梅乾與武士刀：在傳統文化中的古老智慧，揭開大和民族創新獨步的關鍵

 

（點此前往博客來購書）

在日圓不斷貶值、引發美日政府聯手救市的今天，全世界已經對日本國力留下衰退的印象。即使日本首相高市早苗不斷以大破大立的手段，試圖扭轉自泡沫經濟以來積弱不振的日本國內景氣，但壓在頂頭的少子化和僵固的形式主義，依舊像烏雲般籠罩在日本國運上空。

與上述唱衰觀點截然不同，《梅乾與武士刀》一書是由真心熱愛、讚揚日本文物的考古學者樋口清之所著。藉由他窮盡一生的文物探索和歷史考證所累積的發現，向讀者分享日本人獨創、有別於西方科學的、與大自然共生共榮的奇思巧技。身處在地震、颱風等災禍環繞的土地上，加上東方特有的敬天惜物風俗民情交互影響下，日本傳統的科學精神，不是歐美人所持的人定勝天，而是取法自然、順應環境的天人合一。

書中提及許多例子，像是運用石頭、原木所築成的河壩，相較於現代普通的水泥結構，更加環保、耐用，又相當美觀，能呈現出河川自然的風貌。在衣著方面，傳統的蓑衣、木屐，特別適合多雨潮溼的氣候，由自然素材製作，更有著極佳的透氣性和零汙染。反觀今日以塑膠材質製作的雨衣、鞋款，穿著過久容易讓人產生皮膚濕疹等困擾，不禁令你我省思，哪邊才是真正的進步呢？

書中舉出許多看似老掉牙，實際卻優於現代作法的人事物，涵蓋飲食、衣著，甚至是都市規劃等。作者主張，這些證據證明日本民族擁有優秀的開創性思維，能因地制宜地發明出像武士刀一樣兼具美感和實用性的生活用品，獨步全球。在世界各國一味追求經濟發展的今天，人類對自然環境持續造成不可逆的傷害，人和土地、森林和海洋的距離是越拉越遠，只能生活在由水泥、金屬打造而成的現代叢林。對比書中提及的古老智慧，人們真的是更進步嗎？

Today, as the Japanese yen continues to lose value and the Japanese and U.S. governments work together to stabilize the market, people around the world have begun to see Japan as a country in decline. Even though Japan’s Prime Minister, Sanae Takaichi, has repeatedly tried to turn the country’s struggling economy around with major reforms, Japan still faces serious problems such as a declining birth rate and a rigid culture of formalities. These problems continue to hang over the country like dark clouds.

However, *Umeboshi and the Samurai Sword* presents a very different view of Japan. The book was written by Kiyoyuki Higuchi, an archaeologist who deeply loved and admired Japanese cultural artifacts. Based on what he discovered through a lifetime of studying artifacts and researching history, Higuchi shares many unique ideas and practical skills developed by the Japanese. These ideas are different from Western science in some ways and focus on living in harmony with nature. Living in a land surrounded by natural disasters such as earthquakes and typhoons, and influenced by the traditional Eastern values of respecting nature and avoiding waste, the Japanese developed a traditional scientific spirit that was different from the idea of “conquering nature.” Instead, they learned from nature and adapted themselves to their environment.

The book gives many examples. For instance, some river dams were built using stones and logs. Compared with ordinary modern concrete structures, these traditional methods can be more environmentally friendly and durable, while also looking more natural and fitting better into the landscape of the river. When it comes to clothing, traditional rain capes made from natural materials and wooden sandals were especially suitable for Japan’s rainy and humid climate. They were highly breathable and produced almost no pollution. In contrast, modern raincoats and shoes made from plastic materials can cause problems such as skin irritation when worn for too long. This makes us stop and think: **What is real progress?**

The book introduces many things that may seem old-fashioned but can actually be better than some modern methods, including food, clothing, and even urban planning. The author argues that these examples show the Japanese people’s creativity and ability to adapt to their environment. They were able to create practical everyday products, such as the samurai sword, that combined beauty and usefulness in a way that was unique in the world. Today, countries around the world continue to pursue economic growth, while human beings are causing irreversible damage to the natural environment. We are becoming increasingly distant from the land, forests, and oceans, and are left living in a modern jungle made of concrete and metal. Compared with the ancient wisdom described in the book, **are we really more advanced than people in the past?**

Python修仙之路：VS Code + GitHub Copilot──從AI輔助學習到專案開發，全方位提升你的「程式閱讀力」

 

