to repair their balance sheets, while businesses restored confidence and resumed appropriate borrowing and spending. Robin adds that government can lead investment and expand innovation amid uncertainty, giving people—especially the young—confidence in the future and helping Taiwan move forward.

Reading guide

Once an investor meeting is secured, founders need three preparations. First, explain what they intend to do. Investors care about monetization and trends: not yesterday’s opportunities, already irrelevant; not today’s, already crowded; and not tomorrow’s, where mature opportunities favor large institutions. Their favorite is the “day after tomorrow.” Explain the industry’s past, present, and future, the opportunity you see, and your preparations. Second, demonstrate a distinctive edge. Barriers may come from advanced technology, industry experience, or—most often—people. With the right team, a second-rate idea can produce a first-rate outcome. Third, prove execution. Investors increasingly value the second-in-command who turns ideas into action. Use past achievements to show that you move quickly; even running in the wrong direction and correcting promptly is better than standing still.

Reading guide

Six lessons for first-time founders. Hire people more knowledgeable and experienced than yourself to fill gaps in your understanding; do not insist on being the smartest person in the company. Attend to both the big picture and details, balancing direction with execution. Assess every aspect of the business objectively: “getting many things done” is not the same as completing what matters; ask whether a decision is necessary and serves the ultimate goal. Consider opportunity costs, including the cost of not acting or acting later. Validate before investing: gather information, test in small steps, and iterate rather than spending months building something customers do not need. Finally, invest more once the approach proves effective. Apply the Pareto principle and deepen the one or two activities generating 80% of revenue rather than spreading effort too broadly.

Reading guide

Wu Jun continues his discussion of John Doerr. The best time for venture capital is the stealth phase, when an industry and company are still young and most people have not recognized the opportunity. Risk is high and opposition inevitable, but by the time everyone agrees, the best entry point may have passed. VC therefore requires insight and risk-taking rather than committee-style democratic decisions. Reviewing a successful portfolio company, Doerr’s only concern was that its gross margin was too high. SGI had died that way: a 70% margin led it to ignore necessary but lower-margin opportunities, leaving it unable to endure hardship after the bubble burst. In Doerr’s era, KPCB limited each fund to $800 million–$1 billion, invested in fewer than 300 companies, took nearly 100 public, and generated returns of 40–50 times per fund. One secret was encouraging portfolio companies to help and acquire one another, creating a broad alumni network.

Reading guide

Lei Jun’s fundraising advice: the worse your character, the harder raising money becomes. The first funding should be pooled from people you know. American angel rounds may involve twenty people contributing $10,000 or $20,000 each. The real question is not whether wealthy people are around you, but why they should trust you. Fundraising reflects credibility accumulated in work and life. If you cannot persuade parents, siblings, or close friends, why should Lei Jun invest? He openly invests only in acquaintances, extending at most to their acquaintances. He cites YY founder Li Xueling: a journalist who criticized him yet refused press-event travel allowances, had principles and ambitions, and loved technology. Ten years of interaction built trust. In 2005, Lei invested RMB 4.15 million and lent another RMB 4.15 million. Taking your conduct seriously is the first step toward credibility; begin looking for money close to home.

Reading guide

Zhang Lei sets out four elements of Hillhouse’s approach. First, uphold an investment philosophy and a broad perspective. Its three principles emphasize ethical discipline and independence, restraint and depth in familiar areas, and doing the right things so like-minded partners naturally arrive. Consistency is not rigidity: Hillhouse fundamentally invests in change. Second, remain even-tempered and return to common sense. Returns should come from business profits and growth, not fashionable concepts or short-term games. He cites Bogle on long-term, risk- and cost-adjusted returns, and Buffett on not holding a stock for ten minutes if unwilling to own it for ten years. The best investment needs no exit. Third, preserve entrepreneurial spirit, put fund investors first, practice empathy, and support a positive cycle in the real economy. Fourth, develop talent and teams built on trust and mutual inspiration. Talent is an investment that never requires an exit.

