[{"content":"Welcome! \u0026lsquo;Product = Analysis + Process\u0026rsquo; I am an analyst who enjoys exploring diverse areas such as Product Design, Business Models, Business Process Improvement, and Supply Chain. I’m passionate about discovering the essence of these subjects and understanding the underlying \u0026lsquo;why\u0026rsquo;, \u0026lsquo;what\u0026rsquo;, and \u0026lsquo;how\u0026rsquo;. I use this site to hold my profile and share some thoughts from time to time. Read more about me or check out my portfolio, posts if you are interested.\nThe following series are continuously updated (you can find them in \u0026ldquo;Tags\u0026rdquo; section):\nBeyond the XXX: In this series, I share reflections and insights from my professional journey and daily life. Product Decoder: In this series, I analyze exceptional product designs and deconstruct the mechanics behind their success. Book Review: In this series, I review impactful books and share my key takeaways. AI Practice Journal: In this series, I document practical, hands-on projects I have built using AI. Passport to AI Era: In this series, I explore various aspects of the AI industry, such as Data Centers and related topics. ","date":null,"permalink":"https://chunghaolee.com/","section":"","summary":"\u003cp\u003e\u003cem\u003e\u003cstrong\u003eWelcome!\u003c/strong\u003e\u003c/em\u003e\n\u003cbr\u003e\n\u003cbr\u003e\n\u0026lsquo;\u003cstrong\u003eProduct = Analysis + Process\u003c/strong\u003e\u0026rsquo;\n\u003cbr\u003e\nI am an analyst who enjoys exploring diverse areas such as \u003cstrong\u003eProduct Design\u003c/strong\u003e, \u003cstrong\u003eBusiness Models\u003c/strong\u003e, \u003cstrong\u003eBusiness Process Improvement\u003c/strong\u003e, and \u003cstrong\u003eSupply Chain\u003c/strong\u003e. I’m passionate about discovering the essence of these subjects and understanding the underlying \u0026lsquo;why\u0026rsquo;, \u0026lsquo;what\u0026rsquo;, and \u0026lsquo;how\u0026rsquo;.\n\u003cbr\u003e\n\u003cbr\u003eI use this site to hold my profile and share some thoughts from time to time.\n\u003cbr\u003eRead more \n      \n    \u003ca href=\"https://chunghaolee.com/about/\"\u003eabout me\u003c/a\u003e or check out my \n      \n    \u003ca href=\"https://chunghaolee.com/portfolio/\"\u003eportfolio\u003c/a\u003e, \n      \n    \u003ca href=\"https://chunghaolee.com/posts/\"\u003eposts\u003c/a\u003e if you are interested.\u003c/p\u003e","title":""},{"content":"","date":"May 9, 2026","permalink":"https://chunghaolee.com/posts/external-%E6%8E%A2%E8%AE%BF%E7%81%AB%E7%AE%AD%E5%B7%A5%E5%8E%82%E4%B8%AD%E5%9B%BD%E5%95%86%E4%B8%9A%E8%88%AA%E5%A4%A9%E7%83%AD%E4%B8%8B%E7%9A%84%E7%9C%9F%E5%AE%9E%E8%BF%9B%E5%B1%95/","section":"Posts","summary":"","title":"[External Article] - 探訪火箭工廠，中國商業航天熱下的真實進展"},{"content":"","date":null,"permalink":"https://chunghaolee.com/tags/ai/","section":"Tags","summary":"","title":"AI"},{"content":"Recently, due to project demands, I have begun working extensively with the Command Line Interface (CLI). In the lightning-fast world of AI, being even a few months late to a tool can make one feel like a technological dinosaur. Yet, this \u0026ldquo;belated\u0026rdquo; experience has sparked some profound reflections. Much of my career has been dedicated to building human-centric platforms—designing features that simplify complex tasks through visualization and UI/UX optimization. However, with the rise of AI Agents, there is now a surging demand to pivot many of these same functions back into a CLI format.\n最近因為產品推進的關係，我開始深入接觸 CLI（命令列介面）。(在日新月異的 AI 領域裡，晚了幾個月才親手實作這些，感覺自己簡知像還停留在恐龍時代)。但這份略顯「遲來」的體驗讓我產生了一些深刻的思考。我之前的工作是打造以人為本的平台，設計各種功能方便人們操作，因此過去我投入了大量心力在可視化呈現與 UI/UX 優化上。但隨著 AI Agent的普及，現在同樣有強大的需求要把許多功能轉換成 CLI 形式。\nWhen you strip it down, the essence of both GUI and CLI is the same: data. The only difference lies in how we present that data. Humans are biologically wired to process visual information and prefer intuitive, tactile interactions. AI, conversely, excels at processing text—specifically structured content.\n仔細想想，不論是 GUI 還是 CLI，底層的本質都是資料。差別只在於我們用什麼方式呈現這些資料：人類天生擅長處理圖像資訊，習慣直覺式的操作交互；而 AI 則擅長處理文字，特別是結構化的內容。\nThe Data Remains Constant; The \u0026ldquo;Observer\u0026rdquo; Has Changed // 數據沒變，變的是「誰在看」 #Looking back at the history of computing, the evolution of human-computer interaction (HCI) has come full circle.\n回顧計算機發展史，人機交互的模式其實經歷過一次完整的輪迴。\nThe CLI emerged around 1961, an era when computers communicated with humans primarily through plain text. Users had to memorize a vast library of commands and input them via keyboard to operate the machine. While this was incredibly efficient for the computer, the cognitive load and learning curve for humans were prohibitively high.\nCLI（命令列介面）大約在 1961 年開始被使用，那個時代的計算機主要透過純文字與人類互動。使用者必須記住大量指令，透過鍵盤輸入命令來操作電腦。這種方式對電腦來說效率極高，但對人類來說學習成本昂貴。\nThis changed in 1984 when Apple launched the Macintosh, bringing the Graphical User Interface (GUI) into the mainstream. Because humans naturally favor imagery and intuitive interaction, the GUI rapidly conquered the industry. For the next forty years, interaction design tilted almost entirely toward the GUI, leaving the CLI as a niche tool reserved for developers and system administrators.\n直到 1984 年蘋果推出 Macintosh，將 GUI（圖形使用者介面）帶入大眾視野。因為人類天生更喜歡圖像與直覺式的操作交互，GUI 很快就席捲了整個產業。此後近四十年，整個電腦產業的操作與交互設計幾乎完全朝向 GUI 傾斜，CLI 逐漸淪為開發者與系統管理員的專用工具。\nCommand-Line Interface (CLI) Graphical User Interface (GUI) The CLI Renaissance: An Unlikely Reversal in the AI Era // AI 時代的 CLI 復興：一場意料之外的逆轉 #However, modern AI is built upon Large Language Models (LLMs), which are trained from the ground up on the principle of \u0026ldquo;text completion.\u0026rdquo; For an AI, text and structured content are its native tongue. As AI Agents are deployed at scale across various sectors, the CLI—which is naturally compatible with these models—has staged a remarkable comeback. In this text-only logic, even non-textual data like images, audio, or video must be transcribed into text descriptions before an AI can ingest and process it.\n然而，現在的 AI 核心是 LLM（大型語言模型），它們從一開始就是透過「文字接龍」的方式訓練而成。對 AI 來說，文字與結構化的內容就是它的母語。當 AI Agent 開始大規模部署到各種場景中時，為了配合這位新時代的超級大腦，與其天然適配的 CLI 順理成章地迎來了文藝復興。在這個純文本的邏輯裡，即便是圖像、聲音或影片等非文字數據，也要先被轉譯為文字描述， AI 才會接收與處理。\nThis creates a fascinating conflict: the plain text that AI excels at is precisely the format humans process least efficiently. A human can interpret a complex chart far faster than they can read a long string of raw numbers.\n但這就出現了有趣的衝突：AI 擅長處理的純文字，恰恰是人類處理效率較低的格式。人看一張圖表的理解速度，遠比閱讀一長串數字快得多。\nConsequently, we find ourselves in an awkward position: to make AI work more smoothly, we are forced to revert to the interaction method we find most difficult.\n這就造成了一個尷尬的局面：為了讓 AI 工作得更順暢，我們被迫回到了自己最不擅長的交互方式。\nHTML: A Universal Language for Humans and AI Alike // HTML：人與 AI 都能讀懂的共同語言 #This is precisely why the article \u0026ldquo;Using Claude Code: The Unreasonable Effectiveness of HTML\u0026rdquo; has sparked such widespread discussion. It offers a brilliant insight: HTML may be the optimal solution to the friction between human and AI interaction.\n這也正是為什麼，那篇《Using Claude Code: The Unreasonable Effectiveness of HTML》 的文章引起了廣泛討論。文章提出了一個極具洞察力的觀點：HTML 可能是解決人與 AI 交互矛盾的最佳方案。\nThe uniqueness of HTML lies in its dual nature:\nHTML 之所以獨特，在於它同時具備兩種屬性：\nFor AI: It is structured text that LLMs can easily parse, generate, and modify. 對 AI 來說，它是結構化的文字，可以被 LLM 輕鬆解析、生成和修改 For Humans: It can be rendered by a browser into rich images, charts, and interactive interfaces. 對人類來說，它可以被瀏覽器渲染成豐富的圖像、圖表和交互介面 HTML avoids the user-unfriendliness of a raw CLI while bypassing the \u0026ldquo;black box\u0026rdquo; complexity of a pure GUI that is difficult for AI to decompose. It sits in the middle as a \u0026ldquo;bilingual interface\u0026rdquo;—a single source of truth that the model can understand and execute, but a human can intuitively navigate.\nHTML不像純 CLI 那樣對人不友善，也不像純 GUI 那樣對 AI 難以拆解。它站在中間，成為一種「雙語介面」：同一份內容，既能被模型理解執行，也能被人直觀操作。\nThe Road Ahead: Defining a Shared Working Language // 未來的方向：尋找人與 AI 共同的工作語言 #Looking toward the future, the central challenge for product design and engineering will be finding an interface format that allows humans and AI to collaborate seamlessly.\n面對未來，如何找到一個讓人與 AI 都能高效理解、無縫合作的介面形態，將是下一波產品設計與工程領域最核心的命題。\nWe are not here to completely overturn the past, but to grow new branches from established roots. As the famous adage goes: \u0026ldquo;Men make their own history, but they do not make it as they please\u0026hellip; but under circumstances existing already, given and transmitted from the past.\u0026rdquo; We stand today at the intersection of the GUI’s brilliance and the CLI’s legacy, ready to write the next chapter of human-AI collaboration.\n我們並非要完全推翻過去，而是在舊有的基礎上長出新的枝枒。正如那句經典的名言所說：「人們自己創造自己的歷史，但他們並不是隨心所欲地創造……而是在直接碰到的、既定的、從過去承繼下來的條件下創造。」我們正站在這個繼承了 GUI 輝煌與 CLI 基因的交叉口上，寫下人機協作的新歷史。\n© Chung-Hao Lee. All Rights Reserved. All content on this webpage—including but not limited to text, images, design, code, and multimedia materials—is protected under the international copyright treaties. Unauthorized reproduction, modification, distribution, public transmission, or commercial use is strictly prohibited. Legal action will be taken against infringement. © 李崇豪。保留所有權利。 本網頁之內容（包括但不限於文字、圖片、設計、程式碼及多媒體素材）均受國際著作權條約保護。未經書面授權，嚴禁任何形式之複製、改作、散布、公開傳輸或商業利用。侵權者將依法追訴。\n","date":"May 9, 2026","permalink":"https://chunghaolee.com/posts/beyond-the-interface-designing-a-common-language-for-humans-and-ai/","section":"Posts","summary":"\u003cp\u003eRecently, due to project demands, I have begun working extensively with the Command Line Interface (CLI). In the lightning-fast world of AI, being even a few months late to a tool can make one feel like a technological dinosaur. Yet, this \u0026ldquo;belated\u0026rdquo; experience has sparked some profound reflections. Much of my career has been dedicated to building human-centric platforms—designing features that simplify complex tasks through visualization and UI/UX optimization. However, with the rise of AI Agents, there is now a surging demand to pivot many of these same functions back into a CLI format.\u003c/p\u003e","title":"Beyond the Interface: Designing a Common Language for Humans and AI"},{"content":"","date":null,"permalink":"https://chunghaolee.com/tags/beyond-the___/","section":"Tags","summary":"","title":"Beyond The___"},{"content":"","date":null,"permalink":"https://chunghaolee.com/tags/external-article/","section":"Tags","summary":"","title":"External Article"},{"content":"","date":null,"permalink":"https://chunghaolee.com/tags/pm/","section":"Tags","summary":"","title":"PM"},{"content":"","date":null,"permalink":"https://chunghaolee.com/posts/","section":"Posts","summary":"","title":"Posts"},{"content":"","date":null,"permalink":"https://chunghaolee.com/tags/sharing/","section":"Tags","summary":"","title":"Sharing"},{"content":"","date":null,"permalink":"https://chunghaolee.com/tags/","section":"Tags","summary":"","title":"Tags"},{"content":"","date":null,"permalink":"https://chunghaolee.com/tags/%E4%B8%AD%E6%96%87/","section":"Tags","summary":"","title":"中文"},{"content":"✅ You\u0026rsquo;re in! / 訂閱成功! #Thanks for confirming your email. You\u0026rsquo;ll receive a digest every Saturday morning (Asia/Taipei) whenever there are new posts or portfolio entries.\n感謝確認 email。日後只要有新文章或作品集更新,你會在每週六早上(台灣時間)收到摘要。\nIf a week passes with no new content, no email is sent — I won\u0026rsquo;t fill your inbox with empty digests.\n如果整週沒有新內容,當週就不會寄信 — 不會用空白摘要塞爆你的信箱。\nYou can unsubscribe at any time using the link at the bottom of every newsletter.\n任何時候都可以從每封信底部的退訂連結取消訂閱。\n← Back to home / 回首頁\n","date":null,"permalink":"https://chunghaolee.com/subscribed/","section":"Subscribed / 訂閱完成","summary":"\u003ch2 id=\"-youre-in--訂閱成功\" class=\"relative group\"\u003e✅ You\u0026rsquo;re in! / 訂閱成功! \u003cspan class=\"absolute top-0 w-6 transition-opacity opacity-0 -start-6 not-prose group-hover:opacity-100\"\u003e\u003ca class=\"group-hover:text-primary-300 dark:group-hover:text-neutral-700\" style=\"text-decoration-line: none !important;\" href=\"#-youre-in--%e8%a8%82%e9%96%b1%e6%88%90%e5%8a%9f\" aria-label=\"Anchor\"\u003e#\u003c/a\u003e\u003c/span\u003e\u003c/h2\u003e\u003cp\u003eThanks for confirming your email. You\u0026rsquo;ll receive a digest every Saturday morning (Asia/Taipei) whenever there are new posts or portfolio entries.\u003cbr\u003e\n感謝確認 email。日後只要有新文章或作品集更新,你會在每週六早上(台灣時間)收到摘要。\u003c/p\u003e\n\u003cp\u003eIf a week passes with no new content, no email is sent — I won\u0026rsquo;t fill your inbox with empty digests.\u003cbr\u003e\n如果整週沒有新內容,當週就不會寄信 — 不會用空白摘要塞爆你的信箱。\u003c/p\u003e\n\u003cp\u003eYou can unsubscribe at any time using the link at the bottom of every newsletter.\u003cbr\u003e\n任何時候都可以從每封信底部的退訂連結取消訂閱。\u003c/p\u003e","title":"Subscribed / 訂閱完成"},{"content":"✓ You\u0026rsquo;ve been unsubscribed / 你已退訂 #Sorry to see you go. You won\u0026rsquo;t receive any more newsletters from this site.\n很可惜你要離開,你不會再收到本站的 newsletter。\nIf this was a mistake or you change your mind later, you can resubscribe from any page footer.\n如果是誤點,或之後改變主意,可以從任一頁的 footer 重新訂閱。\n← Back to home / 回首頁\n","date":null,"permalink":"https://chunghaolee.com/unsubscribe/","section":"Unsubscribed / 已退訂","summary":"\u003ch2 id=\"-youve-been-unsubscribed--你已退訂\" class=\"relative group\"\u003e✓ You\u0026rsquo;ve been unsubscribed / 你已退訂 \u003cspan class=\"absolute top-0 w-6 transition-opacity opacity-0 -start-6 not-prose group-hover:opacity-100\"\u003e\u003ca class=\"group-hover:text-primary-300 dark:group-hover:text-neutral-700\" style=\"text-decoration-line: none !important;\" href=\"#-youve-been-unsubscribed--%e4%bd%a0%e5%b7%b2%e9%80%80%e8%a8%82\" aria-label=\"Anchor\"\u003e#\u003c/a\u003e\u003c/span\u003e\u003c/h2\u003e\u003cp\u003eSorry to see you go. You won\u0026rsquo;t receive any more newsletters from this site.\u003cbr\u003e\n很可惜你要離開,你不會再收到本站的 newsletter。\u003c/p\u003e\n\u003cp\u003eIf this was a mistake or you change your mind later, you can resubscribe from any page footer.\u003cbr\u003e\n如果是誤點,或之後改變主意,可以從任一頁的 footer 重新訂閱。\u003c/p\u003e\n\u003cp\u003e\n      \n    \u003ca href=\"https://chunghaolee.com/\"\u003e← Back to home / 回首頁\u003c/a\u003e\u003c/p\u003e","title":"Unsubscribed / 已退訂"},{"content":"","date":null,"permalink":"https://chunghaolee.com/tags/networking/","section":"Tags","summary":"","title":"Networking"},{"content":"Attending the Sydney Build Expo 2026 was more than just a visit to an industry trade show for me; it was a profound observation of the future trajectory of the construction and infrastructure sectors.\n這次在 Sydney Build Expo 2026，對我來說不只是一次產業展會的參觀，更像是一場對未來建築產業與基礎設施發展方向的深度觀察。\nAs an industry professional who recently relocated from Taiwan to Sydney, equipped with a background in AI and supply chain management, what I experienced firsthand was not just the vitality of the construction industry, but two powerful forces fundamentally rewriting its rules: the practical application of AI and the structural advantages forged by supply chain clustering.\n作為一位從台灣移居悉尼、同時具備 AI 與供應鏈管理背景的產業從業者，我在現場感受到的，不只是建築產業的活力，更是兩股正在持續改寫產業規則的力量：AI 的實務落地，以及供應鏈聚落所形成的結構性優勢。\n📌 Trend 1: Beyond the Hype — Solving Real Problems with AI // AI 的價值不在於「跟風」，而在於「對焦真正的痛點」 #At this year\u0026rsquo;s expo, AI was ubiquitous. From digital tools in the design phase to smart monitoring on construction sites, numerous solutions are striving to integrate AI into the built environment. However, what truly left a lasting impression on me wasn\u0026rsquo;t the flashy tech demos, but the applications that clearly addressed industry pain points and utilized AI as a genuine problem-solving tool.\n這次展會上，AI 幾乎出現在每一個角落。從設計階段的數位工具，到施工現場的智慧監測，許多方案都在嘗試將 AI 帶入建築產業。然而，真正讓我印象深刻的，並不是那些單純強調技術展示的內容，而是那些能夠明確對應產業痛點、並把 AI 當作解決問題工具的應用。\n\u0026ldquo;AI is not the destination\u0026hellip;. AI is\u0026hellip; a powerful means to achieve goals more efficiently, accurately, and intelligently..\u0026rdquo;\n「AI 從來不是目的\u0026hellip;.而是幫助我們更高效、更精準、更聰明達成目標的手段。」\n— Beyond the Hype: Human Insight + AI is the Real Competitive Edge\nOnly when technology truly comprehends industry workflows, operational bottlenecks, and decision-making logic can its full value be unleashed. This is particularly evident in the construction sector. Whether it\u0026rsquo;s cost control and progress tracking in project management, or risk identification and safety protocols on-site, the real value doesn\u0026rsquo;t stem from the \u0026ldquo;AI\u0026rdquo; label itself. Instead, it lies in whether the technology enables teams to foresee issues earlier, make decisions faster, and allocate limited resources to where they matter most.\n只有當技術真正理解產業流程、工作瓶頸與決策邏輯時，它的價值才會被完整釋放。在建築產業中，這一點尤其明顯。無論是專案管理中的成本控制與進度追蹤，還是施工現場的風險辨識與安全管理，真正有價值的，不是 AI 這個標籤本身，而是它能否幫助團隊更早看見問題、更快完成判斷，並讓有限的資源用在更關鍵的地方。\nWhat I witnessed at the expo aligns perfectly with this direction: technology exists not for the sake of showcasing capabilities, but to make industry operations more stable, precise, and highly efficient.\n我在現場看到的，正是這樣的方向：技術不是為了展示技術，而是為了讓產業運作得更穩、更準，也更有效率。\n📌 Trend 2: The Resurgence of Supply Chain Clusters // 供應鏈聚落效應正在實體產業中再次被看見 #One specific detail in the exhibition area made me pause for a long time: a seemingly ordinary screw was backed by dozens of suppliers offering choices across different specifications, manufacturing processes, and application scenarios. This scene immediately reminded me of the classic characteristics of supply chain clusters. When suppliers, manufacturers, and technical capabilities continuously concentrate in a specific region, the resulting benefit isn\u0026rsquo;t just a price advantage; it\u0026rsquo;s a comprehensive upgrade in delivery speed, technical proficiency, collaboration efficiency, and systemic resilience.\n展區裡有一個細節讓我停留了很久：一顆看似普通的螺絲釘，背後竟然可以對應到數十家供應商，提供不同規格、不同製程與不同應用情境的選擇。這個畫面讓我很直接地想起供應鏈群聚的典型特徵——當供應商、製造商與技術能力在某個區域持續集中，帶來的就不只是價格優勢，而是交付速度、技術熟練度、協作效率與系統韌性的一整套提升。\nI’ve seen this phenomenon frequently in electric vehicle and electronics manufacturing, but witnessing it again within the context of building materials and components still struck me with profound practical significance.\n這種現象，我過去在電動車與電子產品製造領域已經見過不少，但這次在建築材料與構件的場景中再次看見，仍然讓我感受到很強的現實意義。\nFor the construction and infrastructure industries, the density and maturity of the supply chain often directly dictate whether a project can be delivered on time, on quality, and on budget. Particularly in sectors requiring high-level coordination, material dispatching, supply traceability, lead time stability, and cross-vendor collaboration are fundamentally not added values—they are the underlying pillars of project success.\n對營建與基礎設施產業來說，供應鏈的密度與成熟度，往往直接影響專案能否按時、按質、按預算落地。尤其在需要高度協調的產業裡，物料調度、供應可追溯性、交期穩定性與跨廠商協作能力，本質上都不是附加價值，而是專案成功的底層支撐。\nBecause of this, I increasingly understand why an industry\u0026rsquo;s competitiveness doesn\u0026rsquo;t solely originate from isolated technological breakthroughs, but from whether the entire ecosystem can form stable and highly efficient synergies. This mindset is exactly what I have always prioritized.\n也正因如此，我越來越能理解，為什麼一個產業的競爭力，不只是來自單點技術，而是來自整個生態系是否能形成穩定而高效的協同。這樣的思維，也正是我一直很重視的地方。\nThe clearest signal from this expo is that AI and supply chains are jointly rewriting the way the industry operates.\n這次展會最清楚的訊號，是 AI 與供應鏈正在一起改寫產業的運作方式。\nAI makes decision-making smarter, while the supply chain makes execution more stable; AI elevates insight and efficiency, while the supply chain provides structure and resilience. When these two converge, the industry transforms not just its tools, but its entire operational logic and the very composition of the value chain. This is precisely the direction I am most passionate about.\nAI 讓決策更聰明，供應鏈讓執行更穩定；AI 提升的是洞察與效率，供應鏈提供的是結構與韌性。當這兩者結合時，產業改變的就不只是工具，而是整個工作邏輯與價值鏈的組成方式。這也是我特別感興趣的方向。\nI have always believed that truly valuable industry innovation is not merely about the technology itself, but whether that technology can be deployed in the right scenarios, forming a complete system with workflows, resources, supply, and execution capabilities. This applies not only to the construction industry but equally to future data centers, AI infrastructure, and broader high-density operational environments.\n我一直相信，真正有價值的產業創新，不只是技術本身，而是技術能否被放進正確的場景裡，並與流程、資源、供應與執行能力形成完整的系統。這不只適用於建築產業，也同樣適用於未來的數據中心、AI 基礎設施，以及更廣泛的高密度營運場景。\nTo me, the Sydney Build Expo 2026 was more than an exhibition; it was a reminder: as AI and supply chains truly penetrate the core of the industry, the next phase of competitiveness will no longer be about who owns the technology, but who can best integrate technology, resources, and execution.\n對我而言，Sydney Build Expo 2026 不只是一場展會，更像是一個提醒：當 AI 與供應鏈開始真正深入產業核心，下一階段的競爭力，將不再只是誰擁有技術，而是誰能把技術、資源與執行整合得更好。\nConsequently, I deeply cherish the insights and networking opportunities I gained at the Sydney Build Expo 2026. Seeing vendors, technologies, and resources from all over the world converge on a single platform not only highlighted the openness and dynamism of Australian industries, but also fueled my deep anticipation for future cross-disciplinary collaboration and development.\n也因此，我非常珍惜這次在Sydney Build Expo 2026的觀察與交流機會。能夠看到來自世界各地的廠商、技術與資源在同一個平台上匯聚，不僅讓人感受到澳洲產業的開放與活力，也讓我對未來更多跨領域的合作與發展，抱有更深的期待。\n© Chung-Hao Lee. All Rights Reserved. All content on this webpage—including but not limited to text, images, design, code, and multimedia materials—is protected under the international copyright treaties. Unauthorized reproduction, modification, distribution, public transmission, or commercial use is strictly prohibited. Legal action will be taken against infringement. © 李崇豪。保留所有權利。 本網頁之內容（包括但不限於文字、圖片、設計、程式碼及多媒體素材）均受國際著作權條約保護。未經書面授權，嚴禁任何形式之複製、改作、散布、公開傳輸或商業利用。侵權者將依法追訴。\n","date":"May 3, 2026","permalink":"https://chunghaolee.com/posts/sydney-build-expo-2026/","section":"Posts","summary":"\u003cp\u003eAttending the \u003ca href=\"https://www.sydneybuildexpo.com/\" target=\"_blank\" rel=\"noreferrer\"\u003eSydney Build Expo 2026\u003c/a\u003e was more than just a visit to an industry trade show for me; it was a profound observation of the future trajectory of the construction and infrastructure sectors.\u003c/p\u003e\n\u003cp\u003e這次在 Sydney Build Expo 2026，對我來說不只是一次產業展會的參觀，更像是一場對未來建築產業與基礎設施發展方向的深度觀察。\u003c/p\u003e\n\u003cp\u003eAs an industry professional who recently relocated from Taiwan to Sydney, equipped with a background in AI and supply chain management, what I experienced firsthand was not just the vitality of the construction industry, but \u003cstrong\u003etwo powerful forces fundamentally rewriting its rules: the practical application of AI and the structural advantages forged by supply chain clustering.\u003c/strong\u003e\u003c/p\u003e","title":"Sydney Build Expo 2026: AI and Supply Chain Strength as the Next Competitive Edge"},{"content":"","date":"April 26, 2026","permalink":"https://chunghaolee.com/posts/external-article---%E5%85%A8%E5%93%A1%E8%BF%BD%E6%B1%82%E6%9C%80%E5%A4%A7%E5%8C%96token%E4%B8%80%E5%A0%B4%E7%84%A1%E4%BA%BA%E6%95%A2%E5%81%9C%E4%B8%8B%E7%9A%84%E8%BB%8D%E5%82%99%E7%AB%B6%E8%B3%BD/","section":"Posts","summary":"","title":"[External Article] - 全員Token-maxxing，一場無人敢停下的軍備競賽"},{"content":" The release of DeepSeek V4 has undoubtedly sent shockwaves through the technology community. While most analyses focus on leaps in model capabilities, architectural innovations, or expanded application scenarios, we propose a different perspective: this time, let\u0026rsquo;s shift our gaze from \u0026ldquo;what the model can provide\u0026rdquo; to \u0026ldquo;what is provided to the model.\u0026rdquo;\nDeepSeek V4 的發布，無疑在技術圈投下了一顆震撼彈。當多數分析聚焦於模型能力的躍升、架構的創新或應用場景的拓展時，我們想換一個視角：這一次，讓我們把目光從「模型能提供什麼」，轉向「什麼提供給模型」。\nThe technical pipeline of large language models can be broadly divided into two ends: inference and training. The story on the inference side has already been told brilliantly by DeepSeek V4: when executing the same task, V4-Pro requires only 27% of the FLOPs per token for single-token inference compared to V3.2, while KV Cache occupancy is compressed to just 10% [1]. This means that the same hardware resources can now handle denser inference requests—or, put differently, the barrier to achieving equivalent inference performance has been substantially lowered.\n大語言模型的技術鏈路，粗淺可拆為推理（Inference）與訓練（Training）兩端。推理端的故事，DeepSeek V4 已經講得足夠精彩：V4-Pro 在執行同一任務時，對單 Token 推理的 FLOPs 需求僅為 V3.2 的 27%，KV Cache 佔用更壓縮至 10%[1]。這意味著，同樣的硬體資源，現在能承載更密集的推理請求；或者说，達成同樣的推理效果，門檻已大幅降低。\nIntuitively, this might seem to signal a \u0026ldquo;subtraction\u0026rdquo; in compute demand. However, the Jevons Paradox from economics reminds us that when resource efficiency improves and unit costs decline, total consumption often does not decrease—instead, it may rise due to the explosion of application scenarios and expansion of the user base. The efficiency leap on the inference side is likely opening the door to far broader applications.\n直覺上，這似乎預示著算力需求的「減法」。但經濟學裡的「傑文斯悖論」（Jevons Paradox）提醒我們：當資源使用效率提升、單次成本下降，往往不會導致總消耗減少，反而可能因為應用場景的爆發與用戶規模的擴張，推高整體需求。推理端的效率躍遷，很可能正打開一扇更廣闊的應用之門。\nYet today, we want to focus on the other end: training.\n然而，今天我們想聚焦的，是另一端——訓練。\nScale Laws Still Hold: The \u0026ldquo;Arms Race\u0026rdquo; in Parameters and Data // 規模法則依然有效：參數與數據的「軍備競賽」 #Reviewing DeepSeek\u0026rsquo;s version evolution reveals a clear trajectory:\n翻開 DeepSeek 的版本演進表，一條清晰的曲線浮現：\nModel Version\n模型版本 Release Date\n發布時間 Total Parameters\n總參數量 Pre-training Data Volume\n預訓練數據量 DeepSeek-V2 2024-05-06 236B 8.1T tokens DeepSeek-V3 2024-12-26 671B 14.8T tokens DeepSeek-V4-Pro 2026-04-24 1.6T (1600B) 33T tokens Parameters have surged from 236B to 1.6T, while pre-training data has expanded from 8.1T to 33T tokens. This is not merely numerical growth—it represents iterative capability enhancement. This trend resonates intriguingly with findings in neuroscience: the cognitive capacity of primates correlates nearly linearly with neuron count and brain volume [2][3]. In the realm of artificial intelligence, Scale Laws have not only remained valid but continue to raise the compute barrier with even greater force.