（點此前往博客來購書）

自上次研究 Claude Code 後，開始接觸到 AI 代理人程式開發，想當然耳也會想評估一下另一個鼎鼎有名的 GitHub Copilot，試著找出兩者的差異，而這本由陳會安撰寫的《Python 修仙之路：VS Code + GitHub Copilot》就成了我開箱 GitHub Copilot 的入門書。談起陳會安這位作者，真的算是位奇人。我 20 多年前初出社會、謀職時買的第一本電腦書，是講 ASP 網頁程式設計，作者就是他！除了是長青的資訊技術作者外，看到他掛名的著作能從 C 語言跨到 AI 工具，真心覺得這就是十八般武藝都略懂略懂的人吧。

就我來看，這本書的定位不太討喜。對於真正的程式開發者來說，介紹的內容都是日常工作中必備的基礎知識。從程式語言的基礎、程式碼的重構與優化、程式除錯以及版本控制，到最後的容器化部署，這些作業即使不使用 GitHub Copilot，對於工程師來說也不費吹灰之力。反過來說，如果書籍的目標讀者是完全不懂程式的門外漢，那這些內容又顯得節奏過於快速，造成知其然而不知其所以然的隱憂。不過，或許這正是 AI 代理人開發時代的特色，讓每個人都能在不了解細節的情況下，做出可以動起來的軟體，一切都像魔法般神奇。

在完讀這本介紹 GitHub Copilot 的作品後，大略體驗了這個工具的基本功能和能耐。感覺它是從原本程式開發 AI 助理的定位逐步調整，慢慢朝向 Claude Code 的AI 代理人模式前進。換言之，它本身具有能加速程式人員撰寫程式碼的傳統功能，再疊加上現在最熱門的下指令讓 AI 寫程式的技術，變成魚與熊掌都要的 AI 產品。可惜的是，畢竟 AI 代理的思維，感覺是 GitHub Copilot 後來才添加進去的，所以在整體的使用細節和功能完整性上，是落後 Claude Code 的解決方案。如果用汽車來比喻，它就像是油電混合車，既不像傳統方式要求工程師純手動、一行行刻出程式碼，又不完全讓 AI 代理人從無到有產製出近乎成品的專案。對於 IT 部門管理者和第一線的開發者來說，拿捏在什麼情境下要使用自動完成、什麼時候又該尋求 AI 代理，以及在工程師與這類 AI 工具協同工作的情況下，未來程式碼的質、量管理要如何調整、考核，這些都是要額外討論、思考的議題。

After researching Claude Code, I started getting into AI agent-based software development. Naturally, I also wanted to take a look at another famous tool, GitHub Copilot, and see what the differences are between the two. This book, *Python Cultivation: VS Code + GitHub Copilot*, written by Chen Hui-An, became my beginner's guide to trying out GitHub Copilot. Speaking of Chen Hui-An, I really think he is quite an interesting author. More than 20 years ago, when I first entered the workforce and was looking for my first job, the first computer book I bought was about ASP web programming. And guess what? He was the author! Besides being a long-time writer of IT books, I was also surprised to see that his books cover such a wide range of topics, from C language to AI tools. Honestly, I feel that he is the kind of person who knows a little about almost everything.

In my opinion, the positioning of this book is not very appealing. For real software developers, most of the topics covered in the book are basic knowledge that we need in our daily work. From programming language basics, code refactoring and optimization, debugging, and version control, all the way to container deployment, none of these tasks are very difficult for engineers, even without GitHub Copilot. On the other hand, if the target readers are complete beginners who know nothing about programming, the book moves too quickly. There is a risk that readers may learn how to do things without really understanding why they work. However, perhaps this is exactly what makes the age of AI agent-based development special. It allows almost anyone to build software that actually works without understanding all the details. Everything feels almost like magic.

After finishing this book about GitHub Copilot, I got a basic idea of what this tool can do and experienced some of its main features. I feel that it is gradually moving away from its original role as an "AI assistant for software development" and moving toward the AI agent approach of Claude Code. In other words, it still has its traditional features for helping programmers write code faster, while also adding the increasingly popular approach of simply giving AI instructions and letting it write the code. It is trying to have the best of both worlds. The problem is that, since the AI agent approach seems to have been added to GitHub Copilot later, it still falls behind Claude Code in terms of overall usability and completeness. If I use a car as an example, GitHub Copilot is like a hybrid car. It is neither like the traditional approach, where engineers manually write every line of code, nor does it completely let an AI agent build an almost finished project from scratch. For IT managers and frontline developers, deciding when to use traditional code completion and when to turn to an AI agent is now another issue to think about. At the same time, when engineers work together with these AI tools, we also need to rethink how the quality and quantity of future code should be managed and evaluated. These are all issues that require further discussion and thought.

