103 Chinese originals · 90 English translations · Page 3 of 3.
Trends & technologyThematic illustration · AI generated
Topics use the existing Chinese title and full-text keyword rules. English translations retain the same topic membership; articles may belong to more than one topic. Browse the complete Chinese topic →
Reflecting on American divisions and a Facebook article, Robin questions whether greater freedom of expression and information have actually connected people. Algorithms can trap groups in echo chambers, showing only desired information and reinforcing beliefs, even shaping perception. He urges independent judgment and the ability to falsify claims when receiving algorithmic or AI-mediated information. A description of echo chambers explains repeated, distorted voices in closed environments. Seek varied experiences and perspectives while retaining your own reasoning.
Key points
Algorithms can show groups only what they want to see and believe.
The result may be greater division rather than connection.
Develop independent judgment and the ability to challenge and falsify claims.
Explore diverse perspectives while maintaining your own reasoning.
Before expanding overseas, ask whether the market actually wants your project. In Singapore, for example, investment may follow naturally when a project is both a globally recognized Taiwanese strength—such as semiconductors, AI hardware, IC design, or critical components—and something Singapore wants to develop. If neither condition holds, the effort may yield little. The central questions are whether you can prove world-class strength and how badly the market needs you.
Key points
Before expanding overseas, ask whether the market wants your project.
Investment may follow naturally when the project matches both Taiwan’s recognized strengths and the host market’s priorities.
If neither condition holds, the effort may yield little.
Prove that you are world-class and that the market strongly needs you.
Excerpt from 范冰’s 前線部署工程師, with Robin’s commentary (the original post did not name the author)
Reading guide
Enterprise AI delivery can draw on three kinds of partners: cloud providers and large platforms, consultancies and systems integrators, and customers’ own customers and supply chains. Robin adds that potential partners in Taiwan include the Big Four accounting firms, telecom operators, and systems integrators.
Key points
Three partner groups can support enterprise AI delivery: cloud providers and platforms, consultancies and systems integrators, and customers’ customers and supply chains.
Customers’ committed cloud spend is a practical lever for co-selling.
Partners can provide customer relationships that would take three years to build yourself.
Robin: potential partners in Taiwan include the Big Four accounting firms, telecom operators, and systems integrators.
Reprinted compilation (post credits Professor Yang Bin of Tsinghua University’s School of Economics and Management in Beijing), with Robin’s closing commentary
Reading guide
A reposted article by Tsinghua professor Yang Bin proposes a simple test of a good university in the AI era: how much entrepreneurship it fosters. Entrepreneurship education offers embodied learning and wholehearted personal engagement, led by an expanded teaching community of coaches, mentors, entrepreneurs, and investors. It should combine an expanded classroom with extracurricular practice and real business challenges. Universities should preserve their educational purpose rather than count only unicorns. Robin concludes that the best universities will have the strongest entrepreneurship education and the most complete entrepreneurial ecosystems.
Key points
Entrepreneurship education offers university value that AI cannot replace.
“Faculty-plus”: coaches, mentors, entrepreneurs-in-residence, and investors—not only academic professors.
“Classroom-plus”: credit for learning outside formal courses and real business projects that teach through action.
Success means students become self-directed, willing, capable, and effective—not merely producing more unicorns.
Reposted compilation (recommendation-preface excerpt; post credits Yang Bin) + Robin’s closing remarks
Reading guide
This excerpt from Tsinghua University School of Economics and Management professor Yang Bin’s recommendation preface to The Infinity Machine (Chinese edition: 《哈薩比斯：谷歌AI之腦》) discusses three forces driving DeepMind founder Hassabis: purpose (mission), will, and taste. It concludes that the three are interconnected and that purpose need not be a grand narrative. Robin’s closing remarks recommend the book as a way to return to one’s inner compass in an age of AI anxiety.
Key points
Purpose generates will, and will sustains purpose
Taste is both a fruit of purpose and a force that focuses will
Purpose need not be grand; it is genuine if it can continually guide your inner life
This post compiles three qualities Walter Isaacson draws from Leonardo da Vinci, Einstein, Jobs, and Musk: curiosity, passion, and mission. Robin adds a fourth—imagination—and observes that every child has curiosity and imagination, which are sadly often stifled by their environment.
Article by Wan Weigang (as identified in the post title), with Robin’s commentary
Reading guide
Wan Weigang discusses DeepSeek R1 in January 2025: performance approaching o1, very low prices, and publicly available weights, and what these mean in contrast to OpenAI’s closed approach. Robin’s comments on open and closed approaches follow at the end. These views reflect January 2025.
Key points
R1’s key features are performance approaching o1, low cost, and publicly available weights
A comparison of open and closed approaches
Robin’s closing point: openness, innovation, and breadth of vision beat a single-minded effort to block others
Reprint compilation (tagged #得到頭條20260820; author not identified), with Robin’s closing comments
Reading guide
A repost of Dedao Headlines’ three lessons from Unitree founder Wang Xingxing: building one complete thing exposes system-level problems; when conditions change, reassess, holding to your criteria rather than an old conclusion; and define how a forecast will be tested before making it—the “two 80%s.” Robin’s closing comments add figures for Wang’s shareholding and wealth.
Key points
A small, complete thing provides more information than half of a large one
When conditions change, reassess; remain consistent in your criteria
Define acceptance criteria before forecasting: a forecast without them conveys no information
#得到頭條20260304 (Dedao Headlines course script), with Robin’s closing comments
Reading guide
Three points about lateral thinking: rational abandonment matters more than stubborn persistence (the well-digging analogy, illustrated by Bettger dropping third and subsequent sales calls after analyzing his records); quantity matters more than quality (force yourself to generate ten ideas, with Dyson’s 5,127 prototypes as an example); and some disorder matters more than perfect organization (add constraints, bring in outsiders, and stay open to surprises). Robin closes by urging curiosity, openness, and alertness.
Key points
The well-digging analogy: abandon a location once you find it is wrong; test new directions with small bets
With one answer, the brain chooses the safest; forcing ten ideas helps later ones break habitual patterns
Constraints, outsiders, and openness to surprises can break fixed patterns of thought
Bezos shareholder-letter translation (Chinese translation by Vinchent), with Robin’s introduction
Reading guide
The power of wandering: efficiency can be planned; invention requires exploration. Third-party sellers grew from 3% to 58%, “beating us badly,” while AWS reached a US$30 billion annual revenue run rate.
Key points
Third-party sellers: 3% to 58%, with 52% compound annual growth