\n參數量從 236B 躍升至 1.6T，預訓練數據量從 8.1T tokens 擴張至 33T tokens。這不僅是數字的膨脹，更是能力的疊代。這與腦科學研究的結論形成有趣呼應：靈長類動物的認知能力，與其神經元數量、腦容量呈現近線性相關[2][3]。在人工智慧的領域，Scale Law 非但沒有失效，反而以更強悍的姿態，持續拉升著算力的壁壘。 The stronger the model, the \"heavier\" the training. This leads to a direct business implication: **resource demands during the training phase are escalating along an exponential curve.** 模型越強，訓練越「重」。這帶來一個直接的商業命題：訓練階段的資源需求，正在以指數級曲線攀升。\nThe Structure of Training Costs: More Than Just a Chip Game // 訓練成本的結構：不只是芯片的遊戲 #According to the influential report The Rising Costs of Training Frontier AI Models [4], which analyzes pre-training costs for large models (assuming self-built infrastructure), the cost structure breaks down approximately as follows:\n根據 著名的《THE RISING COSTS OF TRAINING FRONTIER AI MODELS》[4]對大模型預訓練成本的分析（假設自建基礎設施），成本結構大致如下：\nChips 芯片 (GPU/ASIC): 44% Server components 服務器組件 (CPU, memory, etc.): 29% Networking equipment 網絡設備: 17% Energy 能源: 9% At first glance, chips remain the largest component. However, when server components, networking, and energy are combined, they account for over 55%. More critically, as cluster scales expand from thousands to tens of thousands of GPUs, the marginal cost weight of power supply, thermal cooling, and network topology will only continue to rise.\n乍看之下，芯片仍是最大頭。但若將服務器組件、網絡與能源相加，占比已超過 55%。更關鍵的是，隨著集群規模從千卡、萬卡走向十萬卡，電力供應、散熱冷卻、網絡拓撲的邊際成本權重，只會越來越高。\nHere lies a frequently overlooked characteristic: training workloads are inherently insensitive to latency. Unlike inference, which requires millisecond-level responsiveness, training can tolerate longer task cycles. This means training infrastructure need not compete for space in traditional cloud hubs characterized by tight power grids, geopolitical congestion, and elevated costs. The siting of training clusters now has room to \u0026ldquo;vote with their feet\u0026rdquo;—locations with lower geopolitical risk, more controllable energy costs, and greater long-term policy stability may well become the next-generation footholds for AI infrastructure.\n這裡有一個常被忽略的特性：訓練任務對延時（Latency）天生不敏感。與推理端需要毫秒級響應不同，訓練可以容忍更長的任務週期，這意味著它不必擠在用電緊張、地緣擁擠、成本高昂的傳統雲端樞紐。訓練集群的選址，從此有了「用腳投票」的空間——哪裡的地緣風險更低、能源成本更可控、長期政策更穩定，哪裡就可能成為下一代 AI 基礎設施的落腳點。\nFrom Technical Marvel to Commercial Essence: The Coordinates of Extreme ROI // 從技術奇蹟到商業本質：極致投入產出比的坐標 #In the real commercial world, scientific marvels achieved regardless of cost may make it into academic papers, but they rarely reshape the course of civilization. What truly drives technological adoption and industrial transformation has always been extreme return on investment (ROI). When training costs become a core constraint on model iteration, site selection ceases to be merely a technical decision for IT departments—it becomes a strategic imperative central to long-term competitiveness.\n在真實的商業世界裡，不計成本的科學奇蹟，或許能寫進論文，但很難改寫文明進程。真正推動技術普及與產業變革的，從來是極致的投入產出比（ROI）。當訓練成本成為模型迭代的核心約束，選址就不再只是 IT 部門的技術決策，而是關乎長期競爭力的戰略命題。\nThus, a long-underestimated coordinate begins to emerge: Sydney.\n於是，一個長期被低估的坐標開始浮現：悉尼。\n### Why Sydney? // 為什麼是悉尼？ A Geopolitical \u0026ldquo;Safe Harbor\u0026rdquo; Australia offers a mature rule of law, stable governance, and—amid today\u0026rsquo;s increasingly tense global landscape—a geographic position distant from major conflict hotspots. It remains one of the few developed economies capable of mitigating geopolitical risk simultaneously on multiple fronts. For AI training infrastructure, which entails long construction cycles and substantial capital commitment, policy certainty and asset security represent more foundational strategic considerations than short-term cost savings.\n地緣政治的「避風港」 澳大利亞具備成熟法治、穩定政局，且在今日世界局勢日益緊張的當下有遠離主要衝突熱點的地理位置，是全球少數能同時規避地緣風險的發達經濟體。對於建設週期長、資本投入大的 AI 訓練基礎設施，政策確定性與資產安全，是比短期成本更底層的戰略考量。\nDual Advantages of Renewable Energy: Cost and Compliance Australia\u0026rsquo;s abundant solar and wind resources support competitively priced renewable energy. This not only directly reduces the energy expenditure component of training costs but also aligns with global technology enterprises\u0026rsquo; commitments to carbon reduction. In the context of green supply chain trends, this serves as an implicit passport to future-proof operations.\n可再生能源的成本與合規雙重優勢 澳洲豐富的太陽能與風能資源，可支援具價格競爭力的可再生能源。這不僅直接降低訓練成本中的能源支出，更契合全球科技企業對節能減碳的要求，在綠色供應鏈趨勢下成為隱形通行證。\nClimate-Driven Cooling Dividends Sydney\u0026rsquo;s temperate maritime climate significantly extends the periods during which data centers can leverage natural cooling, markedly reducing reliance on mechanical refrigeration. For large-scale clusters where thermal management costs constitute an increasingly significant portion of operational expenditure, this translates into tangible efficiency gains and long-term cost resilience.\n氣候賦予的冷卻紅利 悉尼溫和的海洋性氣候，大幅延長數據中心可利用自然冷卻的時段，顯著降低對機械製冷的依賴。對於散熱成本佔比日益提升的大規模集群，這意味著可觀的運營效率優勢與長期成本韌性。\nRobust Infrastructure and Talent Ecosystem As a leading technology and financial hub in the Asia-Pacific region, Sydney boasts a comprehensive submarine cable network enabling efficient connectivity to major global markets. Concurrently, its mature pool of engineering talent and established data center industry ecosystem can provide end-to-end support—from design and construction to operations and maintenance—for ultra-scale AI clusters.\n基礎設施與人才生態的完備性 作為亞太科技與金融樞紐，悉尼擁有完善的海底光纜網絡，可高效連接全球主要市場；同時，當地成熟的工程技術人才與數據中心產業鏈，能為超規模 AI 集群提供從設計、建設到運維的全週期支援。\nConclusion: The Next Stop in the Compute Arms Race // 結語：算力軍備競賽的下一站 #As the AI competition enters a new phase of \u0026ldquo;compute arms race,\u0026rdquo; the decisive factors are no longer limited to innovations in model architecture—they increasingly hinge on strategic infrastructure deployment. DeepSeek V4, with its 1.6 trillion parameters and 33 trillion tokens of training data, reaffirms the resilience of Scale Laws and compels the industry to reconsider the cost structure and geographic logic underpinning the training phase.\n當 AI 競賽進入「算力軍備競賽」的新階段，勝負手不再只是模型架構的創新，更是基礎設施的戰略佈局。DeepSeek V4 用 1.6 萬億參數與 33 萬億 tokens 的事實，再次驗證了 Scale Law 的韌性，也倒逼產業重新思考訓練端的成本結構與地理邏輯。\nTechnological breakthroughs must ultimately return to commercial fundamentals. Extreme ROI remains the authentic engine driving civilizational progress. And when compute power becomes the new era\u0026rsquo;s \u0026ldquo;oil,\u0026rdquo; the locations capable of refining it more safely, economically, and sustainably are precisely those poised to become the nexuses of the next AI wave.\n技術的突破，終究要回歸商業的本質。極致的投入產出比，才是驅動文明進程的真實引擎。而當算力成為新時代的「石油」，哪裡能更安全、更經濟、更可持續地提煉它，哪裡就可能成為下一輪 AI 浪潮的樞紐。\nSydney may well be positioned at the coordinates of that nexus.\n悉尼，或許就在這個樞紐的坐標上。\nReferences / 參考資料\n[1] DeepSeek-AI, DeepSeek-V4: Towards Highly Efficient Million-Token Context Intelligence https://huggingface.co/deepseek-ai/DeepSeek-V4-Pro/blob/main/DeepSeek_V4.pdf\n[2] Herculano-Houzel, S., The Human Brain in Numbers: A Linearly Scaled-up Primate Brain https://pmc.ncbi.nlm.nih.gov/articles/PMC2776484/\n[3] Manger, P. R. et al., The Evolution of Large Brain Size in Mammals: The \u0026lsquo;Over-700-Gram Club Quartet\u0026rsquo; https://doi.org/10.1159/000351975\n[4] Epoch AI, The Rising Costs of Training Frontier AI Models https://arxiv.org/pdf/2405.21015\n© Chung-Hao Lee. All Rights Reserved. All content on this webpage—including but not limited to text, images, design, code, and multimedia materials—is protected under the international copyright treaties. Unauthorized reproduction, modification, distribution, public transmission, or commercial use is strictly prohibited. Legal action will be taken against infringement. © 李崇豪。保留所有權利。 本網頁之內容（包括但不限於文字、圖片、設計、程式碼及多媒體素材）均受國際著作權條約保護。未經書面授權，嚴禁任何形式之複製、改作、散布、公開傳輸或商業利用。侵權者將依法追訴。\n","date":"April 25, 2026","permalink":"https://chunghaolee.com/posts/passport-to-ai-era-special-editionthe-global-ai-training-pivot-positioning-sydney-as-a-strategic-infrastructure-nexus/","section":"Posts","summary":"\u003cp\u003e\n\n\n\n\n\n\u003cfigure\u003e\n    \n    \n\n\n\n\n\n\n\n\n  \n    \u003cpicture\n      class=\"mx-auto my-0 rounded-md\"\n      \n    \u003e\n      \n      \n      \n      \n        \u003csource\n          \n            srcset=\"https://chunghaolee.com/posts/passport-to-ai-era-special-editionthe-global-ai-training-pivot-positioning-sydney-as-a-strategic-infrastructure-nexus/pic_hu_b9f073a7cfb0331d.webp 330w,https://chunghaolee.com/posts/passport-to-ai-era-special-editionthe-global-ai-training-pivot-positioning-sydney-as-a-strategic-infrastructure-nexus/pic_hu_9d7695bdf86de19d.webp 660w\n            \n              ,https://chunghaolee.com/posts/passport-to-ai-era-special-editionthe-global-ai-training-pivot-positioning-sydney-as-a-strategic-infrastructure-nexus/pic_hu_59dd3c6620c54190.webp 1024w\n            \n            \n              ,https://chunghaolee.com/posts/passport-to-ai-era-special-editionthe-global-ai-training-pivot-positioning-sydney-as-a-strategic-infrastructure-nexus/pic_hu_8e0c00f6c92cc64f.webp 1320w\n            \"\n          \n          sizes=\"100vw\"\n          type=\"image/webp\"\n        /\u003e\n      \n      \u003cimg\n        width=\"1448\"\n        height=\"1086\"\n        class=\"mx-auto my-0 rounded-md\"\n        \n        loading=\"lazy\" decoding=\"async\"\n        \n          src=\"https://chunghaolee.com/posts/passport-to-ai-era-special-editionthe-global-ai-training-pivot-positioning-sydney-as-a-strategic-infrastructure-nexus/pic_hu_4f0e9576f13ff6a6.png\" srcset=\"https://chunghaolee.com/posts/passport-to-ai-era-special-editionthe-global-ai-training-pivot-positioning-sydney-as-a-strategic-infrastructure-nexus/pic_hu_bf68b90ae7d19e19.png 330w,https://chunghaolee.com/posts/passport-to-ai-era-special-editionthe-global-ai-training-pivot-positioning-sydney-as-a-strategic-infrastructure-nexus/pic_hu_4f0e9576f13ff6a6.png 660w\n          \n            ,https://chunghaolee.com/posts/passport-to-ai-era-special-editionthe-global-ai-training-pivot-positioning-sydney-as-a-strategic-infrastructure-nexus/pic_hu_dd1dd99ba90b6de8.png 1024w\n          \n          \n            ,https://chunghaolee.com/posts/passport-to-ai-era-special-editionthe-global-ai-training-pivot-positioning-sydney-as-a-strategic-infrastructure-nexus/pic_hu_8f8083d0f0bf5ed1.png 1320w\n          \"\n          sizes=\"100vw\"\n        \n      /\u003e\n    \u003c/picture\u003e\n  \n\n\n\u003c/figure\u003e\n\u003c/p\u003e\n\u003cp\u003eThe release of DeepSeek V4 has undoubtedly sent shockwaves through the technology community. While most analyses focus on leaps in model capabilities, architectural innovations, or expanded application scenarios, we propose a different perspective: this time, let\u0026rsquo;s shift our gaze from \u0026ldquo;what the model can provide\u0026rdquo; to \u0026ldquo;what is provided to the model.\u0026rdquo;\u003c/p\u003e","title":"【Passport to AI Era Special Edition】The Global AI Training Pivot: Positioning Sydney as a Potential Strategic Infrastructure Nexus"},{"content":"","date":null,"permalink":"https://chunghaolee.com/tags/data-center/","section":"Tags","summary":"","title":"Data Center"},{"content":"","date":null,"permalink":"https://chunghaolee.com/tags/passport-to-ai-era/","section":"Tags","summary":"","title":"Passport to AI Era"},{"content":"","date":"April 24, 2026","permalink":"https://chunghaolee.com/posts/external-article-%E6%BC%B2%E5%83%B9%E9%80%B2%E8%A1%8C%E5%BC%8F%E4%B9%9D%E5%A4%A7%E8%A1%8C%E6%A5%AD%E7%9A%84%E5%B8%B3%E6%9C%AC/","section":"Posts","summary":"","title":"[External Article] 漲價進行式：九大行業的帳本"},{"content":"In the AI era, following my previous deep dive into the truly critical strategic moats, this series of reports will guide you through a systematic study of Data Center (DC) analytics. As the inaugural piece of this series, I will take a macro perspective to examine the global DC market landscape, analyzing it primarily through the fundamental framework of supply and demand.\nIn an age of information overload where AI-generated content is ubiquitous, true human value lies in synthesizing massive amounts of data to extract valuable insights. This report begins by examining the demand side, moves to the supply side, and concludes with key insights and future research directions, aiming to provide a clear and valuable reference.\n在 AI 時代，繼先前文章一起深入探討真正重要的護城河之後，本系列報告將陪伴大家系統性地研究資料中心（Data Center，以下簡稱 DC）的各項分析。這篇作為系列的開頭，我們用宏觀的角度來看看全球 DC 市場的全貌，主要從供給與需求的基本框架出發進行分析。\nAI 時代資訊爆炸、內容到處都是，而人的真正價值就在於把這些海量訊息彙整起來，並整理出真正有用的洞見。本報告先從需求端開始看起，接著檢視供給端，最後分享一些關鍵洞見與後續的研究方向，希望能為大家帶來清晰的參考。\nDemand-Side Analysis / 需求端分析 #As the market adage goes, \u0026ldquo;Supply always comes on the heels of demand.\u0026rdquo; Currently, for Transformer-based generative AI, the fundamental unit of computation is the token. Consequently, token usage volume serves as a highly representative proxy for current AI demand intensity. While it may not capture 100% of the real-world picture, OpenRouter\u0026rsquo;s token usage analysis offers significant reference value.\n正如市場上大家常說的「Supply always comes on the heels of demand」，供給永遠是跟在需求後面來的。目前以 Transformer 為主的生成式 AI，最小的運算單位就是 token，因此 token 的使用量可以很代表性地反映出當前 AI 的需求強度。雖然無法 100% 涵蓋所有實際情況，但 OpenRouter 的 token 使用分析已經很有參考價值。\nAccording to OpenRouter\u0026rsquo;s research [1], global token usage has exhibited exponential growth from 2024 through the end of 2025. This clearly indicates that worldwide demand for AI is currently in a phase of explosive hyper-growth.\n根據 OpenRouter 的研究[1]，從 2024 年到 2025 年底，全球 token 使用量呈現指數式成長，這清楚顯示全世界對 AI 的需求正處於超額爆發的階段。\nThe same report highlights distinct regional variations: the Asia-Pacific (APAC) region\u0026rsquo;s share of global token usage continues to climb, demonstrating that AI demand growth in APAC is significantly outpacing the global average.\n同一份報告也讓我們看到明顯的區域差異：亞太地區（APAC）在全球 token 使用佔比中持續上升，顯示亞太區的 AI 需求成長速度明顯高於全球平均。\nFurthermore, Stanford University\u0026rsquo;s AI Index Report [2] points out that APAC\u0026rsquo;s AI adoption rate in 2024 remained relatively low compared to North America (72% vs. 82%). This suggests substantial headroom for future growth in the APAC region.\n此外，史丹佛大學的分析報告[2]指出，2024 年亞太區的 AI 使用佔比相較北美仍然偏低（72% 對比 82%），這也意味著亞太區未來還有很大的成長空間。\nIt is particularly noteworthy that due to the unique situation of the Chinese market, reports from OpenRouter and Stanford often underestimate China\u0026rsquo;s actual contributions. This underlying factor implies that APAC\u0026rsquo;s overall growth momentum is even stronger than it appears on paper.\n特別值得注意的是，由於中國市場的特殊性，OpenRouter 和史丹佛的報告往往會低估中國的實際貢獻，這反而讓亞太區的整體成長動能看起來更強勁。\nKey Demand-Side Insights / 需求端關鍵洞察 # Global AI demand is experiencing explosive, exponential growth.\n全球 AI 需求呈現爆炸式的指數型增長。\nSince 2024, APAC\u0026rsquo;s share of AI demand has steadily increased, solidifying its position as the world\u0026rsquo;s second-largest region for AI demand.\n從 2024 年開始，亞太區的 AI 需求佔比逐步提升，已成為全球第二大 AI 需求區域。\nThe APAC region retains massive potential for future growth in AI demand.\n亞太區對 AI 的需求在未來仍有相當大的成長潛力。\nSupply-Side Analysis / 供給端分析 #According to CBRE\u0026rsquo;s annual trend report [3], markets worldwide have aggressively expanded their DC construction scale over the past two years, with 2025 supply volumes substantially exceeding 2023 levels. North America boasts the highest supply growth rate, which aligns perfectly with our demand-side analysis and confirms it as the region with the most robust AI demand over the last two years.\n根據 CBRE 的年度分析報告[3]，全球各地在過去兩年都積極擴大 DC 的建設規模，2025 年的 DC 供給量相較 2023 年都有大幅增加。其中北美地區的供給增速最高，這也跟前面需求端的分析互相呼應，確實印證北美是近兩年 AI 需求最旺盛的區域。\nMeanwhile, supply volumes and growth rates in Europe and the APAC region remain roughly on par with each other.\n歐洲與亞太區的供給量和增速則大致相當。\nLooking at DC vacancy rates—where lower rates indicate supply struggling to keep pace with demand—North America remains the most intensely constrained market, followed by Europe, and then APAC.\n如果從 DC 空置率來看（空置率越低，就代表供給越追不上需求），北美目前依然是需求最強烈的地區，其次是歐洲，再來是亞太區。\nInterestingly, within the APAC region, Sydney\u0026rsquo;s overall vacancy rate dropped significantly in 2025 compared to 2023, with a steeper decline than other major hubs like Singapore and Tokyo. This reveals exceptionally robust demand for DCs in Sydney over the past two years, with supply evidently lagging behind.\n有趣的是，在亞太區裡，悉尼的整體空置率在 2025 年相較 2023 年明顯下降，而且降幅高於新加坡、東京等其他主要市場，這顯示悉尼近兩年對 DC 的需求非常旺盛，供給明顯跟不上腳步。\nIn terms of rental costs, highly populated, land-constrained metropolitan areas (such as Tokyo, Hong Kong, and Singapore) command significantly higher monthly pricing than other regions. Conversely, cost levels across North America, Europe, and Sydney show no substantial disparities.\n租金成本方面，土地資源有限、人口密集的大型城市（如東京、香港、新加坡）的月租金相對其他地區明顯偏高；而北美、歐洲以及悉尼的租金水準則沒有太大的差距。\nKey Supply-Side Insights / 供給端關鍵洞見 # Despite aggressive efforts on the supply side to keep up with explosive global AI demand, the market remains fundamentally undersupplied, driving a continuous decline in DC vacancy rates.\n雖然全球 AI 需求呈爆炸式增長，供給端已經很努力跟進，但整體來說還是供不應求，導致 DC 的空置率持續下降。\nNorth America maintains its position as the largest region for global supply, closely followed by Europe and APAC.\n北美依然是全球供給規模最大的地區，歐洲與亞太區緊隨其後。\nBased on monthly rental rates (representing supply costs), densely populated cities with limited footprints (e.g., Singapore, Tokyo, Hong Kong) incur higher DC supply costs due to exorbitant land premiums.\n以月租金（也就是供給成本）來看，地小人稠的大型城市（如新加坡、東京、香港）因為土地成本高，DC 的供給成本也相對較高。\nNote / 注意事項: #This current analysis has yet to incorporate other critical markets such as China, Malaysia, and Thailand. These regions exhibit equally robust supply pipelines paired with highly advantageous cost structures, making them essential considerations for future analyses. 這份分析目前尚未納入中國、馬來西亞及泰國等其他重要市場。這些地區同樣供給旺盛，而且成本結構相對具有優勢，未來在分析時絕對不能忽略。\nKey Insights and Future Research Directions #Synthesizing both demand and supply-side analyses, it is evident that the APAC region will be the most powerful growth engine for future AI supply and demand. Consequently, subsequent reports in this series will focus squarely on the APAC region (with particular emphasis on Singapore and Sydney). I will conduct deep-dive explorations into critical topics including power supply, cooling systems, clean energy, policy environments, site selection strategies, and supply chain.\n關鍵洞見與未來研究方向 #綜合需求與供給端的分析，我們可以清楚看到，亞太區域將會是未來 AI 需求與供給最強勁的成長引擎。因此，本系列後續的報告會把焦點放在亞太區域（特別是新加坡與悉尼），進行更深入的專題探討，涵蓋電力供應、冷卻系統、潔淨能源、政策環境、選址策略、供應鏈等重要主題。\nReferences / 參考資料 #[1] OpenRouter, State of AI 2025: 100T Token LLM Usage Study https://openrouter.ai/state-of-ai\n[2] Stanford University, AI Index Report 2025 https://hai.stanford.edu/assets/files/hai_ai_index_report_2025.pdf\n[3] CBRE, Global Data Center Trends 2025 https://www.cbre.com/insights/reports/global-data-center-trends-2025\n© Chung-Hao Lee. All Rights Reserved. All content on this webpage—including but not limited to text, images, design, code, and multimedia materials—is protected under the international copyright treaties. Unauthorized reproduction, modification, distribution, public transmission, or commercial use is strictly prohibited. Legal action will be taken against infringement. © 李崇豪。保留所有權利。 本網頁之內容（包括但不限於文字、圖片、設計、程式碼及多媒體素材）均受國際著作權條約保護。未經書面授權，嚴禁任何形式之複製、改作、散布、公開傳輸或商業利用。侵權者將依法追訴。\n","date":"April 19, 2026","permalink":"https://chunghaolee.com/posts/passport-to-ai-era-1-overview-ai-market/","section":"Posts","summary":"\u003cp\u003eIn the AI era, following \u003ca href=\"/posts/beyond-the-model-the-invisible-moat-of-ai-infrastructure--data-center/\"\u003emy previous deep dive\u003c/a\u003e into the truly critical strategic moats, this series of reports will guide you through a systematic study of Data Center (DC) analytics. As the inaugural piece of this series, I will take a macro perspective to examine the global DC market landscape, analyzing it primarily through the fundamental framework of supply and demand.\u003c/p\u003e\n\u003cp\u003eIn an age of information overload where AI-generated content is ubiquitous, true human value lies in synthesizing massive amounts of data to extract valuable insights. This report begins by examining the demand side, moves to the supply side, and concludes with key insights and future research directions, aiming to provide a clear and valuable reference.\u003c/p\u003e","title":"【Passport to AI Era vol.1】Overview AI Market"},{"content":"In an era of increasingly volatile global geopolitics, risk is no longer just a headline; it has become a tangible force affecting everyone’s life and corporate operations. Inflationary pressures, supply chain disruptions, and energy price fluctuations—these impacts have long since extended from conflict points on a map into our daily decisions and commercial strategies. As someone who has long monitored global affairs, I have come to a profound realization: the ability to grasp geopolitical risks in a timely and accurate manner has become indispensable for supply chain management and business strategy.\n在當今世界地緣政治環境日益動盪的時代，風險不再只是新聞標題，而是真實影響到每個人的生活與企業運作。通膨壓力、供應鏈中斷、能源價格波動……這些影響早已從地圖上的衝突點，延伸到我們的日常決策與商業布局。作為一名長期關注全球局勢的我，深刻體會到：即時、精準掌握地緣風險，已成為供應鏈管理與商業策略不可或缺的能力。\nFor this reason, I no longer passively consume information. Instead, I have proactively collaborated with AI to build a tool entirely my own—My Daily Digest: Geopolitics Risk.\n正因如此，我不再被動接收資訊，而是主動與AI合作，打造了一款完全屬於自己的工具——My Daily Digest - Geopolitics Risk（每日地緣風險摘要）。\nEvery morning at 8:00 AM, this tool automatically generates a structured, high-quality geopolitical risk briefing for me. It distills the content I care about most from reliable, world-class media sources such as Reuters, Financial Times, SCMP, Bloomberg, and Nikkei Asia:\n🌍 Focus on Key Regions: Systematically tracking high-risk hotspots in East Asia, the Middle East, Europe, and the Americas. 📉 Linked Market Insights: Real-time analysis of the potential impact of these geopolitical events on the supply chain.\n這個工具每天早上8點會自動為我生成一份結構化、高品質的地緣風險簡報。它會從路透社、金融時報、南華早報、彭博社、日經新聞等國際一流媒體的可靠來源中，精煉出我最關心的內容：\n🌍 聚焦關鍵區域： 結構化追蹤東亞, 中東與歐美等高風險熱點。 📉 連動市場判斷： 即時分析這些地緣風險事件對供應鏈的潛在影響。\nAll content is presented in a calm, fact-oriented style, free from hype or unnecessary emotion, allowing me to see the full picture in the shortest possible time.\n所有內容以冷靜、事實導向的風格呈現，沒有炒作、沒有多餘情緒，讓我能在最短時間內看清全貌。\nIn today’s information explosion, we no longer face a \u0026ldquo;lack of information\u0026rdquo; but rather \u0026ldquo;too much noise.\u0026rdquo; This tool allows me to truly make information work for me: I can customize the format, regional focus, and analytical perspective based on my needs, fully realizing an automated workflow from raw RSS feeds to daily summaries with the help of AI. The entire process requires zero maintenance; with just a few minutes each day, I gain strictly vetted, high-value insights.\n在資訊爆炸的今天，我們面臨的已不再是「資訊不足」，而是「雜音太多」。這個工具讓我真正做到「資訊為我所用」：我可以根據自身需求客製化格式、區域重點與分析角度，完全由AI協助實現從原始RSS到每日摘要的自動化流程。整個過程零維護，每天只需幾分鐘，就能獲得經過嚴選、去蕪存菁的高價值洞察。\nTo me, this is more than just a personal gadget; it is an extension of professional capability. It helps me maintain a forward-looking perspective in supply chain risk assessment, business decision-making, and strategic planning. When global events threaten to affect raw material prices, logistics routes, or regional investment environments, I am already equipped to make faster judgments.\n對我而言，這不只是個人小工具，更是專業能力的延伸。它幫助我在供應鏈風險評估、商業決策與戰略規劃上，保持領先一步的視野。當全球事件可能影響到原物料價格、物流路線或區域投資環境時，我已能更快做出判斷。\nThe development process has further convinced me that in the AI era, proactively solving pain points and building your own solutions is the most precious competitive edge for any professional. I do not wait for off-the-shelf tools; instead, I choose to work hand-in-hand with AI to rapidly transform needs into actual value.\n這次開發的過程，也讓我更加相信：在AI時代，主動解決痛點、自己動手打造解決方案，正是每位專業人士最珍貴的競爭力。我不等待現成工具，而是選擇與AI攜手，快速將需求轉化為實際價值。\nJoin the Telegram Channel Daily Geopolitics Digest to stay informed with real-time insights.\n歡迎加入在Telegram上的頻道Daily Geopolitics Digest，獲取即時新知。\n© Chung-Hao Lee. All Rights Reserved. All content on this webpage—including but not limited to text, images, design, code, and multimedia materials—is protected under the international copyright treaties. Unauthorized reproduction, modification, distribution, public transmission, or commercial use is strictly prohibited. Legal action will be taken against infringement. © 李崇豪。保留所有權利。 本網頁之內容（包括但不限於文字、圖片、設計、程式碼及多媒體素材）均受國際著作權條約保護。未經書面授權，嚴禁任何形式之複製、改作、散布、公開傳輸或商業利用。侵權者將依法追訴。\n","date":"April 16, 2026","permalink":"https://chunghaolee.com/posts/my-daily-digest-geopolitics-risk/","section":"Posts","summary":"\u003cp\u003eIn an era of increasingly volatile global geopolitics, risk is no longer just a headline; it has become a tangible force affecting everyone’s life and corporate operations. Inflationary pressures, supply chain disruptions, and energy price fluctuations—these impacts have long since extended from conflict points on a map into our daily decisions and commercial strategies. As someone who has long monitored global affairs, I have come to a profound realization: \u003cstrong\u003ethe ability to grasp geopolitical risks in a timely and accurate manner has become indispensable for supply chain management and business strategy.