紅色滲透：中國媒體全球擴張的真相

 

（點此前往博客來購書）

自改革開放後，中國國力日益增長，聚積的經濟實力也展現在國際影響力上，這也成為2009年起大外宣計畫的源起。透過源源不絕的銀彈優勢，滴水穿石地吸引被網路取代的傳統媒體，形成恩庇侍從的組織結構。雖然美國先賢傑弗遜曾說：「與其擁有沒有報紙的政府，寧可擁抱沒有政府的報紙」，強調新聞自由的可貴。但相較於言論受到管制的極權國家，民主國家堅持的出版、傳播自由容易受到惡意操弄，破壞社會團結、國家安定。

這本《紅色滲透：中國媒體全球擴張的真相》，正是揭露源遠流長的中共外宣史，完整、全面地讓讀者瞭解中國如何持續不懈地對外輸出有利於自身的新聞資訊，時刻呼應中國當下的政治目標。早期中國對外的宣傳手法，也只做到出口轉內銷的效果，但隨著執行手法更加熟練，進行外宣期刊／媒體本土化、重金吸引洋人雇傭軍團加盟大外宣……等策略。拿人錢財替人消災，這些外國人也確實抱著宣傳高於真相的信念，努力撰寫出當地國人接受度高的新聞內容，逐步增強宣傳組織的影響力，但也讓美國政府警惕而重啟外國人代理登記法。在1980年代全球中文媒體的政治版圖中，政治取向是相當多元化的。媒體經營者來自大陸背景、港台新聞背景和美國華人投資三大塊。然而時過境遷，在傳統媒體經營困難的大環境下，不計損益的中共官方傳媒演變成一家獨大的局勢，又成立許多海外中文媒體的研究中心、研修班，定期舉辦各項大型國際會議，據以不斷吸收、影響中文新聞圈內的各方人士。掛著國外媒體招牌的傳播筒，成為中共對外傳播除了國內媒體、境外媒體以外的第三陣營。

在香港，更能看出中共對傳播環境不遺餘力的管制與破壞。在港英時代，受惠於英國統治，逐漸呈現媒體自由化的局面；然而在九七回歸後，媒體開始受控、自律。加上所謂的愛國富商積極討好中共當局，投入資金收購媒體，進而影響新聞報導內容。在以商逼政的策略下，媒體報導不親中共者，就斷絕其廣告收入，一步步讓香港媒體喉舌化。暗地又收買黑社會勢力，動輒以暴力威脅不配合的媒體人，形成風聲鶴唳的言論環境，終於達成一言堂般的歌功頌德局面。

台灣面對的操作手法如出一轍，藉由在大陸有巨額投資的富商，以民間投資的身分，不計成本地收購經營困難的媒體，最顯著的例子就是旺旺集團入主中天、中時和中視集團。紅頂商人成功入主媒體後，新聞報導的內容就開始一面倒地宣揚中國政府的高大上，無法堅守必須平衡報導的新聞原則。更惡質者，在關鍵時刻配合中共當局影響台灣社會輿論，成為境外勢力操弄民情的共犯。而中資轉外資洗白後，被買下大量持股的TVBS，則是以另一種報喜不報憂的方法參與中國大外宣。對歐美各國的報導取材十有八九挑選負面議題，反之，中國新聞皆是正面主題，潛移默化地扭曲、影響閱聽眾對世界局勢的觀感，最明顯的例子就是《文茜的世界周報》。

其實，一個國家的國際形象取決於該國內政的社會狀態，極權國家投入再多資源、努力地擦脂抹粉，一樣無法讓眾人真心相信那個地方是流著奶與蜜的樂土。相對地，美國用它自身的民主自由，以及其他領先全球的制度與成就，讓它成為全世界人民心目中移民、置產的首選之地。落實了，嘴裡說的是一套，真的用腳投票才是真相。然而，在AI時代，從更遠的格局來看，大外宣能達成的實質效果可能更加深遠。在大語言模型是以網路文本為原始資料，作為人工智慧基礎的情況下，大外宣計畫無窮盡地散發正面的報導內容，描述中國；反之，受制於西方媒體平衡報導的原則，會對自身國家多所批評而產生負面報導。此消彼長之下，AI將會被誤導而認為中國的一切都優於西方諸國。如果未來的民眾過於依賴AI工具，這種AI幻覺將會對世界局勢產生不良的影響，必須格外留意。

After China began its reform and opening-up, its national power grew rapidly. Its growing economic strength was also reflected in its increasing influence around the world. This became one of the reasons behind the “Great External Propaganda Campaign,” which began in 2009. With an endless supply of financial resources, China gradually attracted traditional media organizations that were struggling because of the rise of the Internet. Over time, this created an organizational structure based on a system of patronage and loyalty. The American founding father Thomas Jefferson once said, “I would rather have newspapers without a government than a government without newspapers,” emphasizing the value of freedom of the press. However, compared with authoritarian countries where speech is strictly controlled, the freedom of publishing and communication enjoyed by democratic countries can also be deliberately manipulated. When this happens, it can damage social unity and national stability.