Reading guide

Retelling Lao Yu’s article: becoming wealthy depends not on how often your judgment is right, but on whether you bet when right and limit losses when wrong. Magnitude matters more than frequency. Druckenmiller says the most important lesson from Soros was how much you make when right and lose when wrong. Buffett advises catching rare windfalls with a bucket, not a thimble, and imagining a lifetime investment card with only twenty punches. Munger says removing their ten best investments would make their record a joke. Behind this is the power law: in Zero to One, Peter Thiel says to seek only the best, because the largest winner outperforms the rest combined. Even a16z’s 312-fold Instagram return barely covered part of a $1.5 billion fund. Zhang Lei concentrated Yale’s $30 million in Tencent and insisted on investing $300 million in JD.com. But how do you identify the moment for a concentrated bet? The author admits, “I don’t know; I even think nobody knows.” The truth may be simple, but it has no operating manual.

Reading guide

Duan Yongping’s philosophy: missing an opportunity is not the concern; taking the wrong one is. The one reason not to sell a stock is “I’ve already made money,” just as its past price is not a reason to buy. The only test is value. When accounting problems led Nasdaq to suspend NetEase in 2001 and its share price fell from $15.50 to $0.64, Duan found a company valued at $20 million with $60 million in cash, only $14 million in liabilities, a fast-growing SMS value-added business, and plans to enter the gaming market he understood. After lawyers assessed the litigation risk as manageable, he used almost all his available $2 million to buy 1.52 million shares in April 2002. Even gains of dozens of times within two years did not prompt a sale. He ultimately held for more than five years and earned over a hundredfold.

Reading guide

When raising capital, your priority is not just to discuss the business but also yourself. First, you are the real competitive advantage. A Stanford survey of hundreds of investors found that founders, management teams, and their experience and passion were the most important investment factors. Introducing yourself reveals capabilities and values, helping investors assess company culture. Second, closeness builds trust. Baylor research involving 600 executives linked trust to leaders’ technical competence, conflict-resolution ability, and integrity. Share your background, inspiration, motivation, experience, and mission. A Harvard-administered global survey found all institutional investors valued ESG and how corporate purpose was communicated. Third, investors want to know whether you can withstand pressure. Nervousness may be read as an inability to handle difficult business decisions. Begin your presentation by speaking confidently about yourself.

Reading guide

Xuan Tian’s empirical research, published in Review of Finance in 2012, examined more than 30,000 US startups founded between 1980 and 2005. Seventy percent received funding from at least two VCs, while 88% of VC-backed companies that went public had syndicate funding. The article uses Didi’s $4.5 billion Series F as a case: Apple, China Life, Ant, Tencent, China Merchants Bank, and SoftBank pooled resources that helped it evolve from a taxi app into a smart-mobility leader. A parenthetical note records its later market-value decline as a point-in-time observation. Syndicates create product-market value through complementary expertise, networks, and information, and greater willingness to fund risky innovators. In financial markets, multiple VC backers send a positive signal affecting IPO underpricing and valuation. But an informed investor with unusually strong conviction may exclude others to retain all the gains; in that case, solo investment could produce better exits.

Reading guide

Excerpts from 領導 (Leadership), coauthored by Manchester United’s legendary manager Ferguson and Sequoia Capital chairman Michael Moritz. Moritz divides the leadership lessons Sequoia learns from other fields into two categories. Internally, they concern recruiting, team building, standards, motivation, avoiding complacency, and removing underperformers. In portfolio companies, Sequoia is a minority shareholder alongside brilliant founders in their thirties who are driven and sometimes volatile; it can help only through persuasion, encouragement, sensible advice, and patience, with results that may not be immediate. Ferguson refused to define Manchester United by its opponents: the club’s destiny was shaped inside Old Trafford, and money could not buy success. The same applies in Silicon Valley. No company becomes great by continually reacting to competitors. Page and Brin built Google because existing search engines dissatisfied them. Successful founders focus on what they control and stay off powerful rivals’ radar for as long as possible. Novell, Netscape, and others learned the cost of attracting Microsoft’s attention at its peak; Google’s low profile meant that by the time Microsoft noticed it, it was too late to respond.