\u003c/strong\u003e\u003c/p\u003e","title":"【AI Side Project Vol.06】My Daily Digest - Geopolitics Risk 🌏"},{"content":"","date":null,"permalink":"https://chunghaolee.com/tags/ai-practice-journal/","section":"Tags","summary":"","title":"AI Practice Journal"},{"content":"","date":null,"permalink":"https://chunghaolee.com/tags/app/","section":"Tags","summary":"","title":"App"},{"content":"","date":null,"permalink":"https://chunghaolee.com/tags/side_project/","section":"Tags","summary":"","title":"Side_Project"},{"content":"","date":null,"permalink":"https://chunghaolee.com/tags/supply-chain/","section":"Tags","summary":"","title":"Supply Chain"},{"content":"","date":null,"permalink":"https://chunghaolee.com/tags/using-ai/","section":"Tags","summary":"","title":"Using AI"},{"content":"Product professionals naturally focus on user experience and output. It wasn’t until I began diving into AI infrastructure that I realized a critical oversight: while the market spends immense time comparing model intelligence and product fluidity, few discuss the underlying architecture that makes it all possible.\nAt Davos 2026, Jensen Huang described AI as a five-layer cake: Energy, Chips, Infrastructure, Models, and Applications. He emphasized: \u0026ldquo;Every successful application pulls on every layer beneath it, all the way down to the power plant that keeps it alive.\u0026rdquo;\nWhile most eyes are fixed on the top three layers, I am increasingly fascinated by the foundation. Over the past year, I have conducted deep-dive research into the bottom three layers of Jensen’s cake—Siting \u0026amp; Deployment, Supply Chain \u0026amp; Geopolitics, and Energy \u0026amp; Decarbonization. Viewed together, they point to a singular conclusion: AI’s true moat is quietly forming at the base.\n做產品的人，習慣將目光聚焦於用戶體驗與產出。直到我開始深入研究 AI 基礎設施才意識到：市場花了大量時間比較模型的聰明程度與產品的流暢度，卻鮮少有人討論讓這一切得以運作的底層架構。\n黃仁勳2026年在達沃斯論壇上提出，AI 是一個五層蛋糕：能源、晶片、基礎設施、模型、應用。他強調：「每一個成功的應用，都牽動著它底下的每一層，一路往下，直到那座讓它活著的發電廠。」\n多數人的目光停在第三層以上，但我對更基礎的地基更有興趣。過去一年，我針對 Jensen 蛋糕最底部的三層——選址與部署、供應鏈與地緣、能源與去碳——進行了深入研究。將這三者結合來看，指向了一個明確的結論：AI 真正的護城河，正悄悄在底層形成。\nI. Data Center Siting: Business Logic Behind Technical Decisions #Last year at GovHack 2025, I utilized public data from AEMO, BOM, and ABS to build a five-factor evaluation framework for Australian data center siting: Power Supply Density, Cooling Efficiency, Network Connectivity, Talent Pool, and Land Acquisition[1].\nThe most valuable insight from the model was the disconnect between the \u0026ldquo;ideal\u0026rdquo; scores and market reality. For instance, while Sydney is not the optimal solution for grid cleanliness, it carries approximately 60% of Australia’s total data center capacity. This outcome is driven by a different set of logic: latency, customer proximity, and the historical legacy of the fibre backbone. This demonstrates that DC siting has no perfect solution; it is a dynamic tug-of-war between grid conditions, commercial viability, and ESG mandates. Only those who master these trade-offs can execute the right configurations at the right nodes.\n一、數據中心選址：技術決策背後的商業權衡\n去年參加 GovHack 2025，我利用 AEMO、BOM、ABS 的公開數據，建構了澳洲資料中心選址的五因子評估框架（電力供應密度、冷卻效率、網路連接性、人才庫、土地取得)[1]。\n跑完模型後，最有價值的發現是分數與市場現況的落差。例如，Sydney 在電力潔淨度上並非最佳解，卻承載了全澳洲約 60% 的資料中心容量。決定這個結果的是另一組邏輯：延遲時間、客戶重心、以及光纖骨幹（fibre backbone）的歷史沉澱。這顯示 DC 的選址從來沒有完美解，而是電網條件、商業可行性與 ESG 之間的動態拉鋸。只有最懂這些 trade-off 的人，才能在正確的節點做正確的配置。\nII. Upstream Infrastructure: Supply Chain and Geopolitical Risks #Building a data center relies heavily on critical metals: copper (cooling and wiring, ~25% of hardware material costs), aluminum (racks and casings, 15%), silicon (chips and solar panels, 15%), and various rare earth elements.\nEarlier this year, I tracked the role of Australian mineral resources within the AI infrastructure supply chain[2]. Australia is the world’s second-largest holder of copper reserves and a global leader in lithium and rare earths, exporting over $410 billion AUD in resources annually—$152 billion of which goes to China. As AI compute demand grows exponentially, the demand for minerals follows. However, copper supply elasticity is extremely low (new mines take 10–15 years to develop), and the geopolitical concentration of rare earths dictates which nations maintain supply chain sovereignty in the AI arms race. Australia’s position between the U.S. and China makes it both an opportunity and a strategic pawn. Understanding these \u0026ldquo;underwater\u0026rdquo; factors is essential to grasping the long-term procurement logic and massive localized bets made by hyperscalers.\n二、基礎設施上游：供應鏈與地緣風險\n建置一座資料中心，高度依賴關鍵金屬：銅（冷卻與配線，佔硬體材料成本約 25%）、鋁（機架與外殼，15%）、矽（晶片與太陽能板，15%），以及多種稀土元素。\n今年早些時候，我做了一個研究：追蹤澳洲礦產資源在 AI 基礎設施供應鏈裡的位置[2]。澳洲是全球第二大銅儲量國，鋰和稀土同樣是世界級供應者，每年資源出口逾 $4,100 億澳幣，其中對中國就有 $1,520 億。當 AI 算力需求以指數成長，礦產的需求曲線跟著走。銅的供應彈性低、建新礦需要 10–15 年。稀土的地緣集中度，決定了哪些國家在 AI 軍備競賽裡有供應鏈自主權。澳洲夾在美中之間的位置，讓它在 AI 供應鏈裡同時是機會和棋子。要了解到這些水面下的因素，才能真正理解 hyperscaler 的長期採購邏輯，也才能看懂為什麼特定時間點特定業者會在特定地點押下大注。\nIII. Clean Energy: The increasing Challenge for the Next Decade #Electricity represents roughly 60–70% of data center operating expenses. With NSW\u0026rsquo;s 2024 average wholesale electricity price at approximately $130/MWh, a 100MW facility faces annual power costs nearing $100 million AUD. This is not a marginal cost; it is the lifeblood of the business model.\nConcurrently, I analyzed the carbon emission structures of the Queensland aluminum industry[3]. In 2024, Rio Tinto’s QLD aluminum chain emissions reached 7.72 Mt CO₂e, leaving a 3.8 Mt gap to meet 2030 reduction targets. While aluminum and data centers seem unrelated, they face identical structural dilemmas: how to decarbonize while maintaining high-intensity baseload power without spiraling costs. The solutions being explored by both industries are nearly identical—long-term PPAs, renewable energy integration solutions. The decarbonization learning curve is cross-industrial; however, this cross-domain perspective remains scarce in the AI field today.\n三、清潔能源：DC 未來十年越來越重要的挑戰\n電力佔 DC 營運成本高達 60–70%。以 NSW 2024 年批發電價均值約 $130/MWh 計算，一座 100MW 的設施，每年電費接近一億澳幣。這已不是單純的邊際成本，而是商業模型的命脈。\n同期，我分析了昆士蘭鋁業的碳排結構[3]（2024 年 Rio Tinto QLD 鋁業鏈排放量達 7.72 Mt CO₂e，距離 2030 減半目標尚缺 3.8 Mt）。鋁業與 DC 表面上毫無交集，卻面臨完全相同的結構性困境：如何在維持高強度基載（baseload）用電的同時達成去碳，並控制成本。雙方摸索出的解法高度一致——長期 PPA、再生能源整合\u0026hellip;等方案。能源去碳的學習曲線是跨產業共通的，而這種跨域視野在目前的 AI 領域仍極為稀缺。\nConclusion: The Depth of the Foundation Dictates the Height of the Tower #While the industry remains fixated on model parameters and end-user experiences, the true strategic pivot is happening at the base. The logic of data center siting, the resilience of critical mineral supply chains, and the path toward decarbonizing high-intensity power—these three factors define the competitive boundary of next-generation AI infrastructure.\nWithout the support of energy, minerals, and infrastructure, even the most advanced models are merely castles in the air. The moat of AI lies not just in parameter counts, but in a profound understanding of underlying commercial trade-offs, geopolitical landscapes, and energy regulations. The long-term winners in the AI era will be those who can see the technical trends at the top, while mastering the industrial logic at the bottom.\n結論：決定高樓高度的，永遠是地基的深度\n當產業目光仍聚焦於模型參數量與終端體驗時，真正的戰略轉折點已在底層悄然發生。資料中心的選址邏輯、關鍵礦產的供應鏈韌性、高強度用電的去碳路徑——這三者共同構成了新一代 AI 基礎設施的競爭邊界。\n沒有堅實的能源、礦產與基礎設施支撐，再先進的模型與產品都只是空中樓閣。AI 的護城河不只在參數的高低，更在於對底層商業權衡、地緣格局與能源規則的深度理解。未來能在 AI 時代長期立足的玩家，必然是既能看懂上層技術趨勢，又能吃透底層產業邏輯的人。\nFurther Reading / 延伸閱讀：\nGovHack 2025: A Data-Driven Framework for Data Center Siting AI Revolution: Australia\u0026rsquo;s Pivotal Role in the Global Mineral Supply Chain Queensland Aluminium Industry Emissions: 2024 Status \u0026amp; Decarbonization Roadmap © Chung-Hao Lee. All Rights Reserved. All content on this webpage—including but not limited to text, images, design, code, and multimedia materials—is protected under the international copyright treaties. Unauthorized reproduction, modification, distribution, public transmission, or commercial use is strictly prohibited. Legal action will be taken against infringement. © 李崇豪。保留所有權利。 本網頁之內容（包括但不限於文字、圖片、設計、程式碼及多媒體素材）均受國際著作權條約保護。未經書面授權，嚴禁任何形式之複製、改作、散布、公開傳輸或商業利用。侵權者將依法追訴。\n","date":"April 7, 2026","permalink":"https://chunghaolee.com/posts/beyond-the-model-the-invisible-moat-of-ai-infrastructure--data-center/","section":"Posts","summary":"\u003cp\u003eProduct professionals naturally focus on user experience and output. It wasn’t until I began diving into AI infrastructure that I realized a critical oversight: while the market spends immense time comparing model intelligence and product fluidity, few discuss the underlying architecture that makes it all possible.\u003c/p\u003e\n\u003cp\u003eAt Davos 2026, Jensen Huang described \u003ca href=\"https://blogs.nvidia.com/blog/ai-5-layer-cake/\" target=\"_blank\" rel=\"noreferrer\"\u003eAI as a five-layer cake\u003c/a\u003e: Energy, Chips, Infrastructure, Models, and Applications. He emphasized: \u0026ldquo;Every successful application pulls on every layer beneath it, all the way down to the power plant that keeps it alive.\u0026rdquo;\u003c/p\u003e","title":"Beyond the Model: The Invisible Moat of AI Infrastructure \u0026 Data Center 🏭"},{"content":"","date":"April 6, 2026","permalink":"https://chunghaolee.com/posts/external-articleai-is-a-5-layer-cake/","section":"Posts","summary":"","title":"[External Article] - AI is a 5-Layer Cake"},{"content":"","date":null,"permalink":"https://chunghaolee.com/tags/product-decoder/","section":"Tags","summary":"","title":"Product Decoder"},{"content":"As a PM, I always like to view products from a user\u0026rsquo;s perspective. This time, I\u0026rsquo;d like to analyze a feature I find brilliantly designed — Douyin\u0026rsquo;s \u0026ldquo;AI Fact-Check\u0026rdquo; (@抖音求真).\n身為產品經理，我習慣從使用者視角出發。這次我想聊聊一個設計得非常精妙的功能：抖音的「AI 求真」(@抖音求真)。\nWhat Is It? #When browsing Douyin, if a user encounters content that seems questionable, they can simply type @抖音求真 in the comments. An AI agent then verifies the claim, compiles sources, and replies with a structured summary—labeling the content as confirmed, disputed, or debunked.\n這是什麼？ #當使用者在滑抖音看到可疑內容時，只要在評論區標記 @抖音求真，後台的 AI 就會介入核實、彙整來源，並直接回覆一份結構化的查核摘要，清楚標示該內容是「屬實」、「存在爭議」或「純屬謠言」。\nWhy This Design Is Brilliant #(1) Solving a Real Problem at the Right Moment #Traditional fact-checking requires users to leave the app and cross-reference sources manually. That’s too much friction. Douyin embeds \u0026lsquo;AI fact-checking\u0026rsquo; directly into the content flow. The moment of doubt is the moment of action.\n(2) A Minimalist Interaction: Just @ It #The interaction leverages an existing behavior: the @ mention. There’s no new UI to learn and no forms to fill out. It’s a textbook example of reducing cognitive load by repurposing a familiar interaction for a new use case.\n(3) Asynchronous by Design: No Waiting Required #AI verification takes time. Rather than making users wait on a loading screen, the interaction is fully asynchronous. Users can keep scrolling, and the system notifies them once the result is posted. This respects the user\u0026rsquo;s \u0026ldquo;infinite scroll\u0026rdquo; habit instead of interrupting it.\n為什麼這個設計很好？ #(1) 在正確的時機解決真正的痛點 #傳統的查證流程需要使用者跳出 App、打開瀏覽器搜尋並交叉比對，這對大多數人來說門檻太高了。「抖音求真」將查核機制直接嵌入消費流程中。在使用者產生懷疑的當下，就能立刻完成操作，不讓資訊落差過夜。\n(2) 極簡互動：一個 @ 就搞定 #它的互動設計極其簡單——就是大家本來就熟悉的 @ 標記。不需要學習新介面，也不用填寫複雜表單。這是一個減少「認知負荷」的典範：利用既有的使用者習慣，來承載全新的功能需求。\n(3) 非同步設計：無需停留等待 #AI 查證需要時間搜尋與彙整。抖音並沒有讓使用者停在加載畫面，而是採用了完全非同步的設計。標記完後，使用者可以繼續滑下一個影片；等查核完成，系統會自動跳出通知。這種設計順應了短影音「不斷下滑」的行為邏輯，而非中斷它。\nThe PM Takeaway #Great product design isn\u0026rsquo;t just about the most powerful technology; it’s about making that technology feel effortless. Douyin’s AI Fact-Check succeeds by being in the right context, using the right interaction, and respecting the user\u0026rsquo;s flow.\nPM 觀點總結 #出色的產品設計不只是追求技術強大，更重要的是讓技術變得**「毫不費力」**。「AI 求真」的成功在於：鎖定正確痛點、出現在正確時機、使用最輕量的互動，並用非同步機制尊重使用者的時間。這就是好的產品思維。\n© Chung-Hao Lee. All Rights Reserved. All content on this webpage—including but not limited to text, images, design, code, and multimedia materials—is protected under the international copyright treaties. Unauthorized reproduction, modification, distribution, public transmission, or commercial use is strictly prohibited. Legal action will be taken against infringement. © 李崇豪。保留所有權利。 本網頁之內容（包括但不限於文字、圖片、設計、程式碼及多媒體素材）均受國際著作權條約保護。未經書面授權，嚴禁任何形式之複製、改作、散布、公開傳輸或商業利用。侵權者將依法追訴。\n","date":"April 4, 2026","permalink":"https://chunghaolee.com/posts/product-decoder_douyin-truth-seeker/","section":"Posts","summary":"\u003cp\u003eAs a PM, I always like to view products from a user\u0026rsquo;s perspective. This time, I\u0026rsquo;d like to analyze a feature I find brilliantly designed — Douyin\u0026rsquo;s \u0026ldquo;AI Fact-Check\u0026rdquo; (@抖音求真).\u003c/p\u003e\n\u003cp\u003e身為產品經理，我習慣從使用者視角出發。這次我想聊聊一個設計得非常精妙的功能：抖音的「AI 求真」(@抖音求真)。\u003c/p\u003e\n\u003cimg src=\"pic1.png\" alt=\"logo of douyin\" width=\"80\" style=\"display: block; margin: 20px auto; border-radius: 8px; box-shadow: 0 2px 8px rgba(0,0,0,0.1);\"\u003e\n\u003chr\u003e\n\u003ch2 id=\"what-is-it\" class=\"relative group\"\u003eWhat Is It? \u003cspan class=\"absolute top-0 w-6 transition-opacity opacity-0 -start-6 not-prose group-hover:opacity-100\"\u003e\u003ca class=\"group-hover:text-primary-300 dark:group-hover:text-neutral-700\" style=\"text-decoration-line: none !important;\" href=\"#what-is-it\" aria-label=\"Anchor\"\u003e#\u003c/a\u003e\u003c/span\u003e\u003c/h2\u003e\u003cp\u003eWhen browsing Douyin, if a user encounters content that seems questionable, they can simply type \u003cstrong\u003e@抖音求真\u003c/strong\u003e in the comments. An AI agent then verifies the claim, compiles sources, and replies with a structured summary—labeling the content as confirmed, disputed, or debunked.\u003c/p\u003e","title":"Product Decoder: Douyin's AI Fact-Check — How a Simple @ Mention Fights Misinformation"},{"content":"","date":null,"permalink":"https://chunghaolee.com/tags/tiktok/","section":"Tags","summary":"","title":"TikTok"},{"content":"Welcome to visit the site to find out MORE ! Fiery Golden Eyes 🏘️\nExcited to introduce my latest side project: Fiery Golden Eyes #Fiery Golden Eyes is an web-based app that has interactive map of property sale prices across Greater Sydney, powered by official NSW Government open data. Navigating the Sydney property market requires high-fidelity data and spatial clarity. This platform provides a comprehensive, interactive visualization of sales across Greater Sydney, powered by official NSW Government open data. 很高興向大家介紹我的最新 Side Project：Fiery Golden Eyes #Fiery Golden Eyes 是一款網頁應用程式，透過互動式地圖呈現大雪梨地區的房產成交價格，數據來源為新南威爾斯州（NSW）政府的官方開放數據。 在雪梨房地產市場中穿梭，需要極高精確度的數據與清晰的空間直覺。本平台透過官方開放數據，提供大雪梨地區成交狀況的全面互動式視覺化呈現。 It delivers: #→ Suburb-level price heatmaps for instantaneous market evaluation → Granular detail panels featuring median, average, and historical price extremes → Longitudinal trend analysis through integrated scatter plots and property-type modeling 它提供以下核心價值：\n#→ 分區價格熱點圖： 協助使用者即時評估市場行情 → 精細的資訊面板： 展示中位數、平均價以及歷史成交價格的極值 → 長期趨勢分析： 透過整合散點圖與房產類型建模，掌握價格走勢 What else you get: #→ Advanced filtering by zoning, price range, and specific property categories → Seamless weekly automation via GitHub Actions to ensure the latest data is always accessible → Geographic search functionality for precise suburb and address lookup 其他進階功能： #→ 進階篩選： 支援土地分區、價格範圍與特定房產類別的精確篩選 → 無縫自動化更新： 透過 GitHub Actions 實現每週自動更新，確保數據始終處於最新狀態 → 地理搜尋功能： 支援精確的行政區與地址查詢 THE ARCHITECTURE: LEVERAGING AI AGENTS FOR FULL-STACK DEVELOPMENT #This project was built from the ground up using Claude Code as a primary development partner. By treating AI as a high-level engineering agent, I was able to bridge the gap between complex legacy government datasets and a modern web experience. → Engineering a Robust Data Pipeline: The NSW Valuer General publishes data in a specialized .DAT format with multi-layered record types. I utilized AI to architect a Python-based pipeline that autonomously parses, and validates these files, converting them into high-performance structures optimized for the frontend. → Rapid Full-Stack Prototyping: The user interface—a React application integrated with Leaflet for spatial mapping and Recharts for data visualization—was developed through iterative, natural language collaboration. This allowed for rapid UI refinements and the deployment of a highly responsive, professional dark-themed design. → Fully Automated Lifecycle: To ensure the platform remains a reliable source of truth, I leveraged AI to configure a sophisticated CI/CD workflow via GitHub Actions. Every Tuesday morning, the system autonomously fetches new public records, updates the underlying data models, and redeploys the live site. 架構設計：運用 AI Agent 實現全端開發 #本專案從概念開發到最終部署，皆是將 Claude Code 作為主要開發夥伴協作完成。透過將 AI 視為高階工程代理人（Engineering Agent），我得以成功跨越政府舊式複雜數據集與現代網頁體驗之間的技術鴻溝。 → 打造強健的數據管線: NSW Valuer General 以特殊的 .DAT 格式發布數據，其結構包含多層次的紀錄類型。我利用 AI 架構了一套 Python 數據管線，能自動解析並驗證這些原始文件，將其轉換為針對前端效能優化的高效結構。 → 快速全端原型開發: 使用者介面採用 React 開發，並整合 Leaflet 進行空間地圖呈現與 Recharts 進行數據視覺化。透過與 AI 的自然語言協作，我們實現了快速的 UI 迭代，並部署了具備專業感且高度響應式的深色主題設計。 → 全自動化生命週期管理: 為確保平台成為可靠的資訊來源，我利用 AI 設定了基於 GitHub Actions 的精精密 CI/CD 工作流。每週二早晨，系統會自動抓取最新的官方開放紀錄、更新底層數據模型並重新部署網站。 As a practitioner focused on delivering actionable insights, I believe this project demonstrates how AI-driven execution can transform raw open data into high-value professional tools. 作為一名專注於提供實質見解的實踐者，我相信這個專案充分展示了 AI 驅動的執行力如何將原始的開放數據，轉化為具備高價值的專業工具。 © Chung-Hao Lee. All Rights Reserved. All content on this webpage—including but not limited to text, images, design, code, and multimedia materials—is protected under the international copyright treaties. Unauthorized reproduction, modification, distribution, public transmission, or commercial use is strictly prohibited. Legal action will be taken against infringement. © 李崇豪。保留所有權利。 本網頁之內容（包括但不限於文字、圖片、設計、程式碼及多媒體素材）均受國際著作權條約保護。未經書面授權，嚴禁任何形式之複製、改作、散布、公開傳輸或商業利用。侵權者將依法追訴。\n","date":"March 22, 2026","permalink":"https://chunghaolee.com/portfolio/fiery_golden_eyes/","section":"Portfolios","summary":"\u003cp\u003eWelcome to visit the site to find out MORE ! \u003ca href=\"https://lch99310.github.io/Fiery-Golden-Eyes/\" target=\"_blank\" rel=\"noreferrer\"\u003eFiery Golden Eyes\u003c/a\u003e 🏘️\u003c/p\u003e\n\u003chr\u003e\n\u003ch2 id=\"excited-to-introduce-my-latest-side-project-fiery-golden-eyes\" class=\"relative group\"\u003eExcited to introduce my latest side project: Fiery Golden Eyes \u003cbr\u003e \u003cspan class=\"absolute top-0 w-6 transition-opacity opacity-0 -start-6 not-prose group-hover:opacity-100\"\u003e\u003ca class=\"group-hover:text-primary-300 dark:group-hover:text-neutral-700\" style=\"text-decoration-line: none !important;\" href=\"#excited-to-introduce-my-latest-side-project-fiery-golden-eyes\" aria-label=\"Anchor\"\u003e#\u003c/a\u003e\u003c/span\u003e\u003c/h2\u003e\u003cp\u003eFiery Golden Eyes is an web-based app that has interactive map of property sale prices across Greater Sydney, powered by official NSW Government open data. \u003cbr\u003e\nNavigating the Sydney property market requires high-fidelity data and spatial clarity. This platform provides a comprehensive, interactive visualization of sales across Greater Sydney, powered by official NSW Government open data. \u003cbr\u003e\u003c/p\u003e","title":"【AI Side Project Vol.05】Fiery Golden Eyes 🏘️"},{"content":"","date":null,"permalink":"https://chunghaolee.com/portfolio/","section":"Portfolios","summary":"","title":"Portfolios"},{"content":"Welcome to visit the site to find out MORE ! SciCover Summary 🔬\nA web-based application that provide Bilingual Summaries of cover articles from scientific journals # 🚀 Every week, premier academic journals like Science, Nature, and Cell publish cover stories that reshape our understanding of the world. However, a significant chasm remains between the rigors of academic publishing and general public accessibility. #每週，《Science》、《Nature》、《Cell》等頂級學術期刊都會發表改變世界的封面研究。然而，在學術出版與公眾認知之間，仍存在著一道難以跨越的鴻溝。\n🚩 Four Structural Pain Points in Knowledge Acquisition # 🔹 Time Constraints \u0026amp; Information Overload In today’s fast-paced professional environment, researchers and industry experts find it increasingly difficult to dedicate hours each week to navigating multiple journal websites just to identify high-value cover stories. 🔹 Cognitive Load \u0026amp; Technical Jargon Academic papers are designed for peer review, not broad communication. High-density terminology and complex syntax create a steep barrier for non-specialists trying to grasp core innovations. 🔹 Linguistic Barriers \u0026amp; Lack of Localization Academic English follows specific conventions that can be daunting for non-native speakers. It’s not just about translation; it\u0026rsquo;s about the lack of high-quality, real-time Chinese interpretations that preserve scientific context. 🔹 Information Fragmentation \u0026amp; Retrieval Inefficiency Top-tier journals update on different schedules across decentralized platforms. The lack of a unified interface makes the knowledge-seeking process fragmented and inefficient. 🚩 讀者的四大認知痛點 (Pain Points) # 🔹 時間成本與資訊過載 現代職場節奏緊湊，研究者與從業人員難以每週耗費大量時間遍歷各個期刊官網，篩選高價值的封面故事。 🔹 認知負荷與術語壁壘 學術論文旨在同行評審，而非大眾傳播。高密度的專有術語與複雜的句式，使得非領域專家難以快速捕捉核心創新點。 🔹 語言障礙與在地化不足 英文學術寫作具有特定的規範與語境，對於非母語讀者而言，不僅是語言翻譯的問題，更是文化與邏輯理解的挑戰。目前市場缺乏即時且高品質的中文科學解讀。 🔹 資訊碎片化與獲取效率 頂級期刊更新週期不一，來源分散。缺乏統一的聚合介面，導致知識獲取過程低效且斷續。 🧠 Development Philosophy: AI-Driven \u0026ldquo;Semantic Translation Engineering\u0026rdquo; #To address these friction points, I utilized the GitHub Models API to build an automated content workflow focused on:\nFrom \u0026ldquo;Translation\u0026rdquo; to \u0026ldquo;Interpretation\u0026rdquo;: Leveraging the cross-domain capabilities of Large Language Models (LLMs) to reconstruct technical abstracts into bilingual \u0026ldquo;popular science\u0026rdquo; content—reducing entry barriers while maintaining scientific integrity. High-Efficiency Technical Architecture: Adopting a Serverless philosophy, the entire site is deployed on GitHub Pages, achieving zero-cost, high-performance content distribution. Professional UI/UX Aesthetics: Drawing inspiration from the visual language of Quanta Magazine, the platform features a responsive, magazine-style layout with Dark Mode support for a premium reading experience. 🧠 開發思路：AI 驅動的「語意轉譯工程」 #為了優化上述痛點，我利用 GitHub Models API 構建了一套全自動化的內容轉譯工作流，重點在於：\n從「翻譯」進化為「解讀」： 利用大語言模型（LLM）的跨域理解能力，將硬核論文摘要重構為「科普級」雙語內容，在保持科學準確性的前提下，極大降低了閱讀門檻。 高效能的技術架構： 採用 Serverless 理念，全站部署於 GitHub Pages，實現零成本、高效率的內容分發。 UI/UX 的專業審美： 借鑒 Quanta Magazine 的視覺美學，設計了雜誌化的排版佈局，支援多端適配與深色模式，提升專業閱讀的舒適度。 💡 The Solution: SciCover Summary #In response to these challenges, I developed SciCover Summary—a web-based bilingual science platform. This is more than a reading tool; it is an attempt to reshape the science communication pipeline using AI.\n💡 解決方案：SciCover Summary #針對上述問題，我開發了 SciCover Summary—— 一個基於 Web 的雙語科學摘要平台。這不僅是一個閱讀工具，更是利用 AI 技術重塑科學傳播流程的嘗試。\n🔬 SciCover Summary: Key Features # Automated Journal Tracking: Real-time monitoring of Science, Nature, and Cell to eliminate manual searching. High-Quality Bilingual Summaries: AI-generated \u0026ldquo;interpretative rewrites\u0026rdquo; that balance natural Chinese fluency with professional English insights. Visual Science Gallery: Displays original cover art and in-text figures with bilingual captions to enhance visual understanding. Historical Archives: A curated scientific timeline allowing users to browse research trends across months and years. 🔬 SciCover Summary 核心功能摘要 # 智能期刊聚合： 自動化追蹤頂刊動態，告別手動檢索，實現「一站式」科研閱覽。 高品質雙語摘要： AI 生成的 interpretative rewrites，兼顧中文流暢度與專業英文語感。 視覺化科學畫廊： 完整呈現期刊封面圖與圖表說明，強化視覺理解。 歷史脈絡存檔： 建立橫跨年份的科學趨勢地圖。 🌐 Our Vision #Scientific progress should not be confined to the lab. My goal with SciCover is to lower the barrier to entry, enabling busy professionals and the public to touch the frontiers of human knowledge more efficiently.\n🌐 願景 #科技的進步不應僅限於實驗室內。我希望透過 SciCover，降低科學知識的獲取門檻，讓忙碌的專業人士與公眾能更高效地觸摸科學前沿。\n🛡️ Commitment to Transparency \u0026amp; Sourcing: The platform strictly aggregates publicly available metadata and provides direct official links to the original publishers for every article. This ensures full transparency, respects intellectual property, and directs traffic back to the primary academic sources.\n🛡️ 透明度與學術尊重： 本平台僅呈現公開發布的資料與摘要資訊，並為所有文章提供指向官方期刊的原始連結。這不僅確保了資訊的透明度，更體現了對原創學術出版物與智慧財產權的尊重。\n© Chung-Hao Lee. All Rights Reserved. All content on this webpage—including but not limited to text, images, design, code, and multimedia materials—is protected under the international copyright treaties. Unauthorized reproduction, modification, distribution, public transmission, or commercial use is strictly prohibited. Legal action will be taken against infringement. © 李崇豪。保留所有權利。 本網頁之內容（包括但不限於文字、圖片、設計、程式碼及多媒體素材）均受國際著作權條約保護。未經書面授權，嚴禁任何形式之複製、改作、散布、公開傳輸或商業利用。侵權者將依法追訴。