The book *Red Infiltration: The Truth Behind China’s Global Expansion of the Media* reveals the long history of the Chinese Communist Party’s external propaganda. It gives readers a complete picture of how China has continuously worked to spread news and information that serves its own interests, while constantly supporting China’s current political goals. In the early days, China’s overseas propaganda was mostly limited to what could be described as “exporting something overseas and then selling it back at home.” But as its methods became more sophisticated, China began localizing overseas propaganda publications and media outlets, offering large amounts of money to recruit foreigners to join its propaganda efforts, and using many other strategies. These foreigners were, in effect, paid to do the job. They also genuinely followed the belief that “propaganda is more important than the truth.” They worked hard to produce news content that would be more acceptable to local audiences, gradually increasing the influence of these propaganda organizations. However, this also raised concerns within the U.S. government and led to renewed attention to the Foreign Agents Registration Act. In the political landscape of Chinese-language media around the world during the 1980s, political views were quite diverse. Media owners mainly came from three groups: people with backgrounds in mainland China, people with journalism backgrounds from Hong Kong and Taiwan, and Chinese Americans who invested in the media. However, times have changed. As traditional media have struggled financially, Chinese state-run media, which do not need to worry about profits and losses, have gradually gained a dominant position. China has also established research centers and training programs for overseas Chinese-language media and regularly organized large international conferences. Through these activities, China has continuously attracted and influenced people from different parts of the Chinese-language media industry. These media outlets, which operate under the name of foreign media organizations, have become a third force in China’s overseas communication efforts, alongside domestic media and foreign media operating outside China.

Hong Kong provides an even clearer example of how the Chinese Communist Party has worked relentlessly to control and damage the media environment. During the British colonial period, Hong Kong gradually developed a relatively free media environment. However, after the handover in 1997, the media gradually came under greater control and began to practice self-censorship. At the same time, so-called “patriotic businessmen” actively sought to please the Chinese government. They invested large amounts of money to acquire media companies and, in doing so, gained influence over their news coverage. Under the strategy of using business interests to influence politics, media outlets that did not support the Chinese Communist Party could lose their advertising revenue. Step by step, Hong Kong’s media became more like a mouthpiece for the government. Behind the scenes, criminal organizations were also allegedly used to threaten journalists and media workers who refused to cooperate. This created an atmosphere of fear in which people became afraid to speak freely. Eventually, Hong Kong developed something close to a one-sided media environment, filled with praise for the Chinese government.

Taiwan has faced very similar methods. Wealthy businessmen with huge investments in mainland China have used the appearance of private investment to acquire media companies that were struggling financially, sometimes without much concern about profitability. One of the most obvious examples is the Want Want Group taking control of CTi, China Times, and CTV. After these businesspeople successfully took control of the media, the news coverage began to increasingly praise the Chinese government, making it difficult to maintain the journalistic principle of balanced reporting. Even worse, at critical moments, some media outlets may cooperate with the Chinese Communist Party to influence public opinion in Taiwan, becoming part of efforts by foreign forces to manipulate public sentiment. Meanwhile, after Chinese investment was transferred through foreign companies to make its origins less obvious, large amounts of shares in TVBS were acquired. This became another way of participating in China’s external propaganda efforts—by “reporting the good news while ignoring the bad.” When reporting on the United States and European countries, negative issues are often selected. In contrast, news about China is often presented in a positive way. Over time, this can subtly distort the audience’s view of world affairs. One of the most obvious examples is *Sisy’s World News*.

In fact, a country’s international image depends largely on the social conditions inside the country. No matter how many resources an authoritarian government spends trying to improve its image, it cannot truly convince everyone that its country is a paradise flowing with milk and honey. By contrast, the United States has used its democracy and freedom, as well as its institutions and achievements in many areas, to become one of the top destinations in the world for people who want to immigrate or buy property. In the end, what people say is one thing, but where they actually choose to live and invest is the real answer. However, in the age of AI, the potential impact of China’s external propaganda may be much greater when viewed from a broader perspective. Large language models are built partly on massive amounts of text from the Internet. If China’s propaganda campaign continuously produces huge amounts of positive content about China, while Western media, following the principle of balanced reporting, also produces criticism and negative reports about their own countries, the two sides may create an imbalance in the information available online. Over time, AI systems could be influenced by this imbalance and develop a distorted view, potentially making China appear better than Western countries in certain areas. If people in the future become too dependent on AI tools for information, this kind of AI-generated distortion could have a negative impact on world affairs. It is therefore something we need to pay close attention to.