Reading guide

The post opens with Peterson’s observation that opportunities ultimately force a choice between immediate and long-term interests, and Jorgenson’s remark that truly smart people do not take shortcuts. It then draws on Sun Taoran’s 創業36條軍規: a company must make money, and entrepreneurs must give their all, but neither justifies winning at any cost. The 36 rules ultimately come down to one: take the ethical path. If there is a shortcut to success, it is the only one over the long run. Acting ethically supports survival: underhanded methods consume energy, create disputes, and can lead to imprisonment. Doing the right thing may mean earning less now, but it avoids risks and builds a reputation. The text argues that reputable companies receive faster listing approvals and better prices when issuing shares. How should a company respond to a rival’s smear campaign or manufactured provocation? Sun proposes three responses: compete through market action rather than verbal battles; put users’ interests first when speaking; and act openly, especially in a crisis. Robin adds that the financial industry’s bottom line is the safety of funds—risk management, controls, and regulatory compliance.

Reading guide

Xiong Fei of Matrix Partners reflects on founder assessment. During the 2017–2018 wave of returnee-led “copy to China” startups, many investors bought into stories about overseas comparables worth tens of billions of dollars without thinking them through. Other founders were less skilled at grand narratives but understood users, worked on the front line solving real problems, had spent years developing and testing their ideas, committed themselves fully, and attracted well-regarded friends and colleagues. Xiong argues for a dynamic assessment of people. If running a company with a few dozen employees has a difficulty of 10, a company with hundreds may be 10 times 10 squared; management complexity rises geometrically, and differences between founders become apparent. Openness and learning ability determine whether a founder can grow from handling a difficulty of 10 to handling 1,000. Robin adds that growth shifts interactions from individuals to teams and departments. As a Taiwanese saying puts it, it is easier to raise a top scholar than a capable businessperson: academic and research ability do not necessarily reveal management ability, which is often more important to entrepreneurial success.

Reading guide

Andrew Chen’s newsletter explores the dark side of being data-driven. Data measures yesterday rather than tomorrow and can produce analysis paralysis: intuition may provide an answer in seconds, whereas A/B tests take weeks. The tyranny of the majority can reduce a product to a compromise around averages. Noise and false signals abound, correlation is not causation, and historical data struggles to predict a future that has never occurred. Startups competing with giants such as FAANG need intuition, speed, and bold moves rather than copying those companies’ data processes. The ideal is to be data-informed, neither blindly data-driven nor data-ignorant.

Reading guide

Cai Yu examines Hefei’s investment-based approach to attracting industry. When BOE considered locating there, Hefei backstopped RMB17.5 billion. When Sharp intervened, Hefei offered to accept both projects, prompting Sharp to withdraw. It also invested in LCD and plasma technologies to hedge competing technical paths. Investments in ChangXin, GigaDevice, and NIO followed. The RMB7 billion NIO investment involved a NIO China joint venture, a RMB420 billion revenue commitment, an IPO within 60 months, and clear exit arrangements. The underlying approach is to follow a coherent industrial strategy, buy promising assets when they are out of favor, and train investment-promotion teams to conduct due diligence like investment managers. Robin suggests that Kaohsiung, Taichung, and Taoyuan could learn from this ambition. A city should first answer four questions: Who am I? Where am I now? Where do I want to go? How should I get there?

Reading guide

A Caijing article discusses the “one-person unicorn” in the AI era. Li Zhifei used AI to produce 40,000 lines of code in two days; Sam Altman described “one person plus 10,000 GPUs.” Pokee AI, associated with Zhu Zheqing, raised a $12 million seed round with only six people. A view attributed to Sequoia suggests hierarchical organizations are giving way to dynamic collaboration. Highly capable generalists and lean AI-powered solo businesses are emerging, while Chen Yiran comments on the employment market. Cal AI, built by high-school students including 17-year-olds, reportedly earns $1.12 million a month. Robin reminds venture capital professionals that they, too, must keep pace with the times.

Reading guide

In his 1998 shareholder letter, Bezos emphasizes fearing customers rather than competitors: customers remain loyal only until someone offers them something better. Hiring involves three questions: Do you admire this person? Will they raise the team’s average level? In what area are they a superstar? Goals for 1999 include expanding distribution and systems capacity, building the brand, broadening product categories, and strengthening the bench and processes. The letter again attaches the 1997 letter to remind shareholders of the company’s original commitments.