\n","date":"March 1, 2026","permalink":"https://chunghaolee.com/portfolio/scicover_summary/","section":"Portfolios","summary":"\u003cp\u003eWelcome to visit the site to find out MORE ! \u003ca href=\"https://lch99310.github.io/SciCover_Summary/\" target=\"_blank\" rel=\"noreferrer\"\u003eSciCover Summary\u003c/a\u003e 🔬\u003c/p\u003e\n\u003chr\u003e\n\u003ch2 id=\"a-web-based-application-that-provide-bilingual-summaries-of-cover-articles-from-scientific-journals\" class=\"relative group\"\u003eA web-based application that provide Bilingual Summaries of cover articles from scientific journals \u003cspan class=\"absolute top-0 w-6 transition-opacity opacity-0 -start-6 not-prose group-hover:opacity-100\"\u003e\u003ca class=\"group-hover:text-primary-300 dark:group-hover:text-neutral-700\" style=\"text-decoration-line: none !important;\" href=\"#a-web-based-application-that-provide-bilingual-summaries-of-cover-articles-from-scientific-journals\" aria-label=\"Anchor\"\u003e#\u003c/a\u003e\u003c/span\u003e\u003c/h2\u003e\u003cp\u003e\n\n\n\n\n\n\u003cfigure\u003e\n    \n    \n\n\n\n\n\n\n\n\n  \n    \u003cpicture\n      class=\"mx-auto my-0 rounded-md\"\n      \n    \u003e\n      \n      \n      \n      \n        \u003csource\n          \n            srcset=\"https://chunghaolee.com/portfolio/scicover_summary/homepage_hu_d0250713ebc9316d.webp 330w,https://chunghaolee.com/portfolio/scicover_summary/homepage_hu_4f3fd2363d73340a.webp 660w\n            \n              ,https://chunghaolee.com/portfolio/scicover_summary/homepage_hu_fe37cace3be79d93.webp 1024w\n            \n            \n              ,https://chunghaolee.com/portfolio/scicover_summary/homepage_hu_a1a646488422faee.webp 1320w\n            \"\n          \n          sizes=\"100vw\"\n          type=\"image/webp\"\n        /\u003e\n      \n      \u003cimg\n        width=\"1890\"\n        height=\"1306\"\n        class=\"mx-auto my-0 rounded-md\"\n        alt=\"Homepage\"\n        loading=\"lazy\" decoding=\"async\"\n        \n          src=\"https://chunghaolee.com/portfolio/scicover_summary/homepage_hu_fb347b95417182d1.png\" srcset=\"https://chunghaolee.com/portfolio/scicover_summary/homepage_hu_c164ff6f98c8ece4.png 330w,https://chunghaolee.com/portfolio/scicover_summary/homepage_hu_fb347b95417182d1.png 660w\n          \n            ,https://chunghaolee.com/portfolio/scicover_summary/homepage_hu_bbef4af1a0bb8062.png 1024w\n          \n          \n            ,https://chunghaolee.com/portfolio/scicover_summary/homepage_hu_67faeaa0ece00ba7.png 1320w\n          \"\n          sizes=\"100vw\"\n        \n      /\u003e\n    \u003c/picture\u003e\n  \n\n\n\u003c/figure\u003e\n\u003c/p\u003e\n\u003ch3 id=\"-every-week-premier-academic-journals-like-science-nature-and-cell-publish-cover-stories-that-reshape-our-understanding-of-the-world-however-a-significant-chasm-remains-between-the-rigors-of-academic-publishing-and-general-public-accessibility\" class=\"relative group\"\u003e🚀 Every week, premier academic journals like \u003cem\u003eScience\u003c/em\u003e, \u003cem\u003eNature\u003c/em\u003e, and \u003cem\u003eCell\u003c/em\u003e publish cover stories that reshape our understanding of the world. However, a significant chasm remains between the rigors of academic publishing and general public accessibility. \u003cspan class=\"absolute top-0 w-6 transition-opacity opacity-0 -start-6 not-prose group-hover:opacity-100\"\u003e\u003ca class=\"group-hover:text-primary-300 dark:group-hover:text-neutral-700\" style=\"text-decoration-line: none !important;\" href=\"#-every-week-premier-academic-journals-like-science-nature-and-cell-publish-cover-stories-that-reshape-our-understanding-of-the-world-however-a-significant-chasm-remains-between-the-rigors-of-academic-publishing-and-general-public-accessibility\" aria-label=\"Anchor\"\u003e#\u003c/a\u003e\u003c/span\u003e\u003c/h3\u003e\u003cp\u003e每週，《Science》、《Nature》、《Cell》等頂級學術期刊都會發表改變世界的封面研究。然而，在學術出版與公眾認知之間，仍存在著一道難以跨越的鴻溝。\u003c/p\u003e","title":"【AI Side Project Vol.04】Develop web app of \"SciCover Summary\" 🔬"},{"content":"I am deeply grateful to my team for their support and dedicated efforts in product development over the past quarter. It is also a great honor to be recognized by TikTok with the 2025Q4 TikTok Spot Bonus Award.\n© Chung-Hao Lee. All Rights Reserved. All content on this webpage—including but not limited to text, images, design, code, and multimedia materials—is protected under the international copyright treaties. Unauthorized reproduction, modification, distribution, public transmission, or commercial use is strictly prohibited. Legal action will be taken against infringement. © 李崇豪。保留所有權利。 本網頁之內容（包括但不限於文字、圖片、設計、程式碼及多媒體素材）均受國際著作權條約保護。未經書面授權，嚴禁任何形式之複製、改作、散布、公開傳輸或商業利用。侵權者將依法追訴。\n","date":"February 25, 2026","permalink":"https://chunghaolee.com/posts/2025q4-tiktok-spot-bonus-award/","section":"Posts","summary":"\u003cp\u003eI am deeply grateful to my team for their support and dedicated efforts in product development over the past quarter. It is also a great honor to be recognized by TikTok with the 2025Q4 TikTok Spot Bonus Award.\u003c/p\u003e\n\u003cp\u003e\n\n\n\n\n\n\u003cfigure\u003e\n    \n    \n\n\n\n\n\n\n\n\n  \n    \u003cpicture\n      class=\"mx-auto my-0 rounded-md\"\n      \n    \u003e\n      \n      \n      \n      \n        \u003csource\n          \n            srcset=\"https://chunghaolee.com/posts/2025q4-tiktok-spot-bonus-award/25Q4_SpotBonus_hu_2d730d4813f27746.webp 330w,https://chunghaolee.com/posts/2025q4-tiktok-spot-bonus-award/25Q4_SpotBonus_hu_edbd4344e3ec8b85.webp 660w\n            \n              \n                ,https://chunghaolee.com/posts/2025q4-tiktok-spot-bonus-award/25Q4_SpotBonus_hu_304325155d535bab.webp 708w\n              \n            \n            \n              \n                ,https://chunghaolee.com/posts/2025q4-tiktok-spot-bonus-award/25Q4_SpotBonus_hu_304325155d535bab.webp 708w\n              \n            \"\n          \n          sizes=\"100vw\"\n          type=\"image/webp\"\n        /\u003e\n      \n      \u003cimg\n        width=\"708\"\n        height=\"1254\"\n        class=\"mx-auto my-0 rounded-md\"\n        alt=\"Award\"\n        loading=\"lazy\" decoding=\"async\"\n        \n          src=\"https://chunghaolee.com/posts/2025q4-tiktok-spot-bonus-award/25Q4_SpotBonus_hu_9f1ba8e57990f249.png\" srcset=\"https://chunghaolee.com/posts/2025q4-tiktok-spot-bonus-award/25Q4_SpotBonus_hu_9a0eb639ee408cae.png 330w,https://chunghaolee.com/posts/2025q4-tiktok-spot-bonus-award/25Q4_SpotBonus_hu_9f1ba8e57990f249.png 660w\n          \n            ,https://chunghaolee.com/posts/2025q4-tiktok-spot-bonus-award/25Q4_SpotBonus.png 708w\n          \n          \n            ,https://chunghaolee.com/posts/2025q4-tiktok-spot-bonus-award/25Q4_SpotBonus.png 708w\n          \"\n          sizes=\"100vw\"\n        \n      /\u003e\n    \u003c/picture\u003e\n  \n\n\n\u003c/figure\u003e\n\u003c/p\u003e\n\u003chr\u003e\n\u003cp\u003e\u003cem\u003e© Chung-Hao Lee. All Rights Reserved.\nAll content on this webpage—including but not limited to text, images, design, code, and multimedia materials—is protected under the international copyright treaties. Unauthorized reproduction, modification, distribution, public transmission, or commercial use is strictly prohibited. Legal action will be taken against infringement.\u003c/em\u003e \u003cbr\u003e\n\u003cem\u003e© 李崇豪。保留所有權利。\n本網頁之內容（包括但不限於文字、圖片、設計、程式碼及多媒體素材）均受國際著作權條約保護。未經書面授權，嚴禁任何形式之複製、改作、散布、公開傳輸或商業利用。侵權者將依法追訴。\u003c/em\u003e\u003c/p\u003e","title":"2025Q4 TikTok Spot Bonus Award"},{"content":"","date":null,"permalink":"https://chunghaolee.com/tags/achievements/","section":"Tags","summary":"","title":"Achievements"},{"content":"","date":"February 22, 2026","permalink":"https://chunghaolee.com/posts/2026-02-22_external-article---%E5%AD%97%E8%8A%82%E9%98%BF%E9%87%8C%E8%85%BE%E8%AE%AF-ai-%E5%A4%A7%E6%88%98%E5%85%A8%E8%AE%B0%E5%BD%95%E4%B8%80%E5%9C%BA%E5%BD%B1%E5%93%8D%E5%91%BD%E8%BF%90%E7%9A%84%E6%88%98%E4%BA%89/","section":"Posts","summary":"","title":"[External Article] - 字节、阿里、腾讯 AI 大战全记录：一场影响命运的战争"},{"content":"Welcome to visit the site to find out MORE ! Daily Underestomated MLBPlayers Ranking ⚾\nNov 2024: I had the vision and the domain expertise, but I faced a technical ceiling. At that time, I could use AI to generate \u0026ldquo;polished UI mockups,\u0026rdquo; but building a live, data-integrated application as a solo creator required a level of manual engineering that was out of reach.\nFeb 2026: The landscape has fundamentally changed. By evolving my workflow to manage autonomous AI agents, I have transitioned from \u0026ldquo;visualizing ideas\u0026rdquo; to shipping a production-ready, full-stack web application.\n⚾️ The Project: MLB Underestimated Player Analyzer #As an avid baseball fan for over a decade (Go Dodgers!). I’ve spent years diving deep into advanced analytics to fuel my Fantasy Baseball obsession and satisfy my inner \u0026ldquo;Armchair GM.\u0026rdquo; To take my analysis to the next level, I am looking to develop a streamlined, interactive platform that allows for efficient, one-stop access to real-time data.\nThe core philosophy behind this project is rooted in the fact that in real-world baseball, a player’s box score results are heavily influenced by external factors—such as defensive positioning, ballpark dimensions, and environmental variance. Consequently, surface-level statistics like wOBA often fail to accurately reflect a player’s underlying skill set or true performance level in real-time. To understand how a player is actually performing, a deeper analysis of “Value vs. Outcome” is required.\nwOBA (Weighted On-Base Average): This represents a player\u0026rsquo;s actual offensive contribution on the field. It is the \u0026ldquo;realized\u0026rdquo; result—the Outcome. xwOBA (Expected wOBA): Using Statcast data, this calculates what a player\u0026rsquo;s output should have been based on the physical characteristics of their hits (Exit Velocity and Launch Angle). This is the Value, a pure measure of contact quality independent of defensive shifts or luck. The Differential ($wOBA - xwOBA$): When realized results significantly lag behind contact quality, a player is Underestimated. Statistical history shows that these gaps are usually temporary; eventually, a player\u0026rsquo;s on-field results will regress to the mean, aligning with their underlying performance profile. My dashboard identifies these statistical anomalies in real-time, highlighting hitters who are poised for a significant rebound.\nConcept outcome developed in 2024 Nov Web app developed in 2026 Feb 🛠 The Orchestration: Architect + AI Agent #The most significant shift isn\u0026rsquo;t just the speed of AI—it’s the Managerial Workflow. I didn\u0026rsquo;t just \u0026ldquo;ask\u0026rdquo; for an app; I orchestrated it through Documentation-Driven Development:\nThe Blueprint: I authored comprehensive technical documentation defining the Design System, User Interaction Logic, and Backend Data Architecture. The Management: I acted as the Project Manager and Lead Architect, guiding AI agents to execute the Full-Stack build (React/Vite Frontend + Python Data Pipeline). The Automation: Together, we implemented an automated workflow that handles daily data ingestion and processing, ensuring the dashboard reflects the most current MLB metrics without any manual intervention. 📈 Case Study: Performance Convergence #Using Tyler O\u0026rsquo;Neill as an example: his early-season batting average appeared low, but the analyzer flagged an elite 16.5% Barrel rate and a massive negative differential. The data suggested his \u0026ldquo;underperformance\u0026rdquo; was a result of statistical variance rather than a decline in skill. As predicted, his outcomes eventually converged with his elite contact quality.\nThe Takeaway #In 2026, the competitive advantage for developers is no longer just \u0026ldquo;writing code\u0026rdquo;—it’s structuring logic and managing agentic execution. One person with a clear vision and rigorous documentation can now deliver the output of an entire engineering team.\n#AI #GenerativeAI #AIAgents #DataAnalytics #FullStack #MLB #SportsTech #ProductManagement #DataVisualization #FutureOfWork\n© Chung-Hao Lee. All Rights Reserved. All content on this webpage—including but not limited to text, images, design, code, and multimedia materials—is protected under the international copyright treaties. Unauthorized reproduction, modification, distribution, public transmission, or commercial use is strictly prohibited. Legal action will be taken against infringement. © 李崇豪。保留所有權利。 本網頁之內容（包括但不限於文字、圖片、設計、程式碼及多媒體素材）均受國際著作權條約保護。未經書面授權，嚴禁任何形式之複製、改作、散布、公開傳輸或商業利用。侵權者將依法追訴。\n","date":"February 21, 2026","permalink":"https://chunghaolee.com/portfolio/daily_underestimated_mlbplayers_ranking/","section":"Portfolios","summary":"\u003cp\u003eWelcome to visit the site to find out MORE ! \u003ca href=\"https://lch99310.github.io/Daily_underestimated_MLBplayer_ranking/\" target=\"_blank\" rel=\"noreferrer\"\u003eDaily Underestomated MLBPlayers Ranking\u003c/a\u003e ⚾\u003c/p\u003e\n\u003chr\u003e\n\u003cp\u003e\u003cstrong\u003eNov 2024:\u003c/strong\u003e I had the vision and the domain expertise, but I faced a technical ceiling. At that time, I could use AI to generate \u0026ldquo;polished UI mockups,\u0026rdquo; but building a live, data-integrated application as a solo creator required a level of manual engineering that was out of reach.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFeb 2026:\u003c/strong\u003e The landscape has fundamentally changed. By evolving my workflow to manage autonomous AI agents, I have transitioned from \u0026ldquo;visualizing ideas\u0026rdquo; to \u003cstrong\u003eshipping a production-ready, full-stack web application.\u003c/strong\u003e\u003c/p\u003e","title":"【AI Side Project Vol.02】Develop web app of \"Daily Underestimated MLB Player Ranking\" ⚾"},{"content":"【From Defining Pain Points to Product Launch: How I Leveraged AI for Automated Development in an Unfamiliar Domain】 #【從定義痛點到產品落地：我如何利用 AI 在陌生領域完成自動化開發】 # In this age of information overload, we often stumble upon high-quality independent blogs or technical columns that lack a \u0026ldquo;Subscribe\u0026rdquo; feature. To avoid missing deep-dive content, I used to check these sites manually—a process that was both inefficient and prone to oversight.\n在資訊爆炸的時代，我們常會遇到高品質的獨立部落格或技術專欄，卻發現它們缺乏「訂閱」功能。為了不漏掉深度內容，過去我只能手動檢查網頁，既低效又容易遺漏。\n\u0026ldquo;Since there is no subscribe button, I’ll build one for myself.\u0026quot;\n「既然沒有訂閱按鈕，我就為自己做一個。」\nThis was the inspiration behind Newsletter Manager—an automated tool powered by GitHub Actions that transforms any website into scheduled email updates.\n這是我開發 Newsletter Manager 的初衷。這是一個基於 GitHub Actions 的自動化工具，能將任何網站轉化為定時 Email 推送。\nFor me, the core value of this project lies not in the code itself, but in how to leverage AI as a force multiplier to solve real-world problems by crossing technical boundaries:\n這份專案對我而言，最核心的價值不在於程式碼本身，而是**「如何透過 AI 槓桿，跨越技術邊界解決實際問題」**：\n🚀 1. Agile Iteration and Problem Definition Early in development, I wasn\u0026rsquo;t fully versed in email protocols or backend automation logic. I treated AI as my \u0026ldquo;Senior Architect.\u0026rdquo; By providing clear requirement definitions (Domain Knowledge), I moved from the core scraping logic of V1 to the automated scheduling of V2, continuously providing feedback and refining the solution through high-frequency iterations with AI.\n🚀 1. 敏捷迭代與問題定義 在開發初期，我對 Email 協議與後端自動化邏輯並不全然熟悉。我將 AI 視為「資深架構師」，透過明確的需求定義（Domain Knowledge），從 V1 的核心抓取邏輯，到 V2 的自動化排程，我不斷針對痛點提出反饋與修正，與 AI 進行高頻的快速迭代。\n🛠️ 2. Lean Architecture Mindset To achieve \u0026ldquo;zero-cost maintenance,\u0026rdquo; I adopted a serverless mindset. By utilizing GitHub Actions as the execution environment combined with the Resend API and JSON data structures, I created an automated closed-loop system that requires no external servers or databases.\n🛠️ 2. 輕量化的架構思維 為了實現「零成本維護」，我選擇了 Serverless 的思維，利用 GitHub Actions 作為執行環境，結合 Resend API 與 JSON 資料結構，實現了無需伺服器、無需資料庫的自動化閉環。\n💡 3. Persistence and a Proactive \u0026ldquo;Get-it-Done\u0026rdquo; Attitude I firmly believe that technical boundaries should never hinder action. In the modern workplace, the most critical skill isn\u0026rsquo;t \u0026ldquo;what you already know,\u0026rdquo; but the speed at which you \u0026ldquo;learn and solve problems\u0026rdquo; when facing the unknown. This side project proves my ability to define requirements from scratch and proactively find the right tools (like AI) to turn an idea into a functional product.\n💡 3. 解決問題的執著與實幹態度 我深信：「技術邊界不應成為行動的阻礙。」 現代職場中，最關鍵的能力不再是「已知多少」，而是面對未知時，「學習並解決問題」的速度。這份 Side Project 證明了我能從零開始定義需求，並主動尋找工具（如 AI）將想法轉化為可執行的產品。\nProject Achievements: ✅ Automated monitoring of website updates ✅ Integrated scheduled tasks via GitHub Actions ✅ Seamless integration of frontend and backend management interfaces (GitHub Pages)\n專案成果： ✅ 自動化監控網站更新 ✅ 整合 GitHub Actions 定時執行任務 ✅ 前後端管理介面整合（GitHub Pages）\nAs a result-oriented practitioner, I am passionate about breaking down complex problems and bringing them to life through innovative tools. If you\u0026rsquo;re interested in AI-collaborative development or workflow automation, let’s connect and exchange ideas!\n身為一個「結果導向」的實踐者，我熱衷於將複雜問題拆解並透過創新工具落地。如果你也對 AI 協作開發或生產力自動化感興趣，歡迎留言交流！\n#AICollaboration #GitHubActions #ProductLanding #SideProject #ProblemSolving #ActionOriented #AutomationWorkflow #LearningByDoing\n© Chung-Hao Lee. All Rights Reserved. All content on this webpage—including but not limited to text, images, design, code, and multimedia materials—is protected under the international copyright treaties. Unauthorized reproduction, modification, distribution, public transmission, or commercial use is strictly prohibited. Legal action will be taken against infringement. © 李崇豪。保留所有權利。 本網頁之內容（包括但不限於文字、圖片、設計、程式碼及多媒體素材）均受國際著作權條約保護。未經書面授權，嚴禁任何形式之複製、改作、散布、公開傳輸或商業利用。侵權者將依法追訴。\n","date":"February 21, 2026","permalink":"https://chunghaolee.com/portfolio/newsletter_manager/","section":"Portfolios","summary":"\u003ch3 id=\"from-defining-pain-points-to-product-launch-how-i-leveraged-ai-for-automated-development-in-an-unfamiliar-domain\" class=\"relative group\"\u003e【From Defining Pain Points to Product Launch: How I Leveraged AI for Automated Development in an Unfamiliar Domain】 \u003cspan class=\"absolute top-0 w-6 transition-opacity opacity-0 -start-6 not-prose group-hover:opacity-100\"\u003e\u003ca class=\"group-hover:text-primary-300 dark:group-hover:text-neutral-700\" style=\"text-decoration-line: none !important;\" href=\"#from-defining-pain-points-to-product-launch-how-i-leveraged-ai-for-automated-development-in-an-unfamiliar-domain\" aria-label=\"Anchor\"\u003e#\u003c/a\u003e\u003c/span\u003e\u003c/h3\u003e\u003ch3 id=\"從定義痛點到產品落地我如何利用-ai-在陌生領域完成自動化開發\" class=\"relative group\"\u003e【從定義痛點到產品落地：我如何利用 AI 在陌生領域完成自動化開發】 \u003cspan class=\"absolute top-0 w-6 transition-opacity opacity-0 -start-6 not-prose group-hover:opacity-100\"\u003e\u003ca class=\"group-hover:text-primary-300 dark:group-hover:text-neutral-700\" style=\"text-decoration-line: none !important;\" href=\"#%e5%be%9e%e5%ae%9a%e7%be%a9%e7%97%9b%e9%bb%9e%e5%88%b0%e7%94%a2%e5%93%81%e8%90%bd%e5%9c%b0%e6%88%91%e5%a6%82%e4%bd%95%e5%88%a9%e7%94%a8-ai-%e5%9c%a8%e9%99%8c%e7%94%9f%e9%a0%98%e5%9f%9f%e5%ae%8c%e6%88%90%e8%87%aa%e5%8b%95%e5%8c%96%e9%96%8b%e7%99%bc\" aria-label=\"Anchor\"\u003e#\u003c/a\u003e\u003c/span\u003e\u003c/h3\u003e\u003cp\u003e\n\n\n\n\n\n\u003cfigure\u003e\n    \n    \n\n\n\n\n\n\n\n\n  \n    \u003cpicture\n      class=\"mx-auto my-0 rounded-md\"\n      \n    \u003e\n      \n      \n      \n      \n        \u003csource\n          \n            srcset=\"https://chunghaolee.com/portfolio/newsletter_manager/pic1_hu_828e23b67e002980.webp 330w,https://chunghaolee.com/portfolio/newsletter_manager/pic1_hu_7d11dce9a78cf952.webp 660w\n            \n              ,https://chunghaolee.com/portfolio/newsletter_manager/pic1_hu_b770b4e55cfc382.webp 1024w\n            \n            \n              ,https://chunghaolee.com/portfolio/newsletter_manager/pic1_hu_86638e19e864a6c5.webp 1320w\n            \"\n          \n          sizes=\"100vw\"\n          type=\"image/webp\"\n        /\u003e\n      \n      \u003cimg\n        width=\"1670\"\n        height=\"1286\"\n        class=\"mx-auto my-0 rounded-md\"\n        alt=\"Newsletter\"\n        loading=\"lazy\" decoding=\"async\"\n        \n          src=\"https://chunghaolee.com/portfolio/newsletter_manager/pic1_hu_4b5f0fc0695b3fc6.png\" srcset=\"https://chunghaolee.com/portfolio/newsletter_manager/pic1_hu_359d004f13195c0e.png 330w,https://chunghaolee.com/portfolio/newsletter_manager/pic1_hu_4b5f0fc0695b3fc6.png 660w\n          \n            ,https://chunghaolee.com/portfolio/newsletter_manager/pic1_hu_bccca4f47b1ac8a6.png 1024w\n          \n          \n            ,https://chunghaolee.com/portfolio/newsletter_manager/pic1_hu_f3e8e1986b1b9882.png 1320w\n          \"\n          sizes=\"100vw\"\n        \n      /\u003e\n    \u003c/picture\u003e\n  \n\n\n\u003c/figure\u003e\n\u003c/p\u003e\n\u003cp\u003eIn this age of information overload, we often stumble upon high-quality independent blogs or technical columns that lack a \u0026ldquo;Subscribe\u0026rdquo; feature. To avoid missing deep-dive content, I used to check these sites manually—a process that was both inefficient and prone to oversight.\u003cbr\u003e\n在資訊爆炸的時代，我們常會遇到高品質的獨立部落格或技術專欄，卻發現它們缺乏「訂閱」功能。為了不漏掉深度內容，過去我只能手動檢查網頁，既低效又容易遺漏。\u003c/p\u003e","title":"【AI Side Project Vol.03】Develop web app of \"Newsletter Manager\" 📩"},{"content":"Beyond the Model: 基礎設施與流程設計才是 AI 迭代的隱形引擎 # In today’s era of rapid AI evolution, many fixate on model architecture and algorithmic innovation. However, the true determinants of iteration speed and competitiveness often lie hidden within the details of Infrastructure (Infra) and Workflow Design. Why? Because every model evolution relies on high-quality data support, and the collection, processing, and analysis of that data depend heavily on real-time human judgment and intervention. If the infrastructure is flawed or the workflow clunky, the team wastes valuable time \u0026ldquo;hunting for data,\u0026rdquo; \u0026ldquo;patching data holes,\u0026rdquo; or \u0026ldquo;clarifying data issues,\u0026rdquo; making rapid iteration impossible.\n在 AI 技術飛速迭代的今天，許多人將目光聚焦在模型架構與演算法創新上，但真正決定迭代速度與競爭力的關鍵，往往藏在 基礎設施（Infra） 與 流程設計 的細節裡。為什麼這麼說？因為模型的每一次進化，都離不開高品質數據的支撐，而數據的採集、處理與分析，又高度依賴人類的即時判斷與介入。如果基礎設施不夠完善，流程設計不夠流暢，團隊就得花大量時間在「找數據」、「補數據」、「釐清數據問題」上，迭代自然快不起來。\n💡 Real-World Case: Laptop Anomaly Detection Model\nTake, for example, a project involving an AI model for detecting laptop anomalies. When the model underperformed during a validation run, engineers needed to identify the cause immediately. Here, the quality of the infrastructure creates a massive efficiency gap:\nHigh Efficiency Scenario: With robust infrastructure, engineers can view all relevant data via a one-stop dashboard, potentially resolving the issue in minutes. Low Efficiency Scenario: If data is scattered across different systems, requiring manual extraction from databases, run SQL, log files, and third-party platforms, the preparation alone could take half a day—even before cross-analysis begins. 💡 實戰案例：筆電異常檢測模型\n舉個例子，之前開發一個用於檢測筆電硬體異常的 AI 模型。當模型在某次驗證中表現不如預期時，工程師需要快速找出原因。這時，Infra 的好壞決定了效率的天壤之別：\n高效場景： 如果基礎設施夠好，工程師能在同一個儀表板上 一站式查看所有相關數據，問題可能幾分鐘內就水落石出。 低效場景： 如果數據散落在不同系統，需要手動從資料庫、跑SQL、log 檔、第三方平台逐一撈取，光是前置準備就可能耗掉半天，更別提後續的交叉分析。 ⚙️ Workflow Design: The Devil is in the Details\nThe impact of workflow design is even more profound. When data flows through the pipeline, a failure in any link drags down downstream analysis efficiency: Is it a sensor malfunction causing missing material data? Is unstable network transmission causing packet loss? Or is it a poorly designed error code classification that prevents anomalies from being correctly labeled? These seemingly minor flaws accumulate to become massive resistance to iteration. Conversely, if we smooth out these pitfalls during the design phase—ensuring reliable data sources, stable transmission, and clear labeling—subsequent RCA (Root Cause Analysis) and data collection become seamless, drastically boosting iteration speed.\n⚙️ 流程設計：魔鬼藏在細節裡\n流程設計的影響更為深遠。數據在 pipeline 中流轉時，任何一個環節出問題，都會拖累後續的分析效率：是感測器故障導致物料數據遺漏？是網路傳輸不穩定造成數據丟包？還是 Error Code 分類設計不良，讓異常行為無法被正確標註？這些看似微小的瑕疵，累積起來就會變成迭代的巨大阻力。反過來說，如果我們能在流程設計階段就把這些坑填平，確保數據來源可靠、傳輸穩定、標籤清晰。後續的 RCA（根本原因分析）與數據收集就能行雲流水，迭代速度自然大幅提升。\n🚀 Conclusion\nThere are no shortcuts to optimizing infrastructure and workflows; it relies on polishing one detail after another. But it is precisely these details that determine whether a team can identify problems faster, acquire new data faster, and form a positive \u0026ldquo;Analyze → Optimize → Verify → Iterate\u0026rdquo; loop. In the competitive AI landscape, the key to standing out lies in a team that builds a solid foundation and keeps the iteration engine running non-stop.\n🚀 結論\n基礎設施與流程的優化，沒有捷徑，靠的就是一個個細節的打磨。但正是這些細節，決定了團隊能否更快發現問題、更快獲取新數據、更快形成 「分析 → 優化 → 驗證 → 迭代」 的正向循環。在 AI 時代競爭中，能夠脫穎而出的關鍵，就在於能把基礎打紮實、讓迭代停不下來的團隊。\n#AI #MachineLearning #MLOps #Infrastructure #DataPipeline #WorkflowOptimization #EngineeringExcellence © Chung-Hao Lee. All Rights Reserved. All content on this webpage—including but not limited to text, images, design, code, and multimedia materials—is protected under the international copyright treaties. Unauthorized reproduction, modification, distribution, public transmission, or commercial use is strictly prohibited. Legal action will be taken against infringement. © 李崇豪。保留所有權利。 本網頁之內容（包括但不限於文字、圖片、設計、程式碼及多媒體素材）均受國際著作權條約保護。未經書面授權，嚴禁任何形式之複製、改作、散布、公開傳輸或商業利用。侵權者將依法追訴。