台灣超越日本，真的嗎？鳳梨、便當、台積電，台日社會文化多樣交流的觀察與思索

 

（點此前往博客來購書）

在日本長崎原爆和平紀念典禮發生將台灣代表座位移到使節團區域外的事件後，打擊了台灣人對台日關係長期友好、以「絆」字所代表的美好想像。其實國際關係原本就錯綜複雜，更何況台灣有著全球罕見的外交困境，遇到官方往來和民間交流呈現的極端溫差，更是應該自立自強，發展、維持國家本身的軟、硬實力。如同這本《台灣超越日本真的嗎？》書名所示，隨著時代變遷、彼消我長的情況下，在東北亞，台灣幾乎快要迎頭趕上南韓、日本兩國，讓我們足以在 1960 年代後，再次在 AI 世代寫出另一篇台灣奇蹟。

作者野島剛是活躍於網路、週刊的媒體人／作者，長期深耕於了解台灣政治體制、產業文化、風俗民情，加上曾任朝日新聞社並外派多國的經歷，讓他在分析、評論日本、台灣兩國的各種議題時，都能以白話流暢的文字，一針見血地替讀者歸納重點。他是位相當友台的日本媒體人，更直白地說，他的政治思維是偏綠，並且是會為民進黨政策美言幾句的作家。不過整體來說，他討論的主題和抱持的立場，都是足夠客觀而且深具價值。

原本只是因為「台灣鐵道族為什麼不受我的家鄉便當？」這篇討論進軍台北車站的日本火車便當業者崎陽軒的文章，才開始翻閱這本書。雖然書籍有二分之一的內容是在討論台、日政治局勢的相關主題，較不受我的青睞；其餘偏向生活化、點出台灣特色的文章，能從日本人的眼中觀察自己，都非常有趣，也能帶來許多反思。台灣人和日本人在做事的風格上，台灣人求快，而日本人求穩。而在高速發展的現代經濟中，台灣人馬上做、錯了再改的方法，反而更能適應變化。

由這麼熱愛台灣的外國人眼光，點出台灣有待修正的缺點，更是值得我們加油改進。像是眾所皆知的行人地獄問題，平時待人和善的台灣人一旦坐上交通工具，便有如雙重人格般，路怒症隨之發作。台灣人求職、做事是以利益為導向；中小企業做事，少有經營事業的長遠眼光；個人謀職也不會有從事志業的職人精神。這是日本依舊優於台灣的觀念。

After the recent incident at the Nagasaki Peace Memorial Ceremony, where Taiwan's representative was moved outside the area reserved for official delegations, many Taiwanese people were disappointed by their long-held image of the friendly relationship between Taiwan and Japan, especially the idea of the bond represented by the Japanese word "kizuna." International relations have always been complicated. And Taiwan faces an unusual diplomatic situation that is rarely seen around the world. When there is such a big difference between official relations and the close exchanges between ordinary people, Taiwan should rely on itself even more. We need to develop and maintain both the soft power and hard power of our country. As the title of the book *Is Taiwan Really Overtaking Japan?* suggests, times are changing, and the balance between countries is also changing. In Northeast Asia, Taiwan is almost catching up with South Korea and Japan. This gives us a chance to create another "Taiwan Miracle" in the age of AI, just as we did after the 1960s.

The author, Tsuyoshi Nojima, is a media writer and author who is active on the Internet and in weekly magazines. He has spent many years studying Taiwan's political system, industries, culture, and way of life. He also worked for *The Asahi Shimbun* and was sent to several countries as a correspondent. Because of this experience, when he analyzes and comments on issues related to Japan and Taiwan, he can explain them in simple and fluent language and clearly point out the key ideas for readers. He is a very pro-Taiwan Japanese journalist. To put it more directly, his political views lean toward Taiwan's green camp, and he is a writer who is willing to say something nice about policies of the Democratic Progressive Party. However, overall, the topics he discusses and the positions he takes are objective enough and still provide great value.

I originally started reading this book because of an article titled "Why Don't Taiwanese Railway Fans Like My Hometown Bento?" It was about Kiyoken, a Japanese train bento company, trying to enter Taipei Main Station. About half of the book discusses political issues between Taiwan and Japan, which are not really my favorite topics. The other half is more about everyday life and the unique characteristics of Taiwan. It is very interesting to look at Taiwan through the eyes of Japanese people, and it gives us a lot to think about. In terms of how they work, Taiwanese people tend to value speed, while Japanese people tend to value stability. In today's fast-changing economy, however, the Taiwanese way of "do it first and fix it if something goes wrong" may actually be better suited to rapid changes.