Reading guide

Robin recommends the 2016 shareholder letter as a must-read. Bezos describes Day 2 as stasis, followed by irrelevance, painful decline, and death: a company must remain in Day 1. Four essentials defend it. First, genuine customer obsession: customers are always “beautifully dissatisfied.” Second, resist proxies: processes and market research cannot replace closeness to customers. Third, embrace external trends—then, machine learning and AI. Fourth, make decisions quickly: act with about 70% of the information, disagree and commit, escalate misalignment promptly, and avoid one-size-fits-all processes.

Reading guide

Cai Yu reviews Yang Haoyong’s battle between Guazi and Renrenche, where capital-driven momentum was a key weapon. A month-long stealth expansion into 40 cities and more thanRMB 100 million in advertising produced a gap of 1,700 versus 700 transactions. A Series A of over$200 million then constrained the rival’s fundraising options. Spending created barriers, and capital momentum drove the market toward a 9:1 split. Do not take comfort in “just do your own job well” or “the market is large enough for everyone.” Cai also cautions that capital is only the foundation; data and algorithmic moats form the deepest layer.

Reading guide

Amazon’s first shareholder letter after its IPO was appended to subsequent annual letters. Its core is long-termism: put customers first; abandon projects without returns and increase investment in promising ones; choose cash flow over attractive financial statements; and retain talent through equity incentives. The letter reports 1997 revenue of$147 million, growth of 838%, a 58% repeat-purchase rate, and 1.51 million users. Bezos says the figures will be surpassed, but the long-term principles remain.

Reading guide

Professional-camera sales fell from 121 million units in 2010 to 8.3 million in 2024, but Insta360 captured 67.2% of the global panoramic-camera market and ranked first for six consecutive years. Self-taught programmer Liu Jingkang declined invitations from Zhou Hongyi and Li Kaifu, and moved from livestreaming software into hardware. After the VR bubble burst, he shifted panoramic cameras from specialist VR equipment toward action-camera use, admitting he had benefited from the market GoPro created. Two lenses addressed a key extreme-sports problem: capturing the moment matters more than framing it beautifully. Across ten years, Liu often followed Ricoh, DJI, and GoPro, but innovated in functions and use cases to pull ahead. Now DJI is following Insta360. Robin suggests Taiwan may be better suited to hardware or integrated hardware–software ventures than pure software, but their competitors will be Chinese manufacturers.

Reading guide

In 2001, after four years pursuing growth and two reducing costs, Amazon finally balanced resources between the two. July’s major price cut—30% off books priced above$20—made lower prices a third pillar alongside selection and convenience. A better cost structure supports lower prices; those drive growth, which spreads costs further and supports additional reductions. Eliminating errors at their roots improved both experience and cost. Bezos’s investment framework again favors cash flow over financial-statement appearance: a share represents future cash flows, and understanding future cash flows and share count indicates present value. Goals for 2002 include positive operating cash flow, free cash flow, and keeping net employee-equity dilution within 3% over five years.

Reading guide

There was no specialist venture capital industry in the late nineteenth century. Breaking with his usual practice, Morgan backed Edison's electric-light venture in 1878 despite his father's objections. At the end of 1880, he invested another US$1 million in a central power station in Manhattan. He also helped Edison restructure his board and remove dissatisfied investors. Their honeymoon ended in 1892: after discovering that Thomson-Houston's profit margin was more than twice Edison's, Morgan backed a merger. Edison learned that he was losing his company without having been consulted; his name disappeared from General Electric. With his stake down to 10%, he angrily sold out. From the founder's perspective, this was capital's ruthlessness; in hindsight, General Electric became one of the twentieth century's great industrial companies. Edison's stubborn opposition to alternating current also suggested that he was not the person to lead the electrical industry into the new century. Robin's takeaway: founders must recognize the limits of their abilities. Success need not be led by me, but I should have a part in it.