\n","date":"February 17, 2026","permalink":"https://chunghaolee.com/posts/beyond-the-model-infrastructure-and-workflow-design-are-the-hidden-engines/","section":"Posts","summary":"\u003ch2 id=\"beyond-the-model-基礎設施與流程設計才是-ai-迭代的隱形引擎\" class=\"relative group\"\u003eBeyond the Model: 基礎設施與流程設計才是 AI 迭代的隱形引擎 \u003cspan class=\"absolute top-0 w-6 transition-opacity opacity-0 -start-6 not-prose group-hover:opacity-100\"\u003e\u003ca class=\"group-hover:text-primary-300 dark:group-hover:text-neutral-700\" style=\"text-decoration-line: none !important;\" href=\"#beyond-the-model-%e5%9f%ba%e7%a4%8e%e8%a8%ad%e6%96%bd%e8%88%87%e6%b5%81%e7%a8%8b%e8%a8%ad%e8%a8%88%e6%89%8d%e6%98%af-ai-%e8%bf%ad%e4%bb%a3%e7%9a%84%e9%9a%b1%e5%bd%a2%e5%bc%95%e6%93%8e\" aria-label=\"Anchor\"\u003e#\u003c/a\u003e\u003c/span\u003e\u003c/h2\u003e\u003cp\u003e\n\n\n\n\n\n\u003cfigure\u003e\n    \n    \n\n\n\n\n\n\n\n\n  \n    \u003cpicture\n      class=\"mx-auto my-0 rounded-md\"\n      \n    \u003e\n      \n      \n      \n      \n        \u003csource\n          \n            srcset=\"https://chunghaolee.com/posts/beyond-the-model-infrastructure-and-workflow-design-are-the-hidden-engines/pic_hu_1da09544fe3bdabf.webp 330w,https://chunghaolee.com/posts/beyond-the-model-infrastructure-and-workflow-design-are-the-hidden-engines/pic_hu_a967e2f4f38cb655.webp 660w\n            \n              ,https://chunghaolee.com/posts/beyond-the-model-infrastructure-and-workflow-design-are-the-hidden-engines/pic_hu_172a97ec776b6d37.webp 1024w\n            \n            \n              ,https://chunghaolee.com/posts/beyond-the-model-infrastructure-and-workflow-design-are-the-hidden-engines/pic_hu_18356523d8511404.webp 1320w\n            \"\n          \n          sizes=\"100vw\"\n          type=\"image/webp\"\n        /\u003e\n      \n      \u003cimg\n        width=\"2730\"\n        height=\"1535\"\n        class=\"mx-auto my-0 rounded-md\"\n        \n        loading=\"lazy\" decoding=\"async\"\n        \n          src=\"https://chunghaolee.com/posts/beyond-the-model-infrastructure-and-workflow-design-are-the-hidden-engines/pic_hu_df02ebd1af44460a.png\" srcset=\"https://chunghaolee.com/posts/beyond-the-model-infrastructure-and-workflow-design-are-the-hidden-engines/pic_hu_9119a6c2cc01fc66.png 330w,https://chunghaolee.com/posts/beyond-the-model-infrastructure-and-workflow-design-are-the-hidden-engines/pic_hu_df02ebd1af44460a.png 660w\n          \n            ,https://chunghaolee.com/posts/beyond-the-model-infrastructure-and-workflow-design-are-the-hidden-engines/pic_hu_4f1401107c082068.png 1024w\n          \n          \n            ,https://chunghaolee.com/posts/beyond-the-model-infrastructure-and-workflow-design-are-the-hidden-engines/pic_hu_f7930907b0b18a5d.png 1320w\n          \"\n          sizes=\"100vw\"\n        \n      /\u003e\n    \u003c/picture\u003e\n  \n\n\n\u003c/figure\u003e\n\u003c/p\u003e\n\u003chr\u003e\n\u003cp\u003eIn today’s era of rapid AI evolution, many fixate on model architecture and algorithmic innovation. However, the true determinants of iteration speed and competitiveness often lie hidden within the details of \u003cstrong\u003eInfrastructure (Infra)\u003c/strong\u003e and \u003cstrong\u003eWorkflow Design\u003c/strong\u003e. Why? Because every model evolution relies on high-quality data support, and the collection, processing, and analysis of that data depend heavily on real-time human judgment and intervention. If the infrastructure is flawed or the workflow clunky, the team wastes valuable time \u0026ldquo;hunting for data,\u0026rdquo; \u0026ldquo;patching data holes,\u0026rdquo; or \u0026ldquo;clarifying data issues,\u0026rdquo; making rapid iteration impossible.\u003c/p\u003e","title":"Beyond the Model: Infrastructure and Workflow Design are the Hidden Engines of AI Iteration"},{"content":"","date":null,"permalink":"https://chunghaolee.com/tags/process/","section":"Tags","summary":"","title":"Process"},{"content":"【AI 實戰手札 Vol.01】回歸數據本質：透過深度錯誤分析 (Error Analysis) 突破 Prompt Engineering 瓶頸 # In the development of high-performance AI systems, it is easy to rely on technical intuition while overlooking fundamental engineering truths. Recently, while working on a Video Content Moderation task, I navigated a deep dialectic between \u0026ldquo;theoretical optimization\u0026rdquo; and \u0026ldquo;data-centric reality.\u0026rdquo; This experience redefined my understanding of \u0026ldquo;pragmatism\u0026rdquo; in AI engineering.\n在開發高效能 AI 系統的過程中，我們往往容易依賴技術直覺，卻忽略了最基本的工程真理。 近期在處理一個影片內容合規性偵測 (Content Moderation) 的任務時，我經歷了從「理論優化」回歸「數據本質」的深度辯證，這讓我對 AI 工程中的「務實思維」有了更深刻的定義。\n🚩 The Challenge: Stuck in a Metric Plateau The objective was to leverage Prompt Engineering to accurately identify policy violations in video keyframes. Initially, I followed the industry-standard Few-shot Learning strategy, integrating multiple sets of carefully curated examples. However, the results were counterintuitive: for two full days, I was stuck in a loop of \u0026ldquo;Instruction Optimization ➔ Evaluation ➔ Performance Stagnation.\u0026rdquo; Despite refining the semantic logic, the model’s $Precision$ remained stubbornly low.\n🚩 挑戰：陷入指標停滯的循環 任務目標是透過 提示詞工程 (Prompt Engineering)，讓模型精準辨識影片關鍵幀中的違規行為。最初，我採用了業界標準的 Few-shot Learning 策略，在提示詞中整合多組精心挑選的正負範例。然而，現實卻不如預期：整整兩天，我陷入了「優化指令 ➔ 測試評估 ➔ 效能停滯」的無限迴圈。儘管不斷微調語意邏輯，模型的 精確率 ($Precision$) 始終無法突破瓶頸。\n🔍 The Turning Point: Shifting from Blind Iteration to Systematic Case Studies When empirical rules failed, I pivoted from trial-and-error to a more rigorous approach: a comprehensive qualitative Error Analysis. By manually inspecting model outputs case by case, I captured the underlying root cause: Context Leakage.\nI discovered the model was misinterpreting the \u0026ldquo;reference examples\u0026rdquo; as the \u0026ldquo;actual target content.\u0026rdquo; Even with clear instruction boundaries, the model experienced cognitive confusion when parsing complex contexts. This explained why simply increasing the number of examples led to an imbalanced Signal-to-Noise Ratio (SNR), ultimately hindering performance.\n🔍 轉機：放棄盲目迭代，啟動系統化 Case Study 當經驗法則走到盡頭，我決定停止無效的指令改版，轉而採取最紮實的工程方法：針對錯誤案例 (Error Cases) 進行詳盡的定性分析。 就在逐一檢視模型輸出的過程中，我捕捉到了隱藏在底層的技術根因：Context Leakage（上下文洩漏）。\n我發現模型將提示詞中的「參考範例」直接誤認為是「當前待判斷的目標內容」。即便指令結構再嚴謹，模型在處理複雜 Context 時仍產生了認知混淆。這解釋了為何單純增加範例數量，反而導致了訊噪比（Signal-to-Noise Ratio）失衡，適得其反。\n📝 Pragmatic Insights for AI Engineers:\nError Analysis is the Soul of AI Development: Instead of \u0026ldquo;guessing\u0026rdquo; the model\u0026rsquo;s internal logic, return to the data. The patience to diagnose data at the front line is what determines the production-grade quality of an AI product. Guarding the Boundaries of Instruction Following: Few-shot is not a silver bullet. Understanding how models weigh context and utilizing structural isolation (e.g., XML Tags) is vital for optimizing complex reasoning tasks. Pragmatism as the Shortest Path: When performance stagnates, don\u0026rsquo;t just rush to tweak parameters. Step into the model\u0026rsquo;s perspective and see what it truly \u0026ldquo;perceives.\u0026rdquo; 📝 務實者的技術洞察：\nError Analysis 是 AI 開發的靈魂：與其依賴技術直覺去「猜測」模型想法，不如回歸數據。在追求自動化的時代，願意深入第一線診斷每筆數據的耐心，是決定 AI 產品落地品質的關鍵。 警惕指令遵循（Instruction Following）的邊界：Few-shot 並非萬靈丹。理解模型如何解析 Context 權重，並透過結構化隔離（如 XML Tags）優化輸入邏輯，對於處理複雜任務至關重要。 務實，才是解決問題最快的路徑：當模型表現不如預期時，與其急著調整參數，不如先站在模型的視角，觀察它到底「看見」了什麼。 🚀 Closing Thoughts Technical depth is often hidden within seemingly tedious detail checks. In the journey of AI tuning, maintaining a reverence for data and an obsession with detail are an engineer\u0026rsquo;s strongest competitive advantages.\n🚀 結語 技術的深度，往往藏在那些看似枯燥的細節檢視中。在 AI 調優的路上，保持對數據的敬畏與對細節的執著，是工程師最強大的競爭力。\n#AI實戰手札 #PromptEngineering #ErrorAnalysis #DataCentricAI #AIEngineer #MachineLearning #技術洞察 #職涯成長 © Chung-Hao Lee. All Rights Reserved. All content on this webpage—including but not limited to text, images, design, code, and multimedia materials—is protected under the international copyright treaties. Unauthorized reproduction, modification, distribution, public transmission, or commercial use is strictly prohibited. Legal action will be taken against infringement. © 李崇豪。保留所有權利。 本網頁之內容（包括但不限於文字、圖片、設計、程式碼及多媒體素材）均受國際著作權條約保護。未經書面授權，嚴禁任何形式之複製、改作、散布、公開傳輸或商業利用。侵權者將依法追訴。\n","date":"February 14, 2026","permalink":"https://chunghaolee.com/posts/ai-practice-journal-vol.01/","section":"Posts","summary":"\u003ch2 id=\"ai-實戰手札-vol01回歸數據本質透過深度錯誤分析-error-analysis-突破-prompt-engineering-瓶頸\" class=\"relative group\"\u003e【AI 實戰手札 Vol.01】回歸數據本質：透過深度錯誤分析 (Error Analysis) 突破 Prompt Engineering 瓶頸 \u003cspan class=\"absolute top-0 w-6 transition-opacity opacity-0 -start-6 not-prose group-hover:opacity-100\"\u003e\u003ca class=\"group-hover:text-primary-300 dark:group-hover:text-neutral-700\" style=\"text-decoration-line: none !important;\" href=\"#ai-%e5%af%a6%e6%88%b0%e6%89%8b%e6%9c%ad-vol01%e5%9b%9e%e6%ad%b8%e6%95%b8%e6%93%9a%e6%9c%ac%e8%b3%aa%e9%80%8f%e9%81%8e%e6%b7%b1%e5%ba%a6%e9%8c%af%e8%aa%a4%e5%88%86%e6%9e%90-error-analysis-%e7%aa%81%e7%a0%b4-prompt-engineering-%e7%93%b6%e9%a0%b8\" aria-label=\"Anchor\"\u003e#\u003c/a\u003e\u003c/span\u003e\u003c/h2\u003e\u003cp\u003e\n\n\n\n\n\n\u003cfigure\u003e\n    \n    \n\n\n\n\n\n\n\n\n  \n    \u003cpicture\n      class=\"mx-auto my-0 rounded-md\"\n      \n    \u003e\n      \n      \n      \n      \n        \u003csource\n          \n            srcset=\"https://chunghaolee.com/posts/ai-practice-journal-vol.01/pic_hu_22d6338662ef8fbf.webp 330w,https://chunghaolee.com/posts/ai-practice-journal-vol.01/pic_hu_c34006ab85ebb3f5.webp 660w\n            \n              ,https://chunghaolee.com/posts/ai-practice-journal-vol.01/pic_hu_456f1e0a1f313bfd.webp 1024w\n            \n            \n              ,https://chunghaolee.com/posts/ai-practice-journal-vol.01/pic_hu_efcdcf3f3f6bedf7.webp 1320w\n            \"\n          \n          sizes=\"100vw\"\n          type=\"image/webp\"\n        /\u003e\n      \n      \u003cimg\n        width=\"2336\"\n        height=\"1760\"\n        class=\"mx-auto my-0 rounded-md\"\n        \n        loading=\"lazy\" decoding=\"async\"\n        \n          src=\"https://chunghaolee.com/posts/ai-practice-journal-vol.01/pic_hu_ec5ac0b48b97c27f.png\" srcset=\"https://chunghaolee.com/posts/ai-practice-journal-vol.01/pic_hu_6f327c31a9d85d53.png 330w,https://chunghaolee.com/posts/ai-practice-journal-vol.01/pic_hu_ec5ac0b48b97c27f.png 660w\n          \n            ,https://chunghaolee.com/posts/ai-practice-journal-vol.01/pic_hu_4210f0e77c030ee.png 1024w\n          \n          \n            ,https://chunghaolee.com/posts/ai-practice-journal-vol.01/pic_hu_8c2dbe3206f2f183.png 1320w\n          \"\n          sizes=\"100vw\"\n        \n      /\u003e\n    \u003c/picture\u003e\n  \n\n\n\u003c/figure\u003e\n\u003c/p\u003e\n\u003chr\u003e\n\u003cp\u003e\u003cstrong\u003eIn the development of high-performance AI systems, it is easy to rely on technical intuition while overlooking fundamental engineering truths.\u003c/strong\u003e Recently, while working on a \u003cstrong\u003eVideo Content Moderation\u003c/strong\u003e task, I navigated a deep dialectic between \u0026ldquo;theoretical optimization\u0026rdquo; and \u0026ldquo;data-centric reality.\u0026rdquo; This experience redefined my understanding of \u0026ldquo;pragmatism\u0026rdquo; in AI engineering.\u003c/p\u003e","title":"【AI Practice Journal Vol.01】Returning to Data Fundamentals: Overcoming Prompt Engineering Bottlenecks through Systematic Error Analysis"},{"content":"","date":null,"permalink":"https://chunghaolee.com/tags/analytics/","section":"Tags","summary":"","title":"Analytics"},{"content":"","date":null,"permalink":"https://chunghaolee.com/tags/self-assessment/","section":"Tags","summary":"","title":"Self-Assessment"},{"content":"貫穿我至今職涯的兩個關鍵字：分析 與 流程 #Over the past seven years of my product management journey—spanning diverse industries and roles—I have found that two capabilities consistently determine whether I can deliver results: analytical rigor and process discipline.\n在過去七年的PM旅程中，雖然跨越不同產業、承擔過多樣化的職務，但我發現有兩項能力始終決定了我能否把事情做成：分析能力與流程管控。\n🔍 Analysis: The Foundation of Solid Decision-Making #Sound business decisions stem from clear judgment, which in turn is grounded in rigorous analysis. Whether through data-driven quantitative methods (data cleansing, metric development) or qualitative assessments based on market dynamics and team context, my approach is to deconstruct information into comparable and measurable components. This enables me to formulate actionable recommendations that empower teams to make swift trade-offs and move forward amid uncertainty.\nFor instance, in a previous role, my team and I observed a 50% sudden drop in the recall rate of our \u0026ldquo;No Power\u0026rdquo; (failure to power on) issue detection model. Rather than attributing this directly to model degradation, we deconstructed the metric\u0026rsquo;s structure—where the numerator represented units predicted by the model as \u0026ldquo;No Power,\u0026rdquo; and the denominator reflected units actually dispatched due to \u0026ldquo;No Power.\u0026rdquo; Through deeper analysis of the gap between user reports and engineering diagnoses, we uncovered the root cause: a significant number of \u0026ldquo;failure to charge\u0026rdquo; cases stemmed from incompatibility between adapter firmware and office systems, preventing the system from recognizing the adapter and thus disabling notebook charging. End users, however, continued to report these as \u0026ldquo;No Power\u0026rdquo; issues. While our model accurately identified adapter-related faults, the mismatch between its classification logic and user perception led to a distorted recall rate.\n🔍 分析：驅動決策的根基 #良好的商業決策來自清晰的判斷，而清晰的判斷源自紮實的分析。不論是數據驅動的量化分析（數據清理、指標建立），或是依照市場與團隊情況做出的定性評估，我的做法都是把資訊拆解成可比較、可衡量的要素，並以此為基礎提出可執行的決策建議，讓團隊能夠在不確定中快速取捨並前進。\n例如在之前的工作時，我跟團隊發現某No Power（無法開機）問題的模型召回率（recall rate）在短期內驟降50%，我們並未直接歸因於模型效能退化。透過拆解指標結構（分子為模型預測出的異常機台數，分母為實際因該問題出貨的機台數），我們進一步分析用戶回報與工程診斷的落差，發現根本原因在於：大量「無法充電」案例源於電源轉接器韌體與辦公系統不相容，導致系統無法辨識轉接器；而用戶端仍將此歸類為「無法開機」。模型準確識別了轉接器問題，卻因分類邏輯與用戶認知不一致，造成召回率失真。\n🔄 Process: Connect and Systematize Complex Tasks #Analysis answers \u0026ldquo;what we should do,\u0026rdquo; while process answers \u0026ldquo;how we will do it.\u0026rdquo; When confronting large-scale or ambiguous projects, I design processes that break them down into manageable and trackable components, thereby constructing reliable pathways to achieve intended outcomes.\nFor example, in a prior work, my team and I built an intelligent operations workflow spanning the entire product lifecycle. This systematically transformed the traditional reactive model of \u0026ldquo;issue occurrence → manual troubleshooting → repair\u0026rdquo; into a proactive closed loop of \u0026ldquo;prediction → diagnosis → knowledge reuse\u0026rdquo;:\nPhase One: Intelligent Diagnosis (Accelerating Response)\nLeveraging user descriptions from repair centers (call logs), we deployed an AI classification model to instantly identify potential failure types (e.g., battery, adapter, panel) with 80% accuracy. This reduced engineers\u0026rsquo; average diagnostic time by 40%, enabling frontline teams to rapidly focus on root causes.\nPhase Two: Predictive Insights (Informing Design Optimization)\nBy integrating device operational logs, we developed a 30-day failure prediction dashboard (Power BI), empowering engineering teams to proactively identify design vulnerabilities. For instance, when adapter anomaly prediction rates rose for a specific laptop batch, the team optimized power management logic in the next-generation design phase—reducing return rates at the source.\nPhase Three: Knowledge Institutionalization (Building Organizational Memory)\nWe established an enterprise-grade issue knowledge base to structurally archive historical cases and their resolutions. Engineers can now leverage semantic search to rapidly retrieve past experiences, eliminating redundant troubleshooting. Critically, the system automatically links predictions from the first two phases to relevant solutions, creating a seamless flow from \u0026ldquo;issue alert → root cause identification → remediation guidance.\u0026rdquo;\nThis three-phase workflow not only integrated discrete tools into a cohesive system but also redefined operational logic—shifting from a cost center (repair) to a value center (prevention and optimization). Ultimately, the process improved the team\u0026rsquo;s efficiency in handling repetitive issues by 50% and provided data-driven insights to inform product design iterations.\n🔄 流程：將複雜任務系統化 #分析回答「我們應該做什麼」，流程則回答「我們要怎麼做」。面對龐大或模糊的項目，透過流程設計將其拆解為可管理、可追蹤的環節，我可以構建出確實的流程來達到預期的目標。\n例如：在過去的某項經歷中，我跟團隊構建了一套貫穿產品生命週期的智能運維流程，將傳統「問題發生→人工排查→修復」的被動模式，系統化升級為「預測→診斷→知識複用」的主動閉環：\n第一階段：智能診斷（縮短響應路徑）\n基於維修中心的用戶描述（call log），部署AI分類模型即時識別潛在故障類型（如battery、adaptor、panel），準確率達80%。此舉將工程師平均診斷時間縮短40%，讓一線團隊能快速聚焦根因。\n第二階段：前瞻預測（驅動設計優化）\n透過整合設備運行日誌，構建30天故障預測儀表板（Power BI），使工程團隊能提前識別設計薄弱點。例如，當某批次筆電的adapter異常預測率上升，團隊可於下一代設計階段即優化電源管理邏輯，從源頭降低返修率。\n第三階段：知識沉澱（形成組織記憶）\n搭建企業級問題知識庫，結構化儲存歷史案例與解決方案。工程師可透過語義搜尋快速調用過往經驗，避免重複排查；更關鍵的是，系統將前兩階段的預測結果自動關聯至相關解決方案，實現「問題預警→根因定位→修復建議」的無縫銜接。 這套三階段流程不僅將單點工具串聯為協同系統，更重構了運營邏輯：從成本中心（維修）轉向價值中心（預防與優化）。最終，該流程幫助團隊將重複性問題處理效率提升50%，並為產品設計迭代提供數據驅動的決策依據。\nI am grateful for the opportunity to hone these two capabilities across diverse scenarios over the past few years. Moving forward, I will continue to apply this methodology—\u0026ldquo;data as the foundation, process as the backbone\u0026rdquo;—to future challenges, enabling teams to build scalable and replicable solutions that consistently deliver stable and predictable value in practice.\n我很感謝過去幾年能在多元場景中鍛鍊這兩項能力。在之後，我也能將這套「以數據為本、以流程為骨」的方法論帶入下一個挑戰，為團隊建立可擴展、可複製的解決方案，並在實務上持續創造穩定且可預期的價值。\n© Chung-Hao Lee. All Rights Reserved. All content on this webpage—including but not limited to text, images, design, code, and multimedia materials—is protected under the international copyright treaties. Unauthorized reproduction, modification, distribution, public transmission, or commercial use is strictly prohibited. Legal action will be taken against infringement. © 李崇豪。保留所有權利。 本網頁之內容（包括但不限於文字、圖片、設計、程式碼及多媒體素材）均受國際著作權條約保護。未經書面授權，嚴禁任何形式之複製、改作、散布、公開傳輸或商業利用。侵權者將依法追訴。\n","date":"February 1, 2026","permalink":"https://chunghaolee.com/posts/the-two-keywords-analysis-and-process/","section":"Posts","summary":"\u003ch2 id=\"貫穿我至今職涯的兩個關鍵字分析-與-流程\" class=\"relative group\"\u003e貫穿我至今職涯的兩個關鍵字：分析 與 流程 \u003cspan class=\"absolute top-0 w-6 transition-opacity opacity-0 -start-6 not-prose group-hover:opacity-100\"\u003e\u003ca class=\"group-hover:text-primary-300 dark:group-hover:text-neutral-700\" style=\"text-decoration-line: none !important;\" href=\"#%e8%b2%ab%e7%a9%bf%e6%88%91%e8%87%b3%e4%bb%8a%e8%81%b7%e6%b6%af%e7%9a%84%e5%85%a9%e5%80%8b%e9%97%9c%e9%8d%b5%e5%ad%97%e5%88%86%e6%9e%90-%e8%88%87-%e6%b5%81%e7%a8%8b\" aria-label=\"Anchor\"\u003e#\u003c/a\u003e\u003c/span\u003e\u003c/h2\u003e\u003cp\u003eOver the past seven years of my product management journey—spanning diverse industries and roles—I have found that two capabilities consistently determine whether I can deliver results: analytical rigor and process discipline.\u003c/p\u003e\n\u003cbr\u003e\n\u003cp\u003e在過去七年的PM旅程中，雖然跨越不同產業、承擔過多樣化的職務，但我發現有兩項能力始終決定了我能否把事情做成：分析能力與流程管控。\u003c/p\u003e\n\u003cp\u003e\n\n\n\n\n\n\u003cfigure\u003e\n    \n    \n\n\n\n\n\n\n\n\n  \n    \u003cpicture\n      class=\"mx-auto my-0 rounded-md\"\n      \n    \u003e\n      \n      \n      \n      \n        \u003csource\n          \n            srcset=\"https://chunghaolee.com/posts/the-two-keywords-analysis-and-process/pic_hu_1dd3654193f4da22.webp 330w,https://chunghaolee.com/posts/the-two-keywords-analysis-and-process/pic_hu_30a3268d5331e189.webp 660w\n            \n              ,https://chunghaolee.com/posts/the-two-keywords-analysis-and-process/pic_hu_a5509ff4560893ed.webp 1024w\n            \n            \n              ,https://chunghaolee.com/posts/the-two-keywords-analysis-and-process/pic_hu_5e06c9805076d495.webp 1320w\n            \"\n          \n          sizes=\"100vw\"\n          type=\"image/webp\"\n        /\u003e\n      \n      \u003cimg\n        width=\"2336\"\n        height=\"1760\"\n        class=\"mx-auto my-0 rounded-md\"\n        \n        loading=\"lazy\" decoding=\"async\"\n        \n          src=\"https://chunghaolee.com/posts/the-two-keywords-analysis-and-process/pic_hu_e1e0f6f2a4c4b2ec.png\" srcset=\"https://chunghaolee.com/posts/the-two-keywords-analysis-and-process/pic_hu_c1906414ee437121.png 330w,https://chunghaolee.com/posts/the-two-keywords-analysis-and-process/pic_hu_e1e0f6f2a4c4b2ec.png 660w\n          \n            ,https://chunghaolee.com/posts/the-two-keywords-analysis-and-process/pic_hu_6740f9b094f797ad.png 1024w\n          \n          \n            ,https://chunghaolee.com/posts/the-two-keywords-analysis-and-process/pic_hu_e89434f63709ff0d.png 1320w\n          \"\n          sizes=\"100vw\"\n        \n      /\u003e\n    \u003c/picture\u003e\n  \n\n\n\u003c/figure\u003e\n\u003c/p\u003e\n\u003chr\u003e\n\u003ch2 id=\"-analysis-the-foundation-of-solid-decision-making\" class=\"relative group\"\u003e🔍 Analysis: The Foundation of Solid Decision-Making \u003cspan class=\"absolute top-0 w-6 transition-opacity opacity-0 -start-6 not-prose group-hover:opacity-100\"\u003e\u003ca class=\"group-hover:text-primary-300 dark:group-hover:text-neutral-700\" style=\"text-decoration-line: none !important;\" href=\"#-analysis-the-foundation-of-solid-decision-making\" aria-label=\"Anchor\"\u003e#\u003c/a\u003e\u003c/span\u003e\u003c/h2\u003e\u003cp\u003eSound business decisions stem from clear judgment, which in turn is grounded in rigorous analysis. Whether through data-driven quantitative methods (data cleansing, metric development) or qualitative assessments based on market dynamics and team context, my approach is to deconstruct information into comparable and measurable components. This enables me to formulate actionable recommendations that empower teams to make swift trade-offs and move forward amid uncertainty.\u003c/p\u003e","title":"The Two Keywords That Define My Career so far: Analysis and Process"},{"content":"在 TikTok 的一年：回顧與心得 # This month marks my one-year milestone at TikTok—a chapter filled with intense learning and unexpected growth. Here are some insights that have defined my journey:\n這個月，我在 TikTok 工作滿一周年，這段經歷充滿緊湊的學習和意想不到的成長。以下是我旅程中收穫的深刻體悟：\n• 𝘚𝘩𝘢𝘳𝘦 𝘰𝘱𝘦𝘯𝘭𝘺 𝘵𝘰 𝘨𝘳𝘰𝘸 𝘤𝘰𝘭𝘭𝘦𝘤𝘵𝘪𝘷𝘦𝘭𝘺:\u2028Inspired by several key influencers, writing LinkedIn posts taught me the power of reflection and community feedback. Documenting my ideas clarifies my own thinking, surfaces new insights from peers, and sparks meaningful discussions that accelerate everyone’s growth. • 公開分享，共同成長：\u2028受幾位重要前輩和意見領袖啟發，我在 LinkedIn 寫文章的過程，讓我深刻感受到反思和社群回饋的力量。記錄自己的想法不僅釐清思路，還能從同事身上獲得新觀點，並激發有意義的討論，加速大家的成長。\n• 𝘈𝘯𝘤𝘩𝘰𝘳 𝘪𝘯 𝘤𝘦𝘳𝘵𝘢𝘪𝘯𝘵𝘺 𝘢𝘮𝘪𝘥 𝘤𝘩𝘢𝘰𝘴:\u2028In a hyper-growth environment like TikTok, ambiguity is the default. I’ve learned that true resilience lies not in avoiding uncertainty, but in creating pockets of certainty within it. The ability to find specific, actionable goals within uncertain situations—and executing them with precision—is the most critical skill I’ve honed this year. • 在混亂中穩住陣腳：\u2028在 TikTok 這樣高速成長的環境，不確定性是常態。我體會到，真正的韌性不在於逃避不確定，而是能在其中創造小範圍的確定感。今年我磨練出最重要的技能，就是在不確定情境下找到具體可行的目標，並精準執行。\n• 𝘋𝘳𝘪𝘷𝘦𝘯 𝘣𝘺 𝘉𝘶𝘪𝘭𝘥𝘦𝘳\u0026rsquo;𝘴 𝘔𝘪𝘯𝘥𝘴𝘦𝘵 (𝘈𝘯𝘢𝘭𝘺𝘴𝘪𝘴 \u0026amp; 𝘖𝘸𝘯𝘦𝘳𝘴𝘩𝘪𝘱):\u2028I am at my best when I am building and driving, fueling my passion for ownership by deconstructing complex challenges through rigorous analysis and mapping out seamless, logical flows. There is a profound sense of achievement in constructing a solution from inception to completion—architecting robust processes that transform high-stakes uncertainty into steady, measurable progress. This is not just the work I love; it is where my analytical edge and commitment to stability deliver the highest strategic value. • 以建設者心態驅動（分析與主人翁精神）：\u2028當我在打造和推動專案時，狀態最為良好；透過嚴謹分析拆解複雜挑戰，規劃流暢的邏輯流程，讓我對專案責任更有熱情。從零開始建構解決方案直到完成，那份成就感格外深刻——設計堅實的流程，將高風險的不確定性轉化為穩健可衡量的進展。這不僅是我鍾愛的工作，更是我發揮分析優勢與追求穩定，創造最大策略價值之處。\n• 𝘌𝘮𝘣𝘳𝘢𝘤𝘦 𝘵𝘳𝘢𝘯𝘴𝘧𝘦𝘳𝘢𝘣𝘭𝘦 𝘴𝘬𝘪𝘭𝘭𝘴 𝘢𝘯𝘥 𝘥𝘪𝘷𝘦𝘳𝘴𝘦 𝘣𝘢𝘤𝘬𝘨𝘳𝘰𝘶𝘯𝘥𝘴:\u2028Coming from manufacturing and supply-chain roles, I discovered many principles transfer directly to content-moderation systems and operations. For example, error-proofing concepts from the production line can be applied to review workflows to reduce human mistakes. Diverse experiences create unexpected leverage — they help you see solutions cross-functionally. • 善用可轉移技能與多元背景：\u2028我過去在製造業和供應鏈累積的經驗，讓我發現許多原則能直接套用到內容審核系統和運營上。例如，生產線的防錯機制，可應用於審核流程來減少人為疏失。多元背景帶來意想不到的優勢——它幫助你跳脫框架，跨部門洞察解決方案。\nIt has been a privilege to learn from such a dynamic environment. I am excited to keep bridging my past experiences with new challenges, focusing on creating stable systems and meaningful impact in the year ahead. 能在如此充滿活力的環境中學習，我深感幸運。我期待持續融合過去經驗與新挑戰，在未來一年專注打造穩定系統，創造真正有意義的影響。\n© Chung-Hao Lee. All Rights Reserved. All content on this webpage—including but not limited to text, images, design, code, and multimedia materials—is protected under the international copyright treaties. Unauthorized reproduction, modification, distribution, public transmission, or commercial use is strictly prohibited. Legal action will be taken against infringement. © 李崇豪。保留所有權利。 本網頁之內容（包括但不限於文字、圖片、設計、程式碼及多媒體素材）均受國際著作權條約保護。未經書面授權，嚴禁任何形式之複製、改作、散布、公開傳輸或商業利用。侵權者將依法追訴。\n","date":"January 24, 2026","permalink":"https://chunghaolee.com/posts/one-year-at-tiktok-reflection-and-what-i-learned/","section":"Posts","summary":"\u003ch2 id=\"在-tiktok-的一年回顧與心得\" class=\"relative group\"\u003e在 TikTok 的一年：回顧與心得 \u003cspan class=\"absolute top-0 w-6 transition-opacity opacity-0 -start-6 not-prose group-hover:opacity-100\"\u003e\u003ca class=\"group-hover:text-primary-300 dark:group-hover:text-neutral-700\" style=\"text-decoration-line: none !important;\" href=\"#%e5%9c%a8-tiktok-%e7%9a%84%e4%b8%80%e5%b9%b4%e5%9b%9e%e9%a1%a7%e8%88%87%e5%bf%83%e5%be%97\" aria-label=\"Anchor\"\u003e#\u003c/a\u003e\u003c/span\u003e\u003c/h2\u003e\u003cimg src=\"pic1.png\" alt=\"logo of TikTok\" width=\"80\" style=\"display: block; margin: 20px auto; border-radius: 8px; box-shadow: 0 2px 8px rgba(0,0,0,0.1);\"\u003e\n\u003cp\u003eThis month marks my one-year milestone at TikTok—a chapter filled with intense learning and unexpected growth. Here are some insights that have defined my journey:\u003c/p\u003e\n\u003cbr\u003e\n\u003cp\u003e這個月，我在 TikTok 工作滿一周年，這段經歷充滿緊湊的學習和意想不到的成長。