It is especially valuable when a foreigner who loves Taiwan points out the things that Taiwan still needs to improve. We should take these comments seriously and do our best to improve. Take the well-known problem of Taiwan's "living hell for pedestrians," for example. Taiwanese people are usually friendly and kind, but once they get behind the wheel or onto the road, it can feel like they have a second personality, and their road rage suddenly appears. When Taiwanese people look for jobs or do business, they often focus on financial benefits. Small and medium-sized companies often lack a long-term vision for building a business. At the individual level, people looking for jobs often do not have the kind of professional spirit that comes from treating a job as a lifelong calling. This is one area where Japan is still better than Taiwan.

上白石萌音的心情點滴

 

（點此前往博客來購書）

在初中一年級就因為參加東寶灰姑娘甄選而步入演藝圈的上白石萌音，因在同屆比賽中和妹妹上白石萌歌一起獲獎，一門雙姝聯手拿下被公認為日本美少女魚躍龍門的大賽，幾乎可說是前無古人，後無來者。更值得稱讚的是，兩姐妹不因天生麗質的外貌而恃寵而驕，認真努力地一步一腳印，在競爭激烈的日本藝能圈裡累積作品和人氣。

《上白石萌音的心情點滴》這本書，是典型出版給粉絲閱讀、珍藏的散文寫真書。其實我對日本藝人也沒有太多了解，這本書是我友人買來放在書架陳列的，我閒來無事，就當成睡前讀物把它翻完。其實也曾經看過她主演的日劇，不過在讀過這本書後，才更進一步對她本人的經歷和個性有了更多的了解。

書中以散文的形式，撰寫出有如日記般的隨筆。一來驚豔於上白石萌音除了表演藝術外，也能寫出漂亮、舒服的文字作品，才貌雙全真的就是用來形容這樣的人。二來從文字中可以感受到，她在待人處事上有種老派而踏實的堅持，有如嚴以待己、寬以待人的大氣。另外，她恰好是一位重症的愛書人，不論是閱讀書籍，或是單純喜歡書的觸感、聞起來的味道，這又增添了我對她的好感。總而言之，對於粉絲來說，這本書不能不收；而對於不是粉絲的人，翻翻書，看看美少女平日在想什麼，也是挺有趣的，不是嗎？

Mone Kamishiraishi entered the entertainment industry when she was only a seventh grader, after taking part in the “Toho Cinderella Audition.” In that same competition, she and her younger sister, Moka Kamishiraishi, both won awards. It was quite a remarkable achievement for two sisters from the same family to win a competition widely regarded as a major stepping stone for young Japanese beauties. You could almost say it was something that had never happened before and may never happen again. What is even more admirable is that the sisters never let their natural beauty go to their heads. Instead, they have worked hard, taking things one step at a time and steadily building up their body of work and popularity in Japan’s fiercely competitive entertainment industry.

*Mone Kamishiraishi’s Little Moments* is a typical photo-essay book published for fans to read and treasure. To be honest, I don’t know that much about Japanese entertainers. A friend of mine bought this book and put it on the bookshelf as part of the collection. One day, with nothing much to do, I picked it up and ended up reading the whole thing as a kind of bedtime read. I had actually seen one of the Japanese TV dramas she starred in before, but it wasn’t until I read this book that I gained a better understanding of her background and personality.

The book consists of essays written in a diary-like, casual style. For one thing, I was impressed to discover that, besides being a talented performer, Mone Kamishiraishi can also write beautifully and comfortably. “Beauty and brains” really does seem like the perfect phrase to describe someone like her. For another, you can sense from her writing that she has an old-fashioned, down-to-earth determination when it comes to dealing with people and conducting herself. There is a certain generosity in the way she holds herself to high standards while being forgiving toward others. She also happens to be a serious book lover. Whether she is actually reading books or simply enjoying the feel of a book in her hands and the smell of its pages, this side of her makes me like her even more. All in all, for her fans, this is definitely a book worth owning. And for those who aren’t fans, it’s still fun to flip through the pages and see what this beautiful young woman thinks about in her everyday life. Isn’t it?