Reading guide

Dedao Headlines recounts LatePost's interview with Li Xiang, introduced by Robin after posts on Gome's decline and Ajinomoto's transformation. Three insights apply beyond the automotive industry. First, endgame thinking starts with an assumption about scale. Wang Xing and Wang Huiwen advised discussing scale before changes in technology, demand, business, and competition. Scale determines platform development—essential for production capacity—and what to develop in-house: standardized batteries and panels are left alone, while differentiators such as head-up displays, air suspension, and core chips require internal development. Second, avoid the “20,000-person trap” through a matrix organization with interlocking objectives: salespeople are also assessed by innovation teams, and selling one new product can count as ten established products. As Wang Huiwen put it, a team cannot reach a World Cup final on one specialty alone. Third, develop three fundamentals: language problems, or communication—Li used a hand-drawn A4 page to persuade Wang Xing to invest; mathematics problems, or measurement—product value equals product appeal times quality divided by price, with over a hundred quantified dimensions of appeal; and philosophy problems, or continually asking why, guided by Russell's truth-seeking and inclusive cooperation.

Reading guide

The 2004 shareholder letter focuses on free cash flow. Robin's introduction reads Bezos as emphasizing Amazon's strong cash-generating ability: fixed-asset capital expenditure was just 4% of revenue; receivables were collected quickly, inventory turned rapidly, and suppliers were paid later. A larger business therefore generated more cash. Bezos explains why the ultimate measure is free cash flow per share rather than earnings: share value reflects discounted future cash flows, which also depend on working capital, capital expenditure, and dilution. His teleportation-machine counterexample produces doubling annual earnings and 150 million in cumulative profit over four years from a US$160 million machine, yet cumulative negative cash flow of 530 million. Earnings can hurt shareholders when growth investment exceeds the present value of the cash flows it produces. Doubling EBITDA is equally misleading: EBITDA is not cash flow. Amazon's cash-generating operating cycle collects from customers before paying suppliers, tying up only 480 million in inventory for nearly US$7 billion in sales. The collected account says 2004 free cash flow rose from 131 million to 477 million, up 38%, with share-based issuance unchanged from 2003.

Reading guide

Yu Minhong argues that mobile internet and the age of intelligence create more opportunities, keeping entrepreneurship a major topic for another twenty or thirty years. There is no perfect time to start, but it must not be an impulsive decision: it is as consequential as marriage or choosing a city in which to settle. Joining as a supporting partner is easier because you do not provide the main resources. Four considerations follow. First, find a genuinely good project that excites you. Ask whether you can make it succeed, rather than whether someone has done it before. When Yu founded New Oriental, China already had tens of thousands of training providers; he trusted his affection for students, teaching ability, and potential for higher earnings. Second, secure funding. Debating whether to use 500,000 to start a business or buy a home suggests you have not clarified why you want to start. Strong projects can attract angels and seed funds; build a model for tens of thousands first. Third, build a team with a real decision-maker, fair rewards, clear responsibilities, shared expectations, and readiness for sustained hardship. Fourth, founders need leadership, decisiveness, and the ability to mobilize resources. In China, deploying people effectively can matter more than specialist expertise. Yu favors gaining experience first; starting immediately after graduation carries a much higher risk of failure.

Reading guide

This book excerpt uses Peter Thiel and PayPal to explain escape velocity: start fast enough and you can leave rivals far behind. Either enter an emerging field without competitors or move so decisively that they cannot catch up. The goal is not merely to defeat competitors but to escape competition. Early PayPal paid users directly: refer a friend and both received US$10. That demonstrated easy transfers while acquiring customers more cheaply and directly than advertising. Before it had a complete business model, Thiel chose scale first and monetization later, burning more than US$10 million a month. Starting with 24 users, PayPal grew 7% daily and passed a million within three months through compounding. Reid Hoffman explains why large companies are often less dangerous rivals: an eBay error could anger millions and attract regulators, making it hesitant. Its payment system took over a year to launch. In 2002, eBay acquired PayPal for US$1.5 billion. Robin's caution: a larger ambition can justify delaying profit, but not evading the question forever. Explain how the company will eventually earn money so investors will accept the risk, wait, and ultimately earn substantial returns.