以下是我旅程中收穫的深刻體悟：\u003c/p\u003e\n\u003chr\u003e\n\u003cp\u003e• 𝘚𝘩𝘢𝘳𝘦 𝘰𝘱𝘦𝘯𝘭𝘺 𝘵𝘰 𝘨𝘳𝘰𝘸 𝘤𝘰𝘭𝘭𝘦𝘤𝘵𝘪𝘷𝘦𝘭𝘺:\u2028\u003cbr\u003e\nInspired by several key influencers, writing LinkedIn posts taught me the power of reflection and community feedback. Documenting my ideas clarifies my own thinking, surfaces new insights from peers, and sparks meaningful discussions that accelerate everyone’s growth. \u003cbr\u003e\u003c/p\u003e","title":"𝗢𝗻𝗲 𝗬𝗲𝗮𝗿 𝗮𝘁 𝗧𝗶𝗸𝗧𝗼𝗸: 𝗥𝗲𝗳𝗹𝗲𝗰𝘁𝗶𝗼𝗻 𝗮𝗻𝗱 𝗪𝗵𝗮𝘁 𝗜 𝗹𝗲𝗮𝗿𝗻𝗲𝗱"},{"content":"“What does it truly feel like to transform a passion into a profession?”\n「將熱情轉化為職業，究竟是什麼樣的感覺？」\nAs someone who was once a power user spending hours daily on the platform, reaching my one-year milestone at TikTok is more than just a career achievement—it is a dream come true.\n作為曾經每天在平台上投入數小時的重度使用者，能在 TikTok 迎來入職一週年的里程碑，對我而言，這不僅僅是一項職涯成就，更是夢想成真的時刻。\nOver the past 365 days, I have transitioned from a viewer to a contributor, gaining a deeper understanding of the mechanics behind the content and the vibrant pulse of our global community. I am incredibly grateful for the platform the company has provided and for the opportunity to work alongside such talented colleagues.\n在過去的 365 天裡，我從單純的閱聽觀眾轉變為內容的推動者，不僅深入了解了內容背後的運作機制，更切實感受到全球社群蓬勃發展的脈動。我非常感謝公司提供的平台，也倍感榮幸能與這麼多優秀的同事並肩作戰。\nHere’s to another year of momentum and growth! 🚀\n期許下一年，持續保持動能，不斷成長！🚀\nhashtag#TikTokLife hashtag#CareerGrowth hashtag#WorkMilestone hashtag#PassionToProfession © Chung-Hao Lee. All Rights Reserved. All content on this webpage—including but not limited to text, images, design, code, and multimedia materials—is protected under the international copyright treaties. Unauthorized reproduction, modification, distribution, public transmission, or commercial use is strictly prohibited. Legal action will be taken against infringement. © 李崇豪。保留所有權利。 本網頁之內容（包括但不限於文字、圖片、設計、程式碼及多媒體素材）均受國際著作權條約保護。未經書面授權，嚴禁任何形式之複製、改作、散布、公開傳輸或商業利用。侵權者將依法追訴。\n","date":"January 6, 2026","permalink":"https://chunghaolee.com/posts/-1-year-at-tiktok--a-journey-of-turning-dreams-into-reality/","section":"Posts","summary":"\u003cp\u003e“What does it truly feel like to transform a passion into a profession?”\u003c/p\u003e\n\u003cbr\u003e\n\u003cp\u003e「將熱情轉化為職業，究竟是什麼樣的感覺？」\u003c/p\u003e\n\u003chr\u003e\n\u003cp\u003eAs someone who was once a power user spending hours daily on the platform, reaching my one-year milestone at TikTok is more than just a career achievement—it is a dream come true.\u003c/p\u003e\n\u003cbr\u003e\n\u003cp\u003e作為曾經每天在平台上投入數小時的重度使用者，能在 TikTok 迎來入職一週年的里程碑，對我而言，這不僅僅是一項職涯成就，更是夢想成真的時刻。\u003c/p\u003e\n\u003chr\u003e\n\u003cp\u003eOver the past 365 days, I have transitioned from a viewer to a contributor, gaining a deeper understanding of the mechanics behind the content and the vibrant pulse of our global community. I am incredibly grateful for the platform the company has provided and for the opportunity to work alongside such talented colleagues.\u003c/p\u003e","title":"✨ 1 Year at TikTok | A Journey of Turning Dreams into Reality"},{"content":" 📥 Download Portfolio PDF 📄 Full Portfolio Document # Your browser doesn't support PDF viewing. Please download the PDF to view it.\nThis report leverages AI-powered analysis of global rocket launch site optimization criteria, applying a multi-criteria decision-making (MCDM) framework across technical, safety, and commercial dimensions to identify Australia\u0026rsquo;s optimal positioning in the global launch market.\nAustralia has established a world-class dual-niche launch strategy with complementary sites—Whalers Way (WWOLC) offering unmatched safety for polar missions and Arnhem Space Centre (ASC) providing near-equatorial efficiency for GTO missions. While WWOLC is operationally mature and immediately scalable, ASC\u0026rsquo;s technical advantages remain constrained by significant infrastructure deficits. Critical to securing independent access to space is closing the medium-lift capability gap (4,000-23,000kg), which is essential for achieving Australia\u0026rsquo;s $12B industry target. Strategic success requires preserving distinct orbital niches, de-risking ASC logistics through targeted infrastructure investment, and rapidly scaling WWOLC\u0026rsquo;s launch cadence to establish Australia\u0026rsquo;s reputation for reliability.\nKey Highlights #🌍 Dual-Niche Strategy: Australia uniquely positions two complementary launch sites—WWOLC for polar/SSO missions with zero population risk and ASC for equatorial GTO missions with 453.2 m/s rotational advantage—capturing distinct orbital markets globally.\n⚖️ Strategic Assessment Framework: Launch site viability determined by Technical Operations (54.2%), Safety \u0026amp; Security (29.2% as hard constraint), and Commercial factors (16.7%), with safety governing permits, insurance costs, and international liability.\n🚀 Critical Capability Gap: Current light-lift capacity (~300 kg to LEO) insufficient to achieve $12B AUD industry target; medium-lift capability (4,000-23,000 kg) development is essential for securing independent access to space.\n⚡ Commercial Readiness Divide: WWOLC demonstrates high maturity (17 km from port, 65.9 rainy days/year); ASC\u0026rsquo;s technical edge is nullified by $355M infrastructure deficit and 500 km isolation from Darwin via unsealed roads.\n🛡️ Safety as Competitive Edge: WWOLC\u0026rsquo;s unrestricted 135°-296° southern corridor eliminates overflight risks, dramatically reducing insurance costs and attracting high-value payloads—Australia\u0026rsquo;s immediate commercial advantage.\n🔧 Strategic Imperatives: 1) Preserve complementary dual-niche positioning, 2) De-risk ASC through Central Arnhem Road upgrades, 3) Maximize WWOLC throughput to establish reliability reputation, 4) Develop medium-lift capability.\n💡 Strategic Pathway: Transition from specialized launch provider to comprehensive space power requires converting launch sites into national industry catalysts while building infrastructure resilience for higher operational tempo and larger vehicle support.\n© Chung-Hao Lee. All Rights Reserved. All content on this webpage—including but not limited to text, images, design, code, and multimedia materials—is protected under the international copyright treaties. Unauthorized reproduction, modification, distribution, public transmission, or commercial use is strictly prohibited. Legal action will be taken against infringement. © 李崇豪。保留所有權利。 本網頁之內容（包括但不限於文字、圖片、設計、程式碼及多媒體素材）均受國際著作權條約保護。未經書面授權，嚴禁任何形式之複製、改作、散布、公開傳輸或商業利用。侵權者將依法追訴。\n","date":"December 13, 2025","permalink":"https://chunghaolee.com/portfolio/australias-space-launch-roadmap/","section":"Portfolios","summary":"\u003cdiv style=\"text-align: center; margin: 20px 0;\"\u003e\n    \u003ca href=\"Australia Space Launch Roadmap.pdf\" download class=\"download-button\" style=\"display: inline-block; padding: 12px 24px; background-color: #007bff; color: white; text-decoration: none; border-radius: 5px; font-weight: bold; transition: background-color 0.3s;\"\u003e\n        📥 Download Portfolio PDF\n    \u003c/a\u003e\n\u003c/div\u003e\n\u003cstyle\u003e\n.download-button:hover {\n    background-color: #0056b3 !important;\n}\n\n.pdf-container {\n    width: 100%;\n    max-width: 900px;\n    margin: 0 auto;\n    padding: 20px 0;\n}\n\n.pdf-embed {\n    width: 100%;\n    height: 600px;\n    border: 1px solid #ddd;\n    border-radius: 8px;\n    box-shadow: 0 2px 10px rgba(0,0,0,0.1);\n    margin-bottom: 30px;\n}\n\n.pdf-page-image {\n    width: 100%;\n    margin-bottom: 20px;\n    border: 1px solid #e0e0e0;\n    border-radius: 8px;\n    box-shadow: 0 2px 8px rgba(0,0,0,0.08);\n}\n\n@media (max-width: 768px) {\n    .pdf-embed {\n        height: 600px;\n    }\n}\n\u003c/style\u003e\n\u003cdiv class=\"pdf-container\"\u003e\n\u003ch3 id=\"-full-portfolio-document\" class=\"relative group\"\u003e📄 Full Portfolio Document \u003cspan class=\"absolute top-0 w-6 transition-opacity opacity-0 -start-6 not-prose group-hover:opacity-100\"\u003e\u003ca class=\"group-hover:text-primary-300 dark:group-hover:text-neutral-700\" style=\"text-decoration-line: none !important;\" href=\"#-full-portfolio-document\" aria-label=\"Anchor\"\u003e#\u003c/a\u003e\u003c/span\u003e\u003c/h3\u003e\u003c!-- Option 1: Using iframe embed (works if PDF is in static folder) --\u003e\n\u003ciframe src=\"Australia Space Launch Roadmap.pdf\" class=\"pdf-embed\" type=\"application/pdf\"\u003e\n    \u003cp\u003eYour browser doesn't support PDF viewing. Please \u003ca href=\"Australia Space Launch Roadmap.pdf\"\u003edownload the PDF\u003c/a\u003e to view it.\u003c/p\u003e","title":"Australia Space Launch Roadmap: Optimizing Geographic Advantage for Global Competitiveness🇦🇺 "},{"content":"","date":null,"permalink":"https://chunghaolee.com/tags/book-review/","section":"Tags","summary":"","title":"Book Review"},{"content":"As 2025 is about to end, I finished reading Apple in China by Patrick McGee. This book is not merely a chronicle of Apple’s supply chain layout over the past two decades; it is a comprehensive retrospective on the coevolution of East Asian tech manufacturing. For me, a veteran of the Taiwanese EMS industry with hands-on end to end production line management experience spanning SMT (Surface Mount Technology), and FATP (Final Assembly, Test, and Pack), every page felt less like history and more like a return to the factory floor. The book provided a macro retrospective, allowing me to re-examine the environments I once worked in. It illuminated the underlying logic behind production line designs and operations that I often missed while focusing on the daily routine jobs.\n在2025歲末年終之際，我讀完了 Apple in China 這部著作。這本書不僅梳理了過去二十年蘋果的供應鏈佈局，更同步回顧了東亞科技製造業的演進史。 身為台灣科技製造業多年從業者、且親歷過從 SMT（表面黏著技術）到 FATP（最終組裝、測試與包裝）的完整產線管理，書中的每一頁對我而言不僅是歷史，更是現場。它讓我從後見之明的宏觀視角，重新審視那些我們曾經身處其中、卻因專注於細節而未見全貌規劃 並讓我理解了當年產線設計與操作背後的底層邏輯。\nDrawing on my own experience in program management, here are my three key takeaways regarding the supply chain coevolution between Apple, China, and Taiwan.\n以下是我結合自身流程管理經驗，對蘋果、中國與台灣供應鏈的三點感想。\nInterwoven History: Supply Chain Reconstruction and Coevolution In the early 2000s, driven by a pursuit of extreme cost efficiency, precision and mass production, Apple consolidated its global production lines into China. This transformation was executed in collaboration with Taiwanese EMS giants (such as Foxconn, Pegatron, and Wistron), leveraging China\u0026rsquo;s tripartite advantages: labor scale, infrastructure maturity, and local government subsidies. The \u0026ldquo;Apple Squeeze\u0026rdquo; model analyzes how brand engineers worked on-site with suppliers to co-optimize assembly processes and quality management, ensuring high-volume, high-efficiency production. Suppliers traded slim margins for technical accumulation and capacity flexibility, while Apple secured cost-effective, large-scale, and high-quality products. This was not just a geographic shift of the supply chain, but an evolutionary history of interdependence and mutual benefit between Apple, Taiwanese EMS giants and China. 歷史交織：供應鏈重構與效率驅動的共生演化 2000年代初，蘋果為追求極致成本與精度，將分散在世界的產線集中至中國。此轉型由台灣EMS巨頭（如富士康、和碩、緯創）協力完成，背後是中國勞動力規模、基建完善性與政策補貼的三重優勢。\n書中剖析的「蘋果擠壓」模式，在實務中體現為品牌方工程師駐廠協作：品牌方與供應商共同優化組裝工藝與品質管理一同努力 確保產品能夠被大量高效地生產出來。供應商以微幅毛利換取技術沉澱與產能彈性，蘋果則獲得貼近成本、大規模、高品質的產出。這不僅是供應鏈的地理轉移，更是蘋果, 台廠與中國相互依存、互利共生的演化史。 Program as the Moat: Cook’s Absolute Control over the Supply Chain People often attribute Apple\u0026rsquo;s success solely to Steve Jobs\u0026rsquo; design genius. However, as a EMS professional, I know that the \u0026ldquo;Program”—translating blueprints into hundreds of millions of high-quality units—is Apple\u0026rsquo;s true, unreplicable moat. Apple\u0026rsquo;s competitiveness lies in the depth of its \u0026ldquo;active engagement.” Unlike traditional brands that passively await RfQs, Apple co-develops new technologies and equipment with suppliers and secures advantages through \u0026ldquo;Capacity Booking.\u0026rdquo; This strategy creates a win-win. Suppliers receive talent and capital injection to achieve technical upgrades (at the cost of being deeply bound), while Apple monopolizes new technologies and gains absolute bargaining power by understanding every detail of the process and cost structure. From product initiation, R\u0026amp;D, NPI (New Product Introduction) to supply chain management and logistics, mastering every program step allows Apple to maximize commercial interests while ensuring quality. This control over program design and execution details is the essence of electronic devices manufacturing management. 流程即護城河：庫克對供應鏈的完全掌握 外界常誤以為蘋果的成功源於賈伯斯的設計天才，但身為在科技製造業從業多年的我來說，深知將設計圖轉化為數億台高品質產品的「流程」（Process），才是蘋果難以被複製的護城河。蘋果的競爭力在於其「主動介入」的深度。不同於傳統品牌廠被動等待報價，蘋果選擇與供應商共同研發新技術與設備，並透過包下產能（Capacity Booking）來鎖定優勢。\n這種策略創造了雙贏：對供應商而言： 獲得了蘋果的人才與資金挹注，實現了技術升級（儘管代價是被深度綁定）。對蘋果而言： 這種做法使其能最大程度地壟斷新技術，並因為對製程細節與成本結構瞭若指掌，在議價時擁有絕對的話語權。一支現代智慧型手機，從立項、研發、NPI（新產品導入）、供應鏈佈局、製造執行到物流配送，是一項漫長且繁雜的工程。唯有像蘋果這樣，將每一個環節、每一道工序（Process）都牢牢掌握，才能在確保品質的同時實現商業利益最大化。這種對 Program 設計與執行細節的掌控能力，正是蘋果脫穎而出的關鍵 也是科技製造業管理的精髓所在。 Undesigned Plan: The Natural Progression of Commercial Motivation The most resonant part of the book for me was a quote from a former Apple VP: \u0026ldquo;Are you sure you’re not overthinking your thesis? You keep talking about geopolitics, but I was there in the 2000s when we were setting up production in China, and I can tell you, we weren’t thinking about geopolitics at all. Precisely.\u0026rdquo; This statement captures the interplay of accident and inevitability in history. Apple, Taiwanese EMS giants, and China were not playing a grand game of geopolitics; they were simply pursuing their own optimal solutions—efficiency, orders, and economic growth, respectively. These pure commercial motives wove together to drive the upgrade of the entire East Asian tech manufacturing industry. It was a mutual symbiosis and a natural historical trend, rather than a pre-designed master plan. 非設計好的規劃：商業邏輯的自然演進 書中最令我有感觸的，是一段前蘋果副總裁的訪談：\u0026ldquo;你確定你沒有過度解讀嗎？你一直在談地緣政治，但我在 2000 年代設廠時，我可以告訴你，我們當時壓根沒想過地緣政治。\u0026rdquo; 這句話道盡了歷史的偶然與必然。對於當年的蘋果、台灣廠商以及中國方面而言，大家並非在下一盤地緣政治的大棋，而是在單純地追求各自發展的最優解：蘋果追求效率與成本，台廠追求訂單與規模，中方追求就業穩定與經濟產出。正是這些單純的商業動機，在歷史的經緯線上交織，最終演變成今日三方共同協力完成了在東亞整體科技製造產業的升級；而中國強大的製造能力，又反過來成為蘋果鞏固商業帝國的基石。這是一種互利共生，也是歷史發展的自然趨勢，而非當初的任何設計。 © Chung-Hao Lee. All Rights Reserved. All content on this webpage—including but not limited to text, images, design, code, and multimedia materials—is protected under the international copyright treaties. Unauthorized reproduction, modification, distribution, public transmission, or commercial use is strictly prohibited. Legal action will be taken against infringement. © 李崇豪。保留所有權利。 本網頁之內容（包括但不限於文字、圖片、設計、程式碼及多媒體素材）均受國際著作權條約保護。未經書面授權，嚴禁任何形式之複製、改作、散布、公開傳輸或商業利用。侵權者將依法追訴。\n","date":"December 6, 2025","permalink":"https://chunghaolee.com/posts/book-review_apple-in-china/","section":"Posts","summary":"\u003cp\u003eAs 2025 is about to end, I finished reading Apple in China by Patrick McGee. This book is not merely a chronicle of Apple’s supply chain layout over the past two decades; it is a comprehensive retrospective on the coevolution of East Asian tech manufacturing. \u003cbr\u003e\nFor me, a veteran of the Taiwanese EMS industry with hands-on end to end production line management experience spanning SMT (Surface Mount Technology), and FATP (Final Assembly, Test, and Pack), every page felt less like history and more like a return to the factory floor. The book provided a macro retrospective, allowing me to re-examine the environments I once worked in. It illuminated the underlying logic behind production line designs and operations that I often missed while focusing on the daily routine jobs.\u003c/p\u003e","title":"Book Review: Apple in China "},{"content":"As a Product Manager, I’ve often found that the spark of inspiration—sometimes arriving late at night when the world is quiet.\nA good idea is important, but I’ve come to realize that the \u0026ldquo;Aha!\u0026rdquo; moment is just the starting line.\nThe true test of a PM’s value isn\u0026rsquo;t just generating the concept; it is the rigorous, complex challenge of landing it—especially when navigating cross-functional teams.\n身為產品經理，我常發現靈感總在萬籟俱寂的深夜悄然降臨。\n好點子固然重要，但我逐漸體悟到：「啊哈！」時刻僅僅是起跑線。\n真正考驗 PM 價值的，從不只是提出概念；而是如何歷經嚴峻且複雜的過程，將它成功落地——尤其是在協調跨職能團隊之時。\nTo successfully operationalize a vision, three non-negotiable traits are required:\nCrystal-Clear Communication\nIt is not enough to simply have the vision. You must evangelize it until every stakeholder—from Engineering to Marketing—understands the \u0026ldquo;why\u0026rdquo; and the \u0026ldquo;what\u0026rdquo; with 100% fidelity. Strategic Agility\nNo plan survives first contact with reality. Whether it’s technical constraints or team feedback, we must possess the flexibility to adjust the roadmap without losing sight of the core objective.\nWe must be rigid on the vision, but flexible on the details. Relentless Iteration\nThe path to launch is rarely a straight line. It is a continuous loop of alignment, modification, and re-alignment. It requires the grit to push through friction, resistance, and unexpected variables. 要將願景成功轉化為現實，有三項不可或缺的特質：\n清晰無誤的溝通\n光有願景還不夠。你必須像傳道者一樣反覆宣導，直到每位利害關係人——從工程到行銷——都能百分之百準確理解「為什麼做」與「做什麼」。 策略上的敏捷性\n沒有一個計畫能在接觸現實後毫髮無傷。無論是技術限制還是團隊回饋，我們都必須保有調整路線圖的彈性，同時緊扣核心目標。\n願景要堅定，細節要靈活。 不懈的迭代精神\n產品上市之路極少是直線。它是一個不斷對齊、調整、再對齊的循環過程，需要足夠的韌性去穿越摩擦、阻力與各種突發變數。 An idea on paper has potential ; An idea executed is value Navigating this \u0026ldquo;messy middle\u0026rdquo; is exactly why the PM role values.\n紙上的點子只有潛力 ; 落實的點子才創造價值 穿越這段「混亂的中段地帶」，正是產品經理這個角色的價值所在。\n© Chung-Hao Lee. All Rights Reserved. All content on this webpage—including but not limited to text, images, design, code, and multimedia materials—is protected under the international copyright treaties. Unauthorized reproduction, modification, distribution, public transmission, or commercial use is strictly prohibited. Legal action will be taken against infringement. © 李崇豪。保留所有權利。 本網頁之內容（包括但不限於文字、圖片、設計、程式碼及多媒體素材）均受國際著作權條約保護。未經書面授權，嚴禁任何形式之複製、改作、散布、公開傳輸或商業利用。侵權者將依法追訴。\n","date":"November 29, 2025","permalink":"https://chunghaolee.com/posts/2025-11-28_beyond-the-aha-momentthe-distance-between-a-good-idea-and-a-great-product-is-measured-in-execution/","section":"Posts","summary":"\u003cp\u003eAs a Product Manager, I’ve often found that the spark of inspiration—sometimes arriving late at night when the world is quiet.\u003cbr\u003e\nA good idea is important, but I’ve come to realize that the \u0026ldquo;Aha!\u0026rdquo; moment is just the starting line.\u003c/p\u003e\n\u003cp\u003eThe true test of a PM’s value isn\u0026rsquo;t just generating the concept; it is the rigorous, complex challenge of landing it—especially when navigating cross-functional teams.\u003c/p\u003e\n\u003cbr\u003e\n\u003cp\u003e身為產品經理，我常發現靈感總在萬籟俱寂的深夜悄然降臨。\u003cbr\u003e\n好點子固然重要，但我逐漸體悟到：「啊哈！」時刻僅僅是起跑線。\u003c/p\u003e\n\u003cp\u003e真正考驗 PM 價值的，從不只是提出概念；而是如何歷經嚴峻且複雜的過程，將它成功落地——尤其是在協調跨職能團隊之時。\u003c/p\u003e","title":"Beyond the Aha! Moment: The distance between a 'Good Idea' and a 'Great Product' is measured in execution."},{"content":"We often treat PM (Product/Program Management) and Operations as separate silos. But here’s the quiet truth: Value isn\u0026rsquo;t created in a silo.\nThe ultimate metric? A user’s willingness to pay with cash or time. To truly validate that value, you must stop designing from the sidelines—you have to be the operator.\n我們常把產品（或專案）管理（PM）與營運視為彼此隔離的孤島。但有一個少有人說破的真相：價值從來不是在孤島中創造的。\n終極的衡量標準是什麼？就是使用者是否願意付出金錢或時間。要真正驗證這份價值，你就不能只站在旁邊設計——你必須親自下場當營運者。\n💡 The Dual Identity: The \u0026ldquo;Berkshire Hathaway Advantage\u0026rdquo; #This is the core of the \u0026ldquo;Berkshire Hathaway Advantage.\u0026rdquo; I strongly resonate with Warren Buffett’s perspective:\n“I am a better investor because I am a businessman, and a better businessman because I am an investor.”\nThis implies a powerful truth for product leaders: Dual Identity = Dual Advantage.\n💡 雙重身分：「波克夏優勢」 #這正是「波克夏優勢」的核心。我非常認同華倫·巴菲特的觀點：\n「我之所以成為更好的投資人，是因為我同時是企業經營者；而我也因為身為投資人，成為更好的企業經營者。」\n對產品領導者而言，這揭示了一個強而有力的真相：擁有雙重身分，就等於擁抱雙重優勢。\n🔄 The Virtuous Cycle: Hybrid Thinking Wins #The best PMs embrace this duality. When you step out of the \u0026ldquo;designer\u0026rdquo; role and into the \u0026ldquo;operator\u0026rdquo; role, you create a necessary feedback loop:\n🛠️ Design informs Ops: Knowing how the tool was built helps you maximize efficiency and output. 💡 Ops informs Design: Feeling the pain points firsthand forces you to build value, not just features. You stop building in a vacuum.\n🔄 正向循環：跨界思維者勝出 #頂尖的產品經理都擁抱這種雙重角色。當你從「設計者」走入「營運者」的身分，就自然建立起一個必要的反饋循環：\n🛠️ 設計指引營運：了解工具是如何建構的，能幫助你最大化效率與產出。 💡 營運反哺設計：親身體驗痛點，才會迫使你打造真正的「價值」，而不只是堆疊功能。 你不再閉門造車。\n🎯 The Strategic Takeaway: Don’t Just Architect, Inhabit #The most valuable leaders leverage operational friction to refine their strategy. The future belongs to the inhabitants, not the architects.\nIn short: Don\u0026rsquo;t just build the machine—operate it to find the real insights.\n🎯 策略啟示：別只當建築師，更要成為居住者 #最有價值的領導者，懂得利用營運中的摩擦來打磨策略。未來屬於那些「親身居住其中」的人，而非僅僅畫出藍圖的建築師。\n簡言之：別只建造機器——要親自操作它，才能挖出真正的洞察。\n© Chung-Hao Lee. All Rights Reserved. All content on this webpage—including but not limited to text, images, design, code, and multimedia materials—is protected under the international copyright treaties. Unauthorized reproduction, modification, distribution, public transmission, or commercial use is strictly prohibited. Legal action will be taken against infringement. © 李崇豪。保留所有權利。 本網頁之內容（包括但不限於文字、圖片、設計、程式碼及多媒體素材）均受國際著作權條約保護。未經書面授權，嚴禁任何形式之複製、改作、散布、公開傳輸或商業利用。侵權者將依法追訴。\n","date":"November 22, 2025","permalink":"https://chunghaolee.com/posts/2025-11-25-beyond-the-job-title-product-strategy-+-operations/","section":"Posts","summary":"\u003cp\u003eWe often treat PM (Product/Program Management) and Operations as separate silos. But here’s the quiet truth: \u003cstrong\u003eValue isn\u0026rsquo;t created in a silo.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe ultimate metric? A user’s willingness to pay with cash or time. To truly validate that value, you must stop designing from the sidelines—\u003cstrong\u003eyou have to be the operator.\u003c/strong\u003e\u003c/p\u003e\n\u003cbr\u003e\n\u003cp\u003e我們常把產品（或專案）管理（PM）與營運視為彼此隔離的孤島。但有一個少有人說破的真相：\u003cstrong\u003e價值從來不是在孤島中創造的。\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e終極的衡量標準是什麼？就是使用者是否願意付出金錢或時間。要真正驗證這份價值，你就不能只站在旁邊設計——\u003cstrong\u003e你必須親自下場當營運者。\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\n\n\n\n\n\n\u003cfigure\u003e\n    \n    \n\n\n\n\n\n\n\n\n  \n    \u003cpicture\n      class=\"mx-auto my-0 rounded-md\"\n      \n    \u003e\n      \n      \n      \n      \n        \u003csource\n          \n            srcset=\"https://chunghaolee.com/posts/2025-11-25-beyond-the-job-title-product-strategy-\u0026#43;-operations/pic_hu_b80cb23bc907a167.webp 330w,https://chunghaolee.com/posts/2025-11-25-beyond-the-job-title-product-strategy-\u0026#43;-operations/pic_hu_fa2410953fc6f869.webp 660w\n            \n              ,https://chunghaolee.com/posts/2025-11-25-beyond-the-job-title-product-strategy-\u0026#43;-operations/pic_hu_7e5fde52a73909d5.webp 1024w\n            \n            \n              ,https://chunghaolee.com/posts/2025-11-25-beyond-the-job-title-product-strategy-\u0026#43;-operations/pic_hu_599e4417543ba9bb.webp 1320w\n            \"\n          \n          sizes=\"100vw\"\n          type=\"image/webp\"\n        /\u003e\n      \n      \u003cimg\n        width=\"2336\"\n        height=\"1760\"\n        class=\"mx-auto my-0 rounded-md\"\n        \n        loading=\"lazy\" decoding=\"async\"\n        \n          src=\"https://chunghaolee.com/posts/2025-11-25-beyond-the-job-title-product-strategy-\u0026#43;-operations/pic_hu_2a0ae502cf808d84.png\" srcset=\"https://chunghaolee.com/posts/2025-11-25-beyond-the-job-title-product-strategy-\u0026#43;-operations/pic_hu_8542283074e9d332.png 330w,https://chunghaolee.com/posts/2025-11-25-beyond-the-job-title-product-strategy-\u0026#43;-operations/pic_hu_2a0ae502cf808d84.png 660w\n          \n            ,https://chunghaolee.com/posts/2025-11-25-beyond-the-job-title-product-strategy-\u0026#43;-operations/pic_hu_a48990ba6fdb6e4f.png 1024w\n          \n          \n            ,https://chunghaolee.com/posts/2025-11-25-beyond-the-job-title-product-strategy-\u0026#43;-operations/pic_hu_600c53f229332b8.png 1320w\n          \"\n          sizes=\"100vw\"\n        \n      /\u003e\n    \u003c/picture\u003e\n  \n\n\n\u003c/figure\u003e\n\u003c/p\u003e","title":"Beyond the Job Title: Product Strategy + Operations (The Berkshire Hathaway Advantage) is the Real Moat 📉📈"},{"content":" 📥 Download Portfolio PDF 📄 Full Portfolio Document # Your browser doesn't support PDF viewing. Please download the PDF to view it.\nKey Highlights #🌍 AI Mineral Demand Surge: The AI revolution is driving unprecedented demand for specific critical minerals, with data centers requiring copper (25%), silicon (15%), and rare earth elements (8%), while energy infrastructure demands coal (45%), copper (20%), and lithium (15%).\n🇦🇺 Australia\u0026rsquo;s Strategic Mineral Advantage: With world-class deposits of copper, lithium, iron ore, and rare earth elements, Australia holds a pivotal position as a foundational materials supplier for the AI era—ranking #2 globally for copper resources and #5 for refined copper exports.\n💰 Growing Export Value Despite Volume Challenges: Australian copper exports demonstrate the sector\u0026rsquo;s increasing strategic importance, with export values forecast to rise significantly due to higher prices—even as export volumes stabilize. Total resource exports reached $410 billion in 2023-2024.