用數據讓客人買不停：Google策略長教你解讀數據，善用對話打造長久顧客關係

 

（點此前往博客來購書）

在今日，談起人工智慧，人人口中都會提到輝達；與之相比，提到數位行銷，沒人能忽略谷歌的影響力。實際上，這家矽谷巨頭的廣告業務收入，占總營收的比重高達七成，而廣義來說，廣告業務的核心技術就是數位行銷。《用數據讓客人買不停》書名本身相當有吸引力，加上作者尼爾．霍恩（Neil Hoyne）曾任 Google 數據策略長，參與、指導過許多企業數位轉型專案，這些都更加增添了讀者對本書內容的期待。

本書的主題環繞在指導企業打造長久、有效的客戶關係，而出乎意料的是，作者認為執行高品質的數位行銷，資料和工具並不是最重要的事。他主張一切應該從簡單著手，別理會商業雜誌、軟體廠商每年強力推銷的時髦術語。在正確的心態下，僅靠手上的客戶交易發票和基本的客戶聯絡資料，就能成就一些驚人的成果。不斷測試與優化，而非原地踏步，是數位行銷能夠長久發展的一項關鍵。一封加上姓名、量身客製化的電子郵件，開信率就能提升 20%；轉換率增加 31%，像這樣的小嘗試不勝枚舉。

除了常見的網站行為追蹤外，直接和客戶對話也是一種能夠了解對方的做法。而實作的重點，在於要問對問題。例如問「你購買本項產品是作為禮物嗎？」、「你在外出用餐方面花費多少錢？」前者回答「是」的人，往往是更認同該品牌的優勢客戶；後者則是藉由了解客戶的消費能力，推算其潛在的成長空間。而像是「你最喜歡我們的地方是什麼？」這樣正面的詢問，比起「你覺得有任何需要改進的地方嗎？」這種負面的問法，更能帶來 8% 的客戶銷售額成長。

從事數位行銷工作，絕對要擁抱消費者行為分析的科學發現，讓人類天性的強處與弱處，都成為你的助力。無論是文字上的微調、強調商品的稀少性，或是營造群眾效應等操作手法，都已被證明是有效影響人們行為的技巧，但也要小心其黑暗面，不要玩過頭。在執行行銷計畫上，作者也提出簡明扼要的指導：一切都要從欲解決的問題著手，不必大費周章製作試算表，也能開始進行。設定目標、辨識信號、建立衡量標準；如果資料真的太過複雜，機器學習便可以派上用場。也請記得多利用市面上已有的工具和方法學，不要重新發明輪子。也要記得世事多變，你所得出的看法並非恆久不變的真理，要持續投資，以找到屬於明天的答案。

做人不要大小眼，但經營一家公司卻不然，把心力花在忠誠客戶身上，遠比花在過路客身上更理性，也更符合實務。有一個簡單、有效的工具，就是計算顧客終身價值，並將80 / 20 法則牢記在心。別期望人們的行為會有大幅度的改變，給予精打細算的客戶更多優惠，只會犧牲自身的淨利，增加不必要的客戶獲取成本。成功的企業總是了解最有價值的客戶是哪一群人，並打造出能夠獲得他們歡心的活動，迎合他們原本的偏好與行為。

Today, when people talk about artificial intelligence (AI), everyone thinks of NVIDIA. In the same way, when people talk about digital marketing, no one can ignore Google's influence. In fact, about 70% of Google's total revenue comes from advertising. In a broad sense, the core technology behind advertising is digital marketing. The title of the book, *Data-Driven Marketing*, is very attractive. Also, the author, Neil Hoyne, was once Google's Chief Strategist for Data and Measurement. He has worked on and guided many companies through digital transformation projects. All of this makes readers expect a lot from this book.

The main purpose of this book is to teach businesses how to build strong and lasting relationships with customers. Surprisingly, the author believes that data and tools are not the most important parts of good digital marketing. He says that everything should start with simplicity. Don't pay too much attention to the trendy terms that business magazines and software companies promote every year. With the right mindset, even simple customer invoices and basic contact information can help you achieve amazing results. The key to long-term success is to keep testing and improving instead of standing still. For example, simply adding a customer's name to an email can increase the open rate by 20%. It can also increase the conversion rate by 31%. There are many small improvements like these.

Besides tracking customer behavior on websites, talking directly with customers is another good way to understand them. The most important thing is to ask the right questions. For example, you can ask, "Are you buying this product as a gift?" or "How much do you spend on eating out?" Customers who answer "yes" to the first question are often more loyal to the brand. The second question helps you understand a customer's spending power and future potential. Another example is asking, "What do you like most about us?" This positive question works better than asking, "Is there anything we should improve?" According to the book, the positive question can lead to an 8% increase in customer sales.

If you work in digital marketing, you should understand what consumer behavior research has taught us. Human strengths and weaknesses can both become useful tools. Simple changes in wording, creating a sense of limited supply, or showing that many other people are buying the same product have all been proven to influence people's decisions. However, these methods also have a dark side, so they should be used carefully. The author also gives simple advice on how to plan a marketing campaign. Start with the problem you want to solve. You do not need complicated spreadsheets before you begin. Set your goals, identify the important signals, and decide how to measure success. If the data becomes too complicated, machine learning can help. Also, remember to use the tools and methods that already exist. There is no need to reinvent the wheel. At the same time, remember that the world is always changing. Your conclusions today may not be true tomorrow. Keep investing in learning so you can find tomorrow's answers.