Reading guide

Dedao Headlines remembers mathematician and investor James Simons, who died on May 10, 2024, aged 86. Robin notes that quantitative investing differs greatly from venture capital, yet its lessons can apply to entrepreneurship. Simons graduated in mathematics from MIT at twenty, earned a Berkeley doctorate at twenty-three, and received the highest honor in mathematical geometry at thirty-seven. He saw himself first as a mathematician: markets were a giant mathematical problem, with money the reward for solving it. According to the account, Renaissance Technologies' Medallion Fund returned an annualized 66% over thirty years after its 1988 launch, or 39% after fees, outperforming Buffett and Soros. Employees signed lifetime confidentiality agreements, and recruits were mathematicians, physicists, and computer scientists rather than conventional Wall Street financiers. Its explanation of high-frequency quantitative trading uses soccer: the final result is uncertain, but intermediate winning probabilities are comparatively determinate. It claims shorter forecast horizons improve accuracy and cites one trade per one sixty-four-millionth of a second, possible only through computers and algorithms. The lesson is to upgrade an intuitive system into an algorithmic one, breaking enormous uncertainty into many intermediate certainties. Robin applies this to founders: accumulate small wins by dividing an ambitious goal into milestones and steadily approaching them.

Reading guide

Amazon's 2006 shareholder letter discusses incubating new businesses: expected returns on capital, potential scale, and differentiation that matters to customers. It uses the Joyo acquisition, FBA, and AWS as examples and explains why large companies need patience to nurture small seeds. The original English letter describes a fulfillment network of 12 million square feet, FBA storage at US$0.45 per cubic foot per month, and a community of more than 240,000 developers. The collected Chinese translation differs in units and headcount; see the source and verification note. The letter says even fast-growing new businesses often need three to seven years to become meaningful to the company as a whole. These are the experiences and strategies of that period, not current pricing or policy.

Reading guide

Fan Deng responds to a nanomaterials startup team: inventors face considerable entrepreneurial risk because they can love their invention more than the social problem the invention should solve. When difficulties arise, they seek to spread the invention everywhere rather than find the best solution, refusing to give up even after reaching a dead end. His example is an automatic car-cover device developed over ten years, whose creator sought investment at a valuation of 100 million. Fan declined for four reasons: assumed demand was unsupported; covering cars was infrequent; a rooftop device could be stolen; and carrying it in the trunk added weight and fuel consumption, undermining value. Real inventors begin with problems. Edison did not invent the light bulb, but his team found a suitable filament. Jobs did not invent the smartphone, but “redefined” it. Where Nokia insisted that computer functions required a computer keyboard, Jobs reportedly demanded one-button simplicity. The excerpt attributes to Bezos, via Allen Zhang, the question of whether innovation serves people or merely displays cleverness: will you show intelligence at any cost, or choose kindness? Inventions designed to demonstrate a founder's ingenuity rather than address a real problem will, in most cases, fail.

Reading guide

Robin opens by observing that it is again time for presidential candidates’ advisers to study seriously and develop policy proposals. He believes: “Encourage innovation, increase investment, and save Taiwan.” The following excerpts cover six chapters of 創新經濟學. First, highly skilled immigrants make outstanding contributions to innovation. Nearly a quarter of U.S. patents have at least one inventor with foreign citizenship. H-1B visas respond to market demand but weaken workers’ bargaining power and ability to become entrepreneurs, while Canada and Australia actively recruit through points-based systems. Second, the division of innovative labor is changing: universities conduct a growing share of research, while research within corporate R&D has fallen from over 35% in the late 1950s to below 20%. Commercialization relies on university technology licensing under the Bayh–Dole Act and technology-oriented startups, but the corporate-laboratory era has lost ground. Third, pharmaceutical innovation involves both push policies, such as NIH-funded research, and pull policies, such as patents, with persistent concerns about patent evergreening. Fourth, incumbents innovate most effectively when they combine existing advantages with competitive pressure; an inverted-U relationship does not justify permissive horizontal-merger policy. Fifth, fewer than 5% of local economic-policy programs have credible evaluations, and fewer than 0.5% have demonstrated positive policy effects, highlighting the need for prospective randomized experiments. Sixth, innovation opportunities are highly unequal across regions. Policy might encourage innovations that expand opportunity in disadvantaged places rather than strengthen those places’ own innovative capacity. All chapter figures are drawn from the research cited in the book.