\n🤝 Strategic Interdependence with Key Markets: China, Japan, and India account for 52% of Australia\u0026rsquo;s resource exports, with China alone receiving 37% ($152 billion), particularly dominating imports of AI-critical minerals like iron ore ($116B), gold ($16B), coal ($10.7B), and lithium ($9.4B).\n🔗 Synergistic Value Chain Partnership: Australia\u0026rsquo;s raw material exports complement China\u0026rsquo;s processing and manufacturing capabilities, creating a powerful economic interdependence that transforms resource wealth into technological advancement.\n🚀 Strategic Horizon for Shared Prosperity: The report positions Australia-China mineral-technology interdependence not as a vulnerability but as a strategic opportunity to co-lead the next wave of global technological advancement through managed collaboration.\n© Chung-Hao Lee. All Rights Reserved. All content on this webpage—including but not limited to text, images, design, code, and multimedia materials—is protected under the international copyright treaties. Unauthorized reproduction, modification, distribution, public transmission, or commercial use is strictly prohibited. Legal action will be taken against infringement. © 李崇豪。保留所有權利。 本網頁之內容（包括但不限於文字、圖片、設計、程式碼及多媒體素材）均受國際著作權條約保護。未經書面授權，嚴禁任何形式之複製、改作、散布、公開傳輸或商業利用。侵權者將依法追訴。\n","date":"November 22, 2025","permalink":"https://chunghaolee.com/portfolio/2025-11-23-powering-the-ai-revolution-australias-pivotal-role-in-the-global-mineral-supply-chain/","section":"Portfolios","summary":"\u003cdiv style=\"text-align: center; margin: 20px 0;\"\u003e\n    \u003ca href=\"AI Revolution - Australia Pivotal Role in the Global Mineral Supply Chain.pdf\" download class=\"download-button\" style=\"display: inline-block; padding: 12px 24px; background-color: #007bff; color: white; text-decoration: none; border-radius: 5px; font-weight: bold; transition: background-color 0.3s;\"\u003e\n        📥 Download Portfolio PDF\n    \u003c/a\u003e\n\u003c/div\u003e\n\u003cstyle\u003e\n.download-button:hover {\n    background-color: #0056b3 !important;\n}\n\n.pdf-container {\n    width: 100%;\n    max-width: 900px;\n    margin: 0 auto;\n    padding: 20px 0;\n}\n\n.pdf-embed {\n    width: 100%;\n    height: 600px;\n    border: 1px solid #ddd;\n    border-radius: 8px;\n    box-shadow: 0 2px 10px rgba(0,0,0,0.1);\n    margin-bottom: 30px;\n}\n\n.pdf-page-image {\n    width: 100%;\n    margin-bottom: 20px;\n    border: 1px solid #e0e0e0;\n    border-radius: 8px;\n    box-shadow: 0 2px 8px rgba(0,0,0,0.08);\n}\n\n@media (max-width: 768px) {\n    .pdf-embed {\n        height: 600px;\n    }\n}\n\u003c/style\u003e\n\u003cdiv class=\"pdf-container\"\u003e\n\u003ch3 id=\"-full-portfolio-document\" class=\"relative group\"\u003e📄 Full Portfolio Document \u003cspan class=\"absolute top-0 w-6 transition-opacity opacity-0 -start-6 not-prose group-hover:opacity-100\"\u003e\u003ca class=\"group-hover:text-primary-300 dark:group-hover:text-neutral-700\" style=\"text-decoration-line: none !important;\" href=\"#-full-portfolio-document\" aria-label=\"Anchor\"\u003e#\u003c/a\u003e\u003c/span\u003e\u003c/h3\u003e\u003c!-- Option 1: Using iframe embed (works if PDF is in static folder) --\u003e\n\u003ciframe src=\"AI Revolution - Australia Pivotal Role in the Global Mineral Supply Chain.pdf\" class=\"pdf-embed\" type=\"application/pdf\"\u003e\n    \u003cp\u003eYour browser doesn't support PDF viewing. Please \u003ca href=\"AI Revolution - Australia Pivotal Role in the Global Mineral Supply Chain.pdf\"\u003edownload the PDF\u003c/a\u003e to view it.\u003c/p\u003e","title":"AI Revolution - Australia Pivotal Role in the Global Mineral Supply Chain for AI 🇦🇺 "},{"content":"I’m truly grateful to The Chartered Institute of Building (CIOB) and Turner \u0026amp; Townsend for hosting such an insightful event on \u0026ldquo;Transforming Project Delivery with AI\u0026rdquo;. It was a fantastic evening of high-level presentations and invaluable networking.\nAs a professional who has just relocated from Taiwan to Sydney, this was my first industry networking event in Australia, and it was an incredibly welcoming experience. I gained significant insights into the current landscape of AI in the local construction sector and had the pleasure of connecting with many dedicated experts.\nA special thanks to George Highfield from Turner \u0026amp; Townsend for his brilliant presentation on two powerful AI applications currently being leveraged:\nAI Agent \u0026ldquo;Dream Teams\u0026rdquo;: I was particularly impressed by the concept of assembling a \u0026ldquo;dream team\u0026rdquo; of specialized AI agents. By dedicating different agents to specific tasks—such as cost management or scheduling master—and having them collaborate, efficiency can be dramatically enhanced.\nRAG for Data-Intensive Roles: The application of Retrieval-Augmented Generation (RAG) is a true game-changer for professions like Quantity Surveying. Historically, a QS would need to manually sift through countless PDFs and spreadsheets to find specific data points from different quotes or schedules. With RAG, they can now use natural language queries to find accurate information instantly. This frees up professionals to focus their time and energy on high-value analysis and strategic decision-making.\nThe key takeaway is clear: AI is a powerful tool for augmenting human capability, creating a win-win synergy that drives efficiency and better outcomes.\nAs a fellow AI professional, I am incredibly excited to be part of Australia\u0026rsquo;s growing AI ecosystem. I am eager to connect, collaborate, and network with more industry leaders here. Let\u0026rsquo;s continue to exchange ideas and build the future of construction together!\n#AIinAustralia #Networking #AI #construction\n© Chung-Hao Lee. All Rights Reserved. All content on this webpage—including but not limited to text, images, design, code, and multimedia materials—is protected under the international copyright treaties. Unauthorized reproduction, modification, distribution, public transmission, or commercial use is strictly prohibited. Legal action will be taken against infringement. © 李崇豪。保留所有權利。 本網頁之內容（包括但不限於文字、圖片、設計、程式碼及多媒體素材）均受國際著作權條約保護。未經書面授權，嚴禁任何形式之複製、改作、散布、公開傳輸或商業利用。侵權者將依法追訴。\n","date":"November 1, 2025","permalink":"https://chunghaolee.com/posts/2025-11-01_1st_au_networking/","section":"Posts","summary":"\u003cp\u003eI’m truly grateful to The Chartered Institute of Building (CIOB) and Turner \u0026amp; Townsend for hosting such an insightful event on \u0026ldquo;Transforming Project Delivery with AI\u0026rdquo;. It was a fantastic evening of high-level presentations and invaluable networking.\u003c/p\u003e\n\u003cp\u003eAs a professional who has just relocated from Taiwan to Sydney, this was my first industry networking event in Australia, and it was an incredibly welcoming experience. I gained significant insights into the current landscape of AI in the local construction sector and had the pleasure of connecting with many dedicated experts.\u003c/p\u003e","title":"AI-Human Synergy in Project Delivery: Reflections on Australia’s Construction Industry Networking Event"},{"content":"身為一位PM,我總是喜歡以一位使用者角度來看一個產品。這次想來跟著大家一起來看我心中一個很好的產品 — Messenger。\n一、 Messenger從Facebook App中獨立出來 #Messenger 本來是 Facebook 的一個通訊功能,但在 2015 年時被 Facebook 分拆出來,要求用戶單獨下載 App 來做使用。\n坦白說,當時我覺得很麻煩,覺得同樣都是 Facebook,為什麼要分兩個 App 做使用 (Facebook: 主要用來瀏覽動態 / Messenger: 主要用來與人溝通)。\n但是後來看,會發現這是非常高明的一步。\n用戶視角 #以用戶視角來看,我很大的需求是想要與我的朋友溝通、分享資訊。\n如果 Messenger 還是被綁定在 Facebook 內部的話,那我要與我的朋友溝通時,總共要幾步呢?\n打開 Facebook app 打開 Messenger 找到我朋友 開始溝通 如果 Messenger 獨立出來成為一個 App,那我要與我的朋友溝通時,總共要幾步呢?\n打開 Messenger 找到我朋友 開始溝通 別看只有一步之差,但因為 Facebook 要同時容納動態與 Messenger,在排版佈局上,勢必會很複雜,因為要兼容很多內容。所以在 Facebook App 裡面要找 Messenger 的入口,會比較沒有那麼友好。\n產品經理視角 #也因為沒那麼友好,從 PM 的視角看,如果 Messenger 當初沒跟 Facebook 分拆的話,Messenger 的留存率一定不會像現在那麼好。\n不論從用戶視角還是PM視角,Messenger 從 Facebook 分拆都是一件非常高明的一步。\n二、 引用回覆 (Quoted Replies) #這是我在 Messenger 中最喜歡的功能了,也認為這是有史以來最棒的設計之一。Messenger 的功能就是與人的對話溝通,在對話的過程中,往往雙方可能會有來來回回,或者我想要回覆之前某個時候的內容,這時候就是「引用回覆」這功能最好的使用時機。\n在 Messenger 的引用回覆有三點,是我認為很偉大的設計:\n(1) 非常直覺的交互方式 #跟別的 App 不同,在 Messenger 使用引用回覆不需要先長按、點出彈窗、再點選引用回覆的功能這麼繁瑣的步驟。只需要在你想要引用的消息,輕輕的按住左滑 or 右滑(取決於這消息是你的還是你朋友的),就可以完成引用回覆,非常的直覺而且省去極度繁瑣的步驟。\n(2) 引用回覆的呈現 #在 Messenger 引用回覆的視覺呈現是:會在你的回覆消息上面,呈現一個小的、顏色較淡的窗口,並放上被你引用的消息 (不論這消息是文字還是圖片,甚至是連結都會用縮圖呈現!)。\n這麼做的好處是,當使用者要查看時,可以第一時間看到被引用的資訊是什麼,就不用像其他 App 一樣,還需要往上滑才能看到被引用的消息,讓使用者整個的閱讀體驗跟效率提升了一個檔次。\n(3) 點選被引用消息的後續交互 #在 Messenger 上,如果使用者想要查看被引用的消息,只需要在被引用消息的小窗口點一下,系統就會自動的往上滾動到被引用消息的原始處。\n這方便使用者可以很快看到被引用消息原始處的上下文;或者,如果被引用消息是一個連結,使用者也可以很快回到原始處的連結,並點一下就可以進到連結。這種交互非常的直覺,且讓使用者查看訊息時的效率提升,提高了使用這體驗。\n也因為上面三種原因,讓我覺得「引用回覆」是一個非常直覺、交互設計優良的功能,大大提升使用者體驗,也是 Messenger 在我看來是一項好產品的主要原因。 © Chung-Hao Lee. All Rights Reserved. All content on this webpage—including but not limited to text, images, design, code, and multimedia materials—is protected under the international copyright treaties. Unauthorized reproduction, modification, distribution, public transmission, or commercial use is strictly prohibited. Legal action will be taken against infringement. © 李崇豪。保留所有權利。 本網頁之內容（包括但不限於文字、圖片、設計、程式碼及多媒體素材）均受國際著作權條約保護。未經書面授權，嚴禁任何形式之複製、改作、散布、公開傳輸或商業利用。侵權者將依法追訴。\n","date":"October 19, 2025","permalink":"https://chunghaolee.com/posts/2025-11-04-product-decoder-messenger-%E4%B8%AD%E6%96%87%E7%89%88/","section":"Posts","summary":"\u003cp\u003e身為一位PM,我總是喜歡以一位使用者角度來看一個產品。這次想來跟著大家一起來看我心中一個很好的產品 — Messenger。\u003c/p\u003e\n\u003cimg src=\"pic1.png\" alt=\"logo of messenger\" width=\"80\" style=\"display: block; margin: 20px auto; border-radius: 8px; box-shadow: 0 2px 8px rgba(0,0,0,0.1);\"\u003e\n\u003chr\u003e\n\u003ch2 id=\"一-messenger從facebook-app中獨立出來\" class=\"relative group\"\u003e一、 Messenger從Facebook App中獨立出來 \u003cspan class=\"absolute top-0 w-6 transition-opacity opacity-0 -start-6 not-prose group-hover:opacity-100\"\u003e\u003ca class=\"group-hover:text-primary-300 dark:group-hover:text-neutral-700\" style=\"text-decoration-line: none !important;\" href=\"#%e4%b8%80-messenger%e5%be%9efacebook-app%e4%b8%ad%e7%8d%a8%e7%ab%8b%e5%87%ba%e4%be%86\" aria-label=\"Anchor\"\u003e#\u003c/a\u003e\u003c/span\u003e\u003c/h2\u003e\u003cp\u003eMessenger 本來是 Facebook 的一個通訊功能,但在 2015 年時被 Facebook 分拆出來,要求用戶單獨下載 App 來做使用。\u003c/p\u003e\n\u003cp\u003e坦白說,當時我覺得很麻煩,覺得同樣都是 Facebook,為什麼要分兩個 App 做使用 (Facebook: 主要用來瀏覽動態 / Messenger: 主要用來與人溝通)。\u003c/p\u003e\n\u003cp\u003e但是後來看,會發現這是非常高明的一步。\u003c/p\u003e\n\u003ch3 id=\"用戶視角\" class=\"relative group\"\u003e用戶視角 \u003cspan class=\"absolute top-0 w-6 transition-opacity opacity-0 -start-6 not-prose group-hover:opacity-100\"\u003e\u003ca class=\"group-hover:text-primary-300 dark:group-hover:text-neutral-700\" style=\"text-decoration-line: none !important;\" href=\"#%e7%94%a8%e6%88%b6%e8%a6%96%e8%a7%92\" aria-label=\"Anchor\"\u003e#\u003c/a\u003e\u003c/span\u003e\u003c/h3\u003e\u003cp\u003e以用戶視角來看,我很大的需求是想要與我的朋友溝通、分享資訊。\u003c/p\u003e\n\u003cp\u003e如果 Messenger 還是被綁定在 Facebook 內部的話,那我要與我的朋友溝通時,總共要幾步呢?\u003c/p\u003e\n\u003col\u003e\n\u003cli\u003e打開 Facebook app\u003c/li\u003e\n\u003cli\u003e打開 Messenger\u003c/li\u003e\n\u003cli\u003e找到我朋友\u003c/li\u003e\n\u003cli\u003e開始溝通\u003c/li\u003e\n\u003c/ol\u003e\n\u003cp\u003e如果 Messenger 獨立出來成為一個 App,那我要與我的朋友溝通時,總共要幾步呢?\u003c/p\u003e","title":"Product Decoder: Messenger 的拆分策略與引用回覆功能解析"},{"content":"As a PM, I always like to view products from a user\u0026rsquo;s perspective. This time, I\u0026rsquo;d like to analyze a product I consider to be excellent — Messenger.\nI. Messenger\u0026rsquo;s Unbundling from the Facebook App #Messenger was originally a messaging feature within Facebook, but around 2015, Facebook unbundled it, requiring users to download a separate app to use it.\nFrankly, I found it inconvenient at the time, wondering why I needed two separate apps from the same company (Facebook: mainly for browsing the feed / Messenger: mainly for communication).\nIn hindsight, however, this turned out to be a brilliant move.\nThe User\u0026rsquo;s Perspective #From a user\u0026rsquo;s point of view, a primary need is to communicate and share information with my friends.\nIf Messenger were still bundled within the Facebook app, how many steps would it take to communicate with a friend?\nOpen the Facebook app Open Messenger Find my friend Start communicating With Messenger as a standalone app, how many steps does it take?\nOpen Messenger Find my friend Start communicating Don\u0026rsquo;t underestimate this one-step difference. Because the Facebook app had to accommodate both the News Feed and Messenger, its layout was bound to be complex to support so much content. Consequently, finding the entry point for Messenger within the Facebook app was less user-friendly.\nThe Product Manager\u0026rsquo;s Perspective #Because of this friction, from a PM\u0026rsquo;s perspective, if Messenger hadn\u0026rsquo;t been unbundled from Facebook, its retention rate would certainly not be as high as it is today.\nFrom both the user\u0026rsquo;s and the PM\u0026rsquo;s perspective, unbundling Messenger from Facebook was a brilliant move.\nII. Quoted Replies #This is my favorite feature in Messenger, and I consider it one of the best designs ever. The core function of Messenger is to converse with people. In a conversation, there\u0026rsquo;s often back-and-forth, or I might want to reply to a specific message from earlier. This is where the \u0026ldquo;quoted reply\u0026rdquo; feature shines.\nThere are three aspects of Messenger\u0026rsquo;s quoted replies that I find brilliant:\n(1) A Highly Intuitive Interaction #Unlike other apps, using quoted replies in Messenger doesn\u0026rsquo;t require a long press, waiting for a pop-up, and then selecting the \u0026lsquo;reply\u0026rsquo; option. You simply swipe left or right on the message you want to quote (depending on whether it\u0026rsquo;s your message or your friend\u0026rsquo;s). This is incredibly intuitive and eliminates tedious steps.\n(2) The Visual Presentation of the Quoted Reply #The visual presentation of a quoted reply in Messenger appears as a smaller, lighter-colored container above your reply, displaying the original message (and it even shows a thumbnail for images or links!).\nThe advantage of this is that the user can immediately see the context of the reply at a glance. This avoids the need to scroll up to find the original message, as required in other apps, elevating the user\u0026rsquo;s reading experience and efficiency to another level.\n(3) The Subsequent Interaction of Tapping the Quoted Message #In Messenger, if a user wants to see the original quoted message, they just need to tap on the small quote container, and the app automatically scrolls up to the original message\u0026rsquo;s location.\nThis allows the user to quickly see the surrounding context of the original message. If the quote is a link, the user can easily return to the original message to click it. This interaction is extremely intuitive and enhances the efficiency of reading messages, significantly improving the user experience.\nFor these three reasons, I believe the \u0026ldquo;quoted reply\u0026rdquo; is a highly intuitive and well-designed interaction feature that greatly enhances the user experience. It\u0026rsquo;s a major reason why I consider Messenger to be such a great product.\n© Chung-Hao Lee. All Rights Reserved. All content on this webpage—including but not limited to text, images, design, code, and multimedia materials—is protected under the international copyright treaties. Unauthorized reproduction, modification, distribution, public transmission, or commercial use is strictly prohibited. Legal action will be taken against infringement. © 李崇豪。保留所有權利。 本網頁之內容（包括但不限於文字、圖片、設計、程式碼及多媒體素材）均受國際著作權條約保護。未經書面授權，嚴禁任何形式之複製、改作、散布、公開傳輸或商業利用。侵權者將依法追訴。\n","date":"October 19, 2025","permalink":"https://chunghaolee.com/posts/2025-11-04-product-decoder-messenger/","section":"Posts","summary":"\u003cp\u003eAs a PM, I always like to view products from a user\u0026rsquo;s perspective. This time, I\u0026rsquo;d like to analyze a product I consider to be excellent — Messenger.\u003c/p\u003e\n\u003cimg src=\"pic1.png\" alt=\"logo of messenger\" width=\"80\" style=\"display: block; margin: 20px auto; border-radius: 8px; box-shadow: 0 2px 8px rgba(0,0,0,0.1);\"\u003e\n\u003chr\u003e\n\u003ch2 id=\"i-messengers-unbundling-from-the-facebook-app\" class=\"relative group\"\u003eI. Messenger\u0026rsquo;s Unbundling from the Facebook App \u003cspan class=\"absolute top-0 w-6 transition-opacity opacity-0 -start-6 not-prose group-hover:opacity-100\"\u003e\u003ca class=\"group-hover:text-primary-300 dark:group-hover:text-neutral-700\" style=\"text-decoration-line: none !important;\" href=\"#i-messengers-unbundling-from-the-facebook-app\" aria-label=\"Anchor\"\u003e#\u003c/a\u003e\u003c/span\u003e\u003c/h2\u003e\u003cp\u003eMessenger was originally a messaging feature within Facebook, but around 2015, Facebook unbundled it, requiring users to download a separate app to use it.\u003c/p\u003e","title":"Product Decoder: Messenger's Unbundling Strategy and Quoted Reply Feature Analysis"},{"content":"PM 的本質：以所有權驅動跨職能協作的整合型領導者與問題解決者 # Regardless of whether the title is Product Manager, Project Manager, or Program Manager, the core essence of the “PM” role does not lie in whether one manages a product, a project, or a program, but in whether one embodies an ownership mindset—taking ultimate accountability for the assigned domain and ensuring its continuous alignment with and advancement toward the organization’s strategic objectives.\n無論職稱為產品經理（Product Manager）、專案經理（Project Manager）或流程經理（Program Manager），「PM」這一角色的核心本質並非取決於其管理對象是產品、專案或流程，而在於是否具備所有權思維（Ownership Mindset）,即對所負責範疇承擔最終責任，並確保其持續朝向企業戰略目標推進。\nThe scope of a PM’s responsibilities can vary significantly in scale: from spearheading the end-to-end development and launch of an entirely new product or designing and driving a comprehensive business process transformation, to optimizing a single user interface element—such as adjusting a button color to improve click-through rates. Regardless of scale, the critical point is that the PM must be accountable for outcomes and ensure every action delivers measurable value.\nPM 的職責範圍可大可小：大至從零開始主導一個全新產品的孵化與上市，或設計並推動一整套端到端的業務流程變革；小至優化一個使用者介面元素（例如調整按鈕顏色以提升點擊率）。無論規模如何，關鍵在於 PM 必須對結果負責，並確保每一項行動皆能創造可衡量的價值。\nIn practice, the primary challenge for a PM is rarely “what to do,” but rather “how to make things happen within a complex organization.” Common obstacles include resource constraints, timeline pressures, misaligned cross-functional priorities, and even blame-shifting or collaboration inertia—collectively referred to as “execution-level problems.” Notably, while both PMs and engineers are problem solvers, their approaches differ fundamentally:\n在實踐過程中，PM 面臨的挑戰往往不在於「要做什麼」，而在於「如何在複雜組織中推動事情發生」。常見障礙包括資源限制、時程壓力、跨部門目標不一致，乃至於責任推諉與協作惰性——這些皆可統稱為「執行層面的問題」。值得注意的是，PM 與工程師雖同為問題解決者，但其解決路徑存在根本差異：\nEngineers tend to approach problems from the perspective of technical feasibility and system logic, focusing on “how to solve the problem from a technical standpoint”;\nPMs, by contrast, must navigate both the “task” and the “people” dimensions: they not only need to understand the intersection of technology and business, but also delve into organizational dynamics—leveraging strategic communication, interest alignment, the art of compromise, and informal influence to resolve conflicts, build consensus, and drive cross-functional teams toward a shared goal, often without formal authority.\n工程師傾向從技術可行性與系統邏輯出發，聚焦於「如何從技術面解決問題」；\nPM則必須同時處理「事」與「人」的雙重維度：不僅需理解技術與商業的交界，更需深入組織行為層面，透過策略性溝通、利益協調、妥協藝術與非正式影響力，在缺乏直接指揮權的情境下，化解衝突、建立共識，並驅動跨職能團隊朝共同目標前進。\nPrecisely for this reason, a PM resembles an “enhanced engineer”: in addition to possessing solid technical competence and domain expertise (hard skills), they must also master advanced interpersonal capabilities such as communication, coordination, negotiation, and influence (soft skills). This dual-track capability enables PMs to bridge the gap between technical feasibility and business value, while simultaneously navigating between organizational realities and aspirational goals.\n正因如此，PM 更像是「加強版的工程師」：除了需具備扎實的技術力與領域專業（硬實力），更必須精通溝通、協調、談判與影響力等高階人際技能（軟實力）。這種雙軌能力使 PM 能在技術可行性與商業價值之間架起橋樑，同時在組織現實與理想目標之間找到可行路徑。\nIn other words, an exceptional PM must not only “understand the problem” and “understand the technology,” but also “understand people.” They excel at balancing limited resources against competing expectations, transforming fragmented stakeholders into collaborative partners through continuous dialogue, strategic concessions, and value-based trade-offs. This capability goes far beyond mere process management or requirement documentation—it represents a high-level integrative skill set that blends business judgment, organizational political acumen, and interpersonal influence.\n換言之，優秀的 PM 不僅要「懂問題」，「懂技術」,更要「懂人」。他們善於在資源與期望之間尋找平衡點，透過持續對話、讓步與價值交換，將分散的利害關係人轉化為協作夥伴。這種能力，遠超單純的流程管理或需求撰寫，而是一種融合商業判斷、組織政治智慧與人際影響力的高階整合技能。\nTherefore, the true value of a PM lies not in controlling processes, but in anchoring efforts in ownership and systematically dismantling organizational execution barriers through technical insight and interpersonal coordination—ensuring that the product, project, or program under their purview, regardless of scale, is effectively delivered and drives strategic outcomes. This positions the PM as an indispensable linchpin in strategic execution within modern enterprises.\n因此，PM 的真正價值不在於控制流程，而在於以所有權為錨點，運用技術理解力與人際協調力，系統性破解組織執行障礙，確保所負責的產品、專案或流程，無論規模大小，皆能有效落地並驅動戰略成果。這使其成為現代企業中不可或缺戰略執行的關鍵樞紐。\n© Chung-Hao Lee. All Rights Reserved. All content on this webpage—including but not limited to text, images, design, code, and multimedia materials—is protected under the international copyright treaties. Unauthorized reproduction, modification, distribution, public transmission, or commercial use is strictly prohibited. Legal action will be taken against infringement. © 李崇豪。保留所有權利。 本網頁之內容（包括但不限於文字、圖片、設計、程式碼及多媒體素材）均受國際著作權條約保護。未經書面授權，嚴禁任何形式之複製、改作、散布、公開傳輸或商業利用。侵權者將依法追訴。\n","date":"October 14, 2025","permalink":"https://chunghaolee.com/posts/2025-10-21-pm-%E7%9A%84%E6%9C%AC%E8%B3%AA/","section":"Posts","summary":"\u003ch2 id=\"pm-的本質以所有權驅動跨職能協作的整合型領導者與問題解決者\" class=\"relative group\"\u003ePM 的本質：以所有權驅動跨職能協作的整合型領導者與問題解決者 \u003cspan class=\"absolute top-0 w-6 transition-opacity opacity-0 -start-6 not-prose group-hover:opacity-100\"\u003e\u003ca class=\"group-hover:text-primary-300 dark:group-hover:text-neutral-700\" style=\"text-decoration-line: none !important;\" href=\"#pm-%e7%9a%84%e6%9c%ac%e8%b3%aa%e4%bb%a5%e6%89%80%e6%9c%89%e6%ac%8a%e9%a9%85%e5%8b%95%e8%b7%a8%e8%81%b7%e8%83%bd%e5%8d%94%e4%bd%9c%e7%9a%84%e6%95%b4%e5%90%88%e5%9e%8b%e9%a0%98%e5%b0%8e%e8%80%85%e8%88%87%e5%95%8f%e9%a1%8c%e8%a7%a3%e6%b1%ba%e8%80%85\" aria-label=\"Anchor\"\u003e#\u003c/a\u003e\u003c/span\u003e\u003c/h2\u003e\u003cp\u003e\n\n\n\n\n\n\u003cfigure\u003e\n    \n    \n\n\n\n\n\n\n\n\n  \n    \u003cpicture\n      class=\"mx-auto my-0 rounded-md\"\n      \n    \u003e\n      \n      \n      \n      \n        \u003csource\n          \n            srcset=\"https://chunghaolee.com/posts/2025-10-21-pm-%E7%9A%84%E6%9C%AC%E8%B3%AA/pic_hu_21b2284d2d68c823.webp 330w,https://chunghaolee.com/posts/2025-10-21-pm-%E7%9A%84%E6%9C%AC%E8%B3%AA/pic_hu_d7cf93ca76c699be.webp 660w\n            \n              ,https://chunghaolee.com/posts/2025-10-21-pm-%E7%9A%84%E6%9C%AC%E8%B3%AA/pic_hu_79e549f7ad27d8a7.webp 1024w\n            \n            \n              ,https://chunghaolee.com/posts/2025-10-21-pm-%E7%9A%84%E6%9C%AC%E8%B3%AA/pic_hu_77efe362f236e1a.webp 1320w\n            \"\n          \n          sizes=\"100vw\"\n          type=\"image/webp\"\n        /\u003e\n      \n      \u003cimg\n        width=\"2364\"\n        height=\"1773\"\n        class=\"mx-auto my-0 rounded-md\"\n        \n        loading=\"lazy\" decoding=\"async\"\n        \n          src=\"https://chunghaolee.com/posts/2025-10-21-pm-%E7%9A%84%E6%9C%AC%E8%B3%AA/pic_hu_44312a14cf79ded5.png\" srcset=\"https://chunghaolee.com/posts/2025-10-21-pm-%E7%9A%84%E6%9C%AC%E8%B3%AA/pic_hu_a7bde19d2e6d639e.png 330w,https://chunghaolee.com/posts/2025-10-21-pm-%E7%9A%84%E6%9C%AC%E8%B3%AA/pic_hu_44312a14cf79ded5.png 660w\n          \n            ,https://chunghaolee.com/posts/2025-10-21-pm-%E7%9A%84%E6%9C%AC%E8%B3%AA/pic_hu_9a57ddc1b7adb78f.png 1024w\n          \n          \n            ,https://chunghaolee.com/posts/2025-10-21-pm-%E7%9A%84%E6%9C%AC%E8%B3%AA/pic_hu_dd8ea308df11ddc5.png 1320w\n          \"\n          sizes=\"100vw\"\n        \n      /\u003e\n    \u003c/picture\u003e\n  \n\n\n\u003c/figure\u003e\n\u003c/p\u003e\n\u003cp\u003eRegardless of whether the title is Product Manager, Project Manager, or Program Manager, the core essence of the “PM” role does not lie in whether one manages a product, a project, or a program, but in whether one embodies an ownership mindset—taking ultimate accountability for the assigned domain and ensuring its continuous alignment with and advancement toward the organization’s strategic objectives.\u003c/p\u003e","title":"The Essence of PM: An Integrative Leader and Problem Solver Driving Cross-Functional Collaboration Through Ownership"},{"content":"","date":"October 11, 2025","permalink":"https://chunghaolee.com/posts/2025-10-12-%E7%BB%93%E6%9D%9F%E7%9A%84%E5%BC%80%E5%A7%8B%E4%B8%A8tiktok-%E4%BA%94%E5%B9%B4%E8%87%AA%E6%95%91%E5%88%B0%E7%BB%88%E7%82%B9%E5%8A%9E%E5%85%AC%E5%AE%A4%E5%AE%89%E9%9D%99%E5%A6%82%E5%B8%B8/","section":"Posts","summary":"","title":"[External Article] - 結束的開始丨TikTok 五年自救到終點，辦公室安靜如常"},{"content":"I’ve had the privilege of working as both an engineer and a product manager—and I’m often asked: “Aren’t those roles totally different?” My answer might surprise you: they’re not opposites. They’re complementary lenses on the same mission.\n我有幸同時擔任過工程師與產品經理，常有人問我：「這兩個角色難道不是天差地遠嗎？」我的答案或許出乎意料：它們並非對立，而是同一使命下的互補視角。\nSurface Differences, Shared Purpose #At first glance, the roles appear distinct:\nEngineers focus on architecture, code, and system performance—turning constraints into feasible, scalable solutions. Product Managers (PMs) focus on user value, product strategy, and cross-team alignment—translating customer needs into prioritized outcomes. Yet beneath these surface distinctions lies a unified core: both roles exist to solve problems. They simply begin from different vantage points.\n表面差異，共同使命 #乍看之下，這兩個角色確實不同：\n工程師聚焦於系統架構、程式碼與效能表現——將限制條件轉化為可行且可擴展的解決方案。 **產品經理(PM)**則聚焦於用戶價值、產品策略與跨團隊協調——將客戶需求轉譯為優先排序的成果。 然而，在這些表象差異之下，藏著一個共通的核心：兩個角色的存在，都是為了「解決問題」，只是出發點不同罷了。\n🧩 Two Lenses, One Mission # Role Core Question Primary Translation Engineer “Is this technically feasible? How do we design a reliable, maintainable solution?” Ambiguity → Implementation Product Manager “What problem truly matters to the user and the business? What should we build first?” User Insight → Prioritized Scope \u0026amp; Resources When teams operate at their best, engineering and product don’t compete—they translate for each other:\nEngineering turns business goals into robust technical plans. Product turns technical possibilities into measurable user value. This symbiotic relationship is the engine of high-performing product teams.\n🧩 雙重視角，同一使命 # 角色 核心問題 主要轉譯方向 工程師 「這在技術上可行嗎？如何設計出可靠且易維護的方案？」 將模糊需求 → 落實為具體實作 產品經理 「對用戶和業務而言，哪個問題真正關鍵？我們該先做什麼？」 將用戶洞察 → 轉化為優先級明確的範疇與資源配置 當團隊運作達到最佳狀態時，工程與產品並非彼此競爭，而是互相翻譯：\n工程將商業目標轉化為堅實的技術規劃。 產品將技術可能性轉化為可衡量的用戶價值。 這種共生關係，正是高效產品團隊的核心引擎。\n🚀 The Strategic Takeaway: The “Translator” Is Your MVP #The most valuable skill in modern product development isn’t deep specialization alone—it’s fluency across domains.\nThe superpower lies in the ability to:\n• Translate business intent into technical strategy, and\n• Translate technical constraints into product trade-offs.\nThis dual fluency unlocks:\nVelocity — fewer miscommunications, faster iteration Trust — mutual respect between disciplines Innovation — creative solutions born at the intersection of feasibility and desirability 🚀 策略啟示：「翻譯者」才是真正的 MVP #在當代產品開發中，最有價值的能力不僅是深度專精，而是跨領域的流暢溝通力。\n真正的超能力在於：\n• 將「商業意圖」轉譯為「技術策略」，\n• 並將「技術限制」轉化為「產品取捨」。\n這種雙語能力能釋放：\n速度：減少誤解，加速迭代 信任：建立跨職能間的相互尊重 創新：在「可行性」與「渴望度」的交會處，催生創造性解決方案 Cultivating Hybrid Thinkers #If you lead a cross-functional team, ask yourself:\nAre we fostering a culture where:\n• Engineers think like product owners—considering user impact and business context?\n• PMs speak the language of systems—understanding scalability, tech debt, and architectural trade-offs?\nThe future of product excellence doesn’t belong to “engineers vs. PMs.”\nIt belongs to hybrid thinkers—those who see the whole picture, bridge the gaps, and transform tension into synergy.\nTrue innovation lives not in silos, but in the space between.\n培養跨界思維者 #如果你帶領一支跨職能團隊，請自問：\n我們是否正在打造這樣的文化：\n• 工程師能像產品負責人一樣思考——關注用戶影響與商業脈絡？\n• 產品經理能說系統的語言——理解可擴展性、技術債與架構取捨？\n卓越產品的未來，不屬於「工程師對上產品經理」的對立，\n而屬於跨界思維者——那些能看見全局、填補鴻溝，並將張力轉化為協同效應的人。\n真正的創新，從不誕生於孤島之中，而是滋長於領域之間的縫隙。\n© Chung-Hao Lee. All Rights Reserved. All content on this webpage—including but not limited to text, images, design, code, and multimedia materials—is protected under the international copyright treaties. Unauthorized reproduction, modification, distribution, public transmission, or commercial use is strictly prohibited. Legal action will be taken against infringement. © 李崇豪。保留所有權利。 本網頁之內容（包括但不限於文字、圖片、設計、程式碼及多媒體素材）均受國際著作權條約保護。未經書面授權，嚴禁任何形式之複製、改作、散布、公開傳輸或商業利用。侵權者將依法追訴。\n","date":"October 6, 2025","permalink":"https://chunghaolee.com/posts/2025-10-07_beyond-the-divide-the-two-sides-of-the-same-coin-/","section":"Posts","summary":"\u003cp\u003e\u003cem\u003eI’ve had the privilege of working as both an engineer and a product manager—and I’m often asked: “Aren’t those roles totally different?” My answer might surprise you: they’re not opposites. They’re complementary lenses on the same mission.\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003e我有幸同時擔任過工程師與產品經理，常有人問我：「這兩個角色難道不是天差地遠嗎？」我的答案或許出乎意料：它們並非對立，而是同一使命下的互補視角。\u003c/p\u003e\n\u003chr\u003e\n\u003ch2 id=\"surface-differences-shared-purpose\" class=\"relative group\"\u003eSurface Differences, Shared Purpose \u003cspan class=\"absolute top-0 w-6 transition-opacity opacity-0 -start-6 not-prose group-hover:opacity-100\"\u003e\u003ca class=\"group-hover:text-primary-300 dark:group-hover:text-neutral-700\" style=\"text-decoration-line: none !important;\" href=\"#surface-differences-shared-purpose\" aria-label=\"Anchor\"\u003e#\u003c/a\u003e\u003c/span\u003e\u003c/h2\u003e\u003cp\u003eAt first glance, the roles appear distinct:\u003c/p\u003e\n\u003cul\u003e\n\u003cli\u003e\u003cstrong\u003eEngineers\u003c/strong\u003e focus on architecture, code, and system performance—turning constraints into feasible, scalable solutions.\u003c/li\u003e\n\u003cli\u003e\u003cstrong\u003eProduct Managers (PMs)\u003c/strong\u003e focus on user value, product strategy, and cross-team alignment—translating customer needs into prioritized outcomes.\u003c/li\u003e\n\u003c/ul\u003e\n\u003cp\u003eYet beneath these surface distinctions lies a unified core: \u003cstrong\u003eboth roles exist to solve problems\u003c/strong\u003e. They simply begin from different vantage points.\u003c/p\u003e","title":"Beyond the Divide: The Two Sides of the Same Coin"},{"content":"","date":"September 30, 2025","permalink":"https://chunghaolee.com/posts/2025-09-30_external-article---how-we-generated-millions-of-content-annotations/","section":"Posts","summary":"","title":"[External Article] - How We Generated Millions of Content Annotations (Spotify)"},{"content":"","date":"September 26, 2025","permalink":"https://chunghaolee.com/posts/2025-09-26_external-article---why-instagram-is-losing-gen-z-we-asked-100-users-to-compare-tiktok-vs.-reels/","section":"Posts","summary":"","title":"[External Article] - Why Instagram is Losing Gen Z: We Asked 100 Users to Compare TikTok vs. Reels"},{"content":"","date":"September 26, 2025","permalink":"https://chunghaolee.com/posts/2025-09-26_external-article---how-tiktok-is-evolving-the-next-generation-of-search/","section":"Posts","summary":"","title":"[External Article] - How TikTok is Evolving the Next Generation of Search"},{"content":" It’s easy to get swept up in the daily flood of AI announcements and assume that artificial intelligence is the ultimate solution to everything. But here’s the quiet truth no one’s shouting about: AI is not the destination. It’s the vehicle.\nI strongly resonate with Jack Ma’s perspective:\n“AI will change everything—but that doesn’t mean AI decides everything. Technology matters, yes. But what truly determines who wins in the future is deep, nuanced understanding of real-world domain needs.”\nWhat truly sets us apart isn’t the tool—it’s the insight. The real value lies in our grasp of domain-specific knowledge, our ability to discern critical needs, and to translate understanding into meaningful solutions. AI is an amplifier: a powerful means to achieve goals more efficiently, accurately, and intelligently.\n我們很容易被每日如潮水般湧來的 AI 新聞淹沒，進而誤以為人工智慧是萬能解方。但有個少有人高聲疾呼的真相：AI 不是目的地，而是載具。\n我非常認同馬雲的觀點：\n「AI 將改變一切——但這不代表 AI 決定一切。科技當然重要，但真正決定誰能在未來勝出的，是對現實世界領域需求的深刻且細膩的理解。」\n真正讓我們脫穎而出的，從來不是工具本身，而是洞察力。真正的價值，來自我們對特定領域知識的掌握、辨識關鍵需求的能力，以及將理解轉化為有意義解決方案的本領。AI 只是放大器——一個讓我們更高效、精準且聰明地達成目標的強大手段。\n🚀 Domain Knowledge is Your Superpower (AI Can’t Replicate It… Yet) #I’ve lived this principle firsthand. In a previous role, I once led end-to-end electronics manufacturing—from incoming quality control through SMT, FATP, to final shipping. The challenge? Designing foolproof monitoring systems to prevent human error at scale. I used poka-yoke (mistake-proofing), layered checkpoints, and intelligent resource allocation.\nLater, when I transitioned to managing massive volumes of content for anomaly detection.\nOn the surface, manufacturing hardware and curating content couldn\u0026rsquo;t be more different. Yet, the core thinking was identical:\nHow do you design a workflow to triage massive volumes? How do you allocate resources to the most critical items? How do you build in poka-yoke systems to prevent human error and oversight? This is the skill AI currently lacks: the creative leap of applying a lesson from a factory floor to a digital workflow. It\u0026rsquo;s the ability to recognize a similar problem structure in a different disguise.\n🚀 領域知識才是你的超能力（AI 目前還無法複製） #我親身實踐過這項原則。在先前職涯中，我曾主導電子產品製造全流程——從進料品質檢驗、SMT（表面貼裝技術）、FATP（最終組裝、測試與包裝）一直到出貨。當時的挑戰？設計一套萬無一失的監控系統，在大規模生產中防止人為疏失。我運用防呆設計（poka-yoke）、多重檢查點，以及智能化的資源配置策略。\n後來，當我轉向管理海量數位內容的異常偵測工作時，表面看來，硬體製造與內容審核天差地遠。然而，核心思維卻如出一轍：\n如何設計工作流程，以快速篩檢海量內容？ 如何把資源優先分配給最關鍵的項目？ 如何建置防呆機制，避免人為疏漏與監管盲區？ 這正是 AI 目前所缺乏的能力：將工廠現場學到的經驗，創造性地遷移到數位流程中的跨界聯想力。它體現在——能一眼看穿不同領域表象下，相似的問題結構。\n🎯 The Strategic Takeaway: Hybrid Talent Wins the Future #The message isn\u0026rsquo;t to dismiss AI, but to look beyond the hype.\nThe most valuable professionals in the AI era won’t be those who just know how to prompt a model or fine-tune a transformer.\nThey’ll be the ones who:\n→ Master their domain deeply\n→ Understand the operational, human, and business realities behind the data\n→ Leverage AI as a precision tool — not a magic wand\n→ See connections across disciplines and translate them into scalable solutions\n🎯 策略啟示：複合型人才將贏得未來 #這並非要我們否定 AI，而是提醒我們穿透 hype（炒作）看清本質。\n在 AI 時代最具價值的專業人士，不會只是懂得如何輸入提示詞或微調 Transformer 模型的人。\n真正引領未來的，會是那些：\n→ 深度掌握自身領域核心知識的人\n→ 理解數據背後的營運、人性與商業現實的人\n→ 將 AI 視為精密工具，而非神奇魔法棒的人\n→ 能跨領域察覺共通模式，並將其轉化為可擴展解決方案的人\nIn short: The future belongs to hybrid thinkers — those who speak both the language of their industry and the language of AI.\n簡言之：未來屬於「跨界思維者」——那些既能流利說出自身產業語言，也能與 AI 對話的人。\n© Chung-Hao Lee. All Rights Reserved. All content on this webpage—including but not limited to text, images, design, code, and multimedia materials—is protected under the international copyright treaties. Unauthorized reproduction, modification, distribution, public transmission, or commercial use is strictly prohibited. Legal action will be taken against infringement. © 李崇豪。保留所有權利。 本網頁之內容（包括但不限於文字、圖片、設計、程式碼及多媒體素材）均受國際著作權條約保護。未經書面授權，嚴禁任何形式之複製、改作、散布、公開傳輸或商業利用。侵權者將依法追訴。\n","date":"September 23, 2025","permalink":"https://chunghaolee.com/posts/2025-09-23_human-insight-+-ai-is-the-real-competitive-edge/","section":"Posts","summary":"\u003cp\u003e\n\n\n\n\n\n\u003cfigure\u003e\n    \n    \n\n\n\n\n\n\n\n\n  \n    \u003cpicture\n      class=\"mx-auto my-0 rounded-md\"\n      \n    \u003e\n      \n      \n      \n      \n        \u003csource\n          \n            srcset=\"https://chunghaolee.com/posts/2025-09-23_human-insight-\u0026#43;-ai-is-the-real-competitive-edge/pic_hu_be31722712353cf4.webp 330w,https://chunghaolee.com/posts/2025-09-23_human-insight-\u0026#43;-ai-is-the-real-competitive-edge/pic_hu_dab05edee73587ec.webp 660w\n            \n              ,https://chunghaolee.com/posts/2025-09-23_human-insight-\u0026#43;-ai-is-the-real-competitive-edge/pic_hu_b310c6cbb3e79568.webp 1024w\n            \n            \n              ,https://chunghaolee.com/posts/2025-09-23_human-insight-\u0026#43;-ai-is-the-real-competitive-edge/pic_hu_b7a2510bb64f7299.webp 1320w\n            \"\n          \n          sizes=\"100vw\"\n          type=\"image/webp\"\n        /\u003e\n      \n      \u003cimg\n        width=\"2336\"\n        height=\"1760\"\n        class=\"mx-auto my-0 rounded-md\"\n        \n        loading=\"lazy\" decoding=\"async\"\n        \n          src=\"https://chunghaolee.com/posts/2025-09-23_human-insight-\u0026#43;-ai-is-the-real-competitive-edge/pic_hu_1ba083f91f84e266.png\" srcset=\"https://chunghaolee.com/posts/2025-09-23_human-insight-\u0026#43;-ai-is-the-real-competitive-edge/pic_hu_504886e597068d6c.png 330w,https://chunghaolee.com/posts/2025-09-23_human-insight-\u0026#43;-ai-is-the-real-competitive-edge/pic_hu_1ba083f91f84e266.png 660w\n          \n            ,https://chunghaolee.com/posts/2025-09-23_human-insight-\u0026#43;-ai-is-the-real-competitive-edge/pic_hu_43dcf1713413b52b.png 1024w\n          \n          \n            ,https://chunghaolee.com/posts/2025-09-23_human-insight-\u0026#43;-ai-is-the-real-competitive-edge/pic_hu_c192b4a11165a73d.png 1320w\n          \"\n          sizes=\"100vw\"\n        \n      /\u003e\n    \u003c/picture\u003e\n  \n\n\n\u003c/figure\u003e\n\u003c/p\u003e\n\u003cp\u003eIt’s easy to get swept up in the daily flood of AI announcements and assume that artificial intelligence is the ultimate solution to everything. But here’s the quiet truth no one’s shouting about: AI is not the destination. It’s the vehicle.\u003c/p\u003e\n\u003cp\u003eI strongly resonate with Jack Ma’s perspective:\u003c/p\u003e\n\u003cblockquote\u003e\n\u003cp\u003e“AI will change everything—but that doesn’t mean AI decides everything. Technology matters, yes. But what truly determines who wins in the future is deep, nuanced understanding of real-world domain needs.”\u003c/p\u003e","title":"Beyond the Hype: Human Insight + AI is the Real Competitive Edge 🧠🚀"},{"content":" 📥 Download Portfolio PDF 📄 Full Portfolio Document # Your browser doesn't support PDF viewing. Please download the PDF to view it.\nKey Highlights #🇦🇺 As Australia positions itself to become a global AI superpower, the critical question emerges: where do we build the massive data centers that will power this digital revolution? For GovHack 2025, my latest analysis provides a data-driven framework to pinpoint the most strategic locations for Australia\u0026rsquo;s next wave of data infrastructure.\n📍 A Data-Driven Framework for Siting Australia\u0026rsquo;s Future Data Centers 📍 #My analysis identifies and scores Australia’s major regions against five non-negotiable factors for data center success: Power, Cooling, Connectivity, Talent, and Land. Location with great potential? Brisbane and ACT emerge as the undisputed champions for immediate development, scoring a perfect and near-perfect 5 and 4.5 out of 5, respectively. Sydney and Melbourne, while slightly behind on power readiness, are hot on their heels with scores of 4, signaling massive future potential.\nPower Availability \u0026amp; Reliability: Data centers are energy hogs, consuming millions of kWh monthly. My framework, powered by AEMO grid data, reveals Brisbane, ACT, and Adelaide as current clean energy leaders, while Melbourne and Armidale show strong future potential. Without reliable, low-carbon power, even the best location fails.\nCooling \u0026amp; Thermal Management: AI generates scorching waste heat. My analysis cross-references BoM temperature data with water storage maps to identify ACT, Melbourne, Sydney, and Brisbane as the “Goldilocks zones”—regions with moderate summer heat and abundant water for efficient liquid cooling.\nConnectivity: Positioned between East Asia and the US, Australia’s destiny is to be a global data hub. My mapping of submarine cables and fiber networks shows Sydney, Brisbane, and Perth as current connectivity kings, while the ACT and Melbourne are primed to leapfrog with strategic investment.\nTalent \u0026amp; Labor Force: A data center is only as good as its engineers. Leveraging Tech Council data, my analysis confirms Sydney, Melbourne, Brisbane, and the ACT as the nation’s tech talent powerhouses, boasting not just large current workforces but explosive growth rates (up to 98% in Queensland) to sustain long-term operations.\nLand \u0026amp; Infrastructure: Forget dense CBDs and remote outback. My framework uses population density as a proxy to find the perfect suburban balance: large, affordable land parcels with mature utilities. The Sydney→ACT→Melbourne corridor and outer Brisbane are the clear winners.\n🤖 AI as My Strategic Co-Pilot 🤖 #This entire framework was accelerated by AI, which helped synthesize complex geospatial, meteorological, and labor market datasets into a clear, actionable scoring system. But the real magic? Applying human domain expertise to weight the factors and interpret the nuances—because the best decisions come from human-AI partnership, not AI alone. In my next piece, I’ll break down exactly how I engineered this collaboration for maximum strategic impact.\n🔗 Let’s connect and explore how data-driven site selection can turn Australia into the world’s next AI superpower!\nFor more details, please visit GovHack page https://hackerspace.govhack.org/projects/australia_s_ai_future_a_data_driven_framework_for_data_center_siting\n© Chung-Hao Lee. All Rights Reserved. All content on this webpage—including but not limited to text, images, design, code, and multimedia materials—is protected under the international copyright treaties. Unauthorized reproduction, modification, distribution, public transmission, or commercial use is strictly prohibited. Legal action will be taken against infringement. © 李崇豪。保留所有權利。 本網頁之內容（包括但不限於文字、圖片、設計、程式碼及多媒體素材）均受國際著作權條約保護。未經書面授權，嚴禁任何形式之複製、改作、散布、公開傳輸或商業利用。侵權者將依法追訴。\n","date":"August 30, 2025","permalink":"https://chunghaolee.com/portfolio/govhack2025---data-center-challenging/","section":"Portfolios","summary":"\u003cdiv style=\"text-align: center; margin: 20px 0;\"\u003e\n    \u003ca href=\"GovHack 2025_ Australia's AI Future - A Data-Driven Framework for Data Center Siting.pdf\" download class=\"download-button\" style=\"display: inline-block; padding: 12px 24px; background-color: #007bff; color: white; text-decoration: none; border-radius: 5px; font-weight: bold; transition: background-color 0.3s;\"\u003e\n        📥 Download Portfolio PDF\n    \u003c/a\u003e\n\u003c/div\u003e\n\u003cstyle\u003e\n.download-button:hover {\n    background-color: #0056b3 !important;\n}\n\n.pdf-container {\n    width: 100%;\n    max-width: 900px;\n    margin: 0 auto;\n    padding: 20px 0;\n}\n\n.pdf-embed {\n    width: 100%;\n    height: 1200px;\n    border: 1px solid #ddd;\n    border-radius: 8px;\n    box-shadow: 0 2px 10px rgba(0,0,0,0.1);\n    margin-bottom: 30px;\n}\n\n.pdf-page-image {\n    width: 100%;\n    margin-bottom: 20px;\n    border: 1px solid #e0e0e0;\n    border-radius: 8px;\n    box-shadow: 0 2px 8px rgba(0,0,0,0.08);\n}\n\n@media (max-width: 768px) {\n    .pdf-embed {\n        height: 600px;\n    }\n}\n\u003c/style\u003e\n\u003cdiv class=\"pdf-container\"\u003e\n\u003ch3 id=\"-full-portfolio-document\" class=\"relative group\"\u003e📄 Full Portfolio Document \u003cspan class=\"absolute top-0 w-6 transition-opacity opacity-0 -start-6 not-prose group-hover:opacity-100\"\u003e\u003ca class=\"group-hover:text-primary-300 dark:group-hover:text-neutral-700\" style=\"text-decoration-line: none !important;\" href=\"#-full-portfolio-document\" aria-label=\"Anchor\"\u003e#\u003c/a\u003e\u003c/span\u003e\u003c/h3\u003e\u003c!-- Option 1: Using iframe embed (works if PDF is in static folder) --\u003e\n\u003ciframe src=\"GovHack 2025_ Australia's AI Future - A Data-Driven Framework for Data Center Siting.pdf\" class=\"pdf-embed\" type=\"application/pdf\"\u003e\n    \u003cp\u003eYour browser doesn't support PDF viewing. Please \u003ca href=\"GovHack 2025_ Australia's AI Future - A Data-Driven Framework for Data Center Siting.pdf\"\u003edownload the PDF\u003c/a\u003e to view it.\u003c/p\u003e","title":"GovHack2025 - Data Center Challenging"},{"content":"✨ Honored to take part in #GovHack2025 this weekend. Huge thanks to Vikash Kumar for hosting and guiding the event 🙏\nThe atmosphere was fantastic: working side by side with many teams under tight deadlines, brainstorming solutions, and pushing through to solve real challenges was truly inspiring 🚀\nHaving recently relocated from Taiwan to Sydney 🇦🇺, I’m passionate about using data to analyze problems and drive practical solutions. GovHack offered this great opportunity for me to apply those skills, contribute to the Australian community, and learn from others who share the same curiosity and drive\nI chose to tackle the “data center” challenge. As Scott Farquhar recently noted at the National Press Club, Australia has a unique opportunity to become a leader in the data center space, with advantages like abundant resources, a reliable power supply, a skilled workforce, and a strategic location between East Asia and the USA.⚡️🌏 However, a key question remains: how do we select the best locations, and what is the best way to evaluate them? There is currently no clear solution.\nIt was an honor to work on this challenge, and over two intensive days the analytical work may help advance Australia’s digital infrastructure and support the broader community’s technological needs 💡📊\n© Chung-Hao Lee. All Rights Reserved. All content on this webpage—including but not limited to text, images, design, code, and multimedia materials—is protected under the international copyright treaties. Unauthorized reproduction, modification, distribution, public transmission, or commercial use is strictly prohibited. Legal action will be taken against infringement. © 李崇豪。保留所有權利。 本網頁之內容（包括但不限於文字、圖片、設計、程式碼及多媒體素材）均受國際著作權條約保護。未經書面授權，嚴禁任何形式之複製、改作、散布、公開傳輸或商業利用。侵權者將依法追訴。\n","date":"August 30, 2025","permalink":"https://chunghaolee.com/posts/2025-09-23_govhack2025-participating/","section":"Posts","summary":"\u003cp\u003e\u003cstrong\u003e✨ Honored to take part in #GovHack2025 this weekend. Huge thanks to Vikash Kumar for hosting and guiding the event 🙏\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe atmosphere was fantastic: working side by side with many teams under tight deadlines, brainstorming solutions, and pushing through to solve real challenges was truly inspiring 🚀\u003c/p\u003e\n\u003cp\u003eHaving recently relocated from Taiwan to Sydney 🇦🇺, I’m passionate about using data to analyze problems and drive practical solutions. GovHack offered this great opportunity for me to apply those skills, contribute to the Australian community, and learn from others who share the same curiosity and drive\u003c/p\u003e","title":"GovHack2025 participating"},{"content":"AI’s purpose isn’t to impress with flashy technology—it’s to solve real business problems 🎯 #In today’s AI-driven landscape, I’ve learned that the most impactful implementations aren’t always the most sophisticated ones. The key lies in understanding genuine business pain points and selecting the most appropriate solution, rather than chasing the latest trends that may look impressive but fail to deliver results.\nAI 的價值不在炫技，而在解決真實的商業問題 🎯 #在當今由 AI 驅動的環境中，我體會到：最具影響力的 AI 實踐，往往並非技術最複雜的那一種。關鍵在於深入理解真實的業務痛點，並選擇最恰當的解決方案，而非盲目追逐看似亮眼卻難以落地的技術潮流。\nA Real Example: #I learned this firsthand during my research at the University of Maryland, where I worked on a project to predict the likelihood of minor league baseball players making it to the MLB. Instead of a complex deep learning model, I used XGBoost. Why? Because it was the right tool for the job. It was efficient, effective, and provided the clarity I needed.\nThe validation of my approach came through tangible results: publication in Springer’s Analytics Enabled Decision Making and coverage by Maryland Today, the University of Maryland’s official media, demonstrating that solving real problems garners genuine recognition.\n一個真實案例： #我在馬里蘭大學進行研究時親身驗證了這一點——當時我負責一個預測小聯盟棒球選手晉升大聯盟（MLB）機率的專案。我沒有選擇複雜的深度學習模型，而是採用 XGBoost。原因很簡單：它才是最適合這項任務的工具——高效、有效，且結果清晰易解。\n這套方法的價值，最終透過具體成果得到驗證：研究成果獲刊登於 Springer 出版的《Analytics Enabled Decision Making》專書，並獲馬里蘭大學官方媒體 Maryland Today 報導。這證明：真正解決問題的方案，才會贏得實質的認可。\nHere’s how I approach AI work in practice: # Problem first → tool second\nDefine the business outcome (revenue, time-to-value, risk reduction). Only then pick the method that best and most efficiently achieves that outcome. Fit-for-purpose modeling\nSometimes a well-tuned XGBoost or classical model beats a large deep network because it’s faster to train, easier to explain, and produces better performance on the true objective. Don’t confuse novelty with value. Measure real impact, not model glamour\nValidate against business KPIs, production constraints, and operational readiness — not just flashy demo metrics. 我在實務中如何應用 AI： # 先釐清問題，再選擇工具\n先明確定義商業目標（例如提升收入、縮短實現價值的時間、降低風險），再挑選最有效且高效達成目標的方法。 以任務為導向的建模思維\n有時，一個調校得當的 XGBoost 或傳統模型，反而比龐大的深度神經網路更勝一籌——因為它訓練更快、解釋性更強，且在真實目標上表現更佳。切勿將「新穎」誤認為「價值」。 衡量實際影響，而非模型光環\n驗證標準應是商業 KPI、生產環境限制與營運就緒度，而非僅僅炫目的展示指標。 The Strategic Takeaway: #When implementing AI solutions, resist the temptation to deploy the most cutting-edge technology simply because it exists. Instead, focus on understanding the specific business challenge, evaluating available approaches objectively, and selecting the method that delivers the most reliable and actionable results.\n策略啟示： #導入 AI 解決方案時，務必克制「只因技術存在就非用不可」的衝動。與其追求技術前沿，不如專注於理解具體的業務挑戰，客觀評估可行方案，並選擇能穩定產出可執行成果的方法。\n\u0026ldquo;Traditional approaches may overlook these nuances, but those who blend domain expertise with targeted AI will lead the game.\u0026rdquo;\n「傳統做法常忽略這些細微卻關鍵的差異，唯有將領域專業與精準的 AI 應用相結合的人，才能在競爭中領先群倫。」\n© Chung-Hao Lee. All Rights Reserved. All content on this webpage—including but not limited to text, images, design, code, and multimedia materials—is protected under the international copyright treaties. Unauthorized reproduction, modification, distribution, public transmission, or commercial use is strictly prohibited. Legal action will be taken against infringement. © 李崇豪。保留所有權利。 本網頁之內容（包括但不限於文字、圖片、設計、程式碼及多媒體素材）均受國際著作權條約保護。未經書面授權，嚴禁任何形式之複製