In daily life, we should treat everyone fairly. But in business, things are different. It makes more sense to spend your time and effort on loyal customers than on one-time buyers. One simple and useful tool is Customer Lifetime Value (CLV). Another important idea is the 80/20 rule. Don't expect people to change their behavior too much. Giving bigger discounts to customers who already look for the best deals will only reduce your profits and increase your customer acquisition cost. Successful companies always know who their most valuable customers are. They create products, services, and marketing activities that these customers truly like, instead of trying to change their natural preferences and behavior.

Claude Code Vibe Coding 開發手冊 第 2 版

 

（點此前往博客來購書）

即使不像對岸因高失業率引發的各行各業「內捲」那般慘烈，在台灣，近年來資訊科技相關工作的日子，也越來越不輕鬆自在了。除了像 Claude Code 這類強大的 Vibe Coding 工具緊追在後之外，IT 技術本身的突破與迭代速度也快得令人措手不及。以前可能學個 COBOL，就能安穩地寫二十年的銀行系統；學個 SQL，也能在中小企業的 MIS 部門一路做到退休，這種好日子已經一去不復返。現在資訊技術的保鮮期，多則不到兩年；像 AI 領域更誇張，新東西可能不到半年就過時，又得重新學起。

以這本《Claude Code Vibe Coding 開發手冊 第 2 版》為例，上市日期是 2026/06，離今天大概也才一個月。書中提到建立 subagent 的 /agents 指令，就已經被原廠停用、廢棄了。雖然就學習、使用的角度來說，這並不會造成太大的困擾，依舊能改用口語式指令操作，但對出版社、作者等教學方而言，撰寫、出版 AI 技術書籍的性價比肯定不高。畢竟，相較於討論底層核心觀念、歷久彌新的電腦書，這類 AI 熱門主題書如果不能在出版當年賣掉，大概也只能被秤重回收了。所以看電腦書，尤其是主題偏向 AI 實作時，近來得出的心得是：有需要再買，買了就馬上看，千萬不要積書！

由於我才剛看過這本書的第一版，因此能很明確地比對出新增加的章節內容。整體而言，基本概念與操作方式變化不大，實作練習則略有增加，翻閱時本質上比較像是在看別人做習題。因此，本書真正有價值的地方，在於 Claude Code 新功能的介紹，像是將特定執行流程 SOP 化的 Skills，以及跨平台的遠端操控功能。另外，內建 MCP 的 Claude in Chrome，也讓我眼睛一亮，更確切地體認到，在 AI 工具的協助下，未來的網路環境中，爬蟲機器人將會是再也無法阻擋的必然存在。還是那句老話，程式開發人員已經不能不懂 Vibe Coding。來一本吧！

Even though Taiwan is not facing the same serious "involution" as China, where high unemployment has made competition much tougher, working in the IT industry here has also become more stressful in recent years. Powerful Vibe Coding tools like Claude Code are moving forward very quickly. At the same time, new IT technologies are changing faster than ever. In the past, you could learn COBOL and spend twenty years writing banking systems. If you learned SQL, you could also work in the MIS department of a small or medium-sized company until retirement. Those days are gone. Today, many IT skills stay useful for less than two years. The AI field changes even faster. A new technology may become old in less than six months, and then you have to learn something new again.

Take this book, *Claude Code Vibe Coding Developer's Guide, 2nd Edition*, as an example. It was published in June 2026, only about one month ago. However, one command for creating a subagent has already been removed by the official developer. Luckily, this does not cause much trouble for readers. We can still do the same thing by using natural language commands instead. Still, for publishers and authors, writing AI books is probably not a very good business. Unlike classic computer books that teach basic ideas and stay useful for many years, AI books can become outdated very quickly. If they do not sell soon after publication, they may end up being recycled. Because of this, I have learned one important lesson about computer books, especially AI books: buy them only when you really need them, and read them right away. Don't let them sit on your bookshelf for too long.

I finished the first edition not long ago, so it was easy for me to see what had changed in this new edition. Overall, the basic ideas and operations are almost the same. There are a few more hands-on exercises, but reading them feels a little like watching someone else solve homework problems. The most valuable part of this book is its introduction to Claude Code's new features. For example, it explains Skills, which can turn common workflows into standard operating procedures (SOPs). It also introduces remote control across different platforms. Another feature that impressed me was Claude in Chrome, which comes with a built-in MCP. It helped me realize that, with the help of AI, web-crawling bots will become an unavoidable part of the future Internet. As I always say, software developers can no longer ignore Vibe Coding. Go get a copy of this book!