Reading guide

Robin’s opening question is what kind of education we need in the AI era. Having raised related educational issues over the preceding few days, he turns to Wan Weigang’s perspective. Wan argues that traditional education takes a top-down approach: schools, teachers, and parents hold the initiative. Once likened to a gardener cultivating plants, it now has a more precise analogy—training AI. Defining the learning scope in advance, feeding in materials labeled right or wrong, and assessing the results is supervised learning. Yet students taught this way fall short even of GPT, which uses self-supervised learning, discovers patterns for itself, and is designed to know things its teachers do not. Traditional education and AI share three shortcomings that innovators do not: rewards depend on administrators’ approval, whereas innovators gain recognition from society and consumers, without fixed standards, and may even define what society considers good; mistakes and shoring up weaknesses receive great attention, whereas innovators need exceptional strengths; and new things outside the defined scope attract little interest, whereas innovators pursue novelty because it may confer a competitive advantage. The biggest problem with AI-style education is students’ lack of autonomy. A machine’s great advantage is its inherent passivity; a defining human difference is the desire to take the initiative. The conclusion: in an age with AI, we must think carefully about nurturing a person rather than training one. Creative people can be nurtured but cannot be controlled. To hope for unexpected good, we must tolerate unexpected “bad”—things we did not want them to do. AI can be controlled; with talented people, we must learn to tolerate their independence.

Reading guide

These book excerpts explain the central idea behind a “small and beautiful” business: profitability first. Rather than pursuing expansion single-mindedly, do everything possible to become profitable. The author was Pinterest’s second employee. In 2011, he left before cashing out his equity to launch Gumroad, a platform for creators to sell their work online. It attracted more than 50,000 visitors on its first day, then raised over US$8 million from prominent Silicon Valley investors and recruited dozens of employees from internet giants in pursuit of a unicorn dream. Once the funding ran out, growth stalled. Nine months of fundraising attempts failed, forcing him to lay off most employees. After moving to a quiet small city, friends helped him recognize that Gumroad had actually remained profitable and was appropriately sized for its target market. Thousands of creators used it to sell courses, ebooks, and software to supplement their incomes. It might have been a poor investment for Silicon Valley investors, but it was a valuable platform for its customers. He realized that many businesses raise substantial venture funding early because they have not found a sustainable profit model, falling into a cycle of pursuing size for its own sake. Entrepreneurship is like sailing: relying entirely on fundraising is like jumping into the sea, treading water, and waiting for a lifeboat. The sustainable approach is to build your own boat—profitability. Robin urges Taiwanese startups to establish that their business models can generate profits before expanding. As in the Buffett view he had shared, making money is the only way to develop sustainably.

Reading guide

Building on the three requirements for a good fundraising story—engaging, credible, and concise—this post explains its structure. The most suitable framework is the familiar confrontation between a hero and a villain, followed by a happy ending. That may sound overly simple, but its simplicity is precisely its fundamental strength. In an entrepreneurial story, the villain is not a particular person but a real problem: hunger, ignorance, disease, waste, inefficiency, or a poor customer experience. A business idea without a clearly defined villain cannot raise funding, because it does not solve a real problem. Google’s villain is the lack of ways to access information, its hero is its search algorithm, and its happy ending is the business empire built through tens of billions of searches. Uber’s villain is being unable to find a ride; its happy ending is a ride for everyone and income from otherwise idle vehicles. Great business stories, however, are never pure fiction. The next step is to substantiate the plot with examples and factual research. First, bring the villain vividly to life: an office worker leaving at 9 p.m., exhausted by clients and a boss, with no dinner arranged. Then introduce the hero—a food-delivery website—and explain just as thoroughly why it can win, whether through breakthrough technology or service innovation. A happy ending requires a reasonable market forecast: the market share to be gained, how long that will take, how much it will cost, how much will be earned, and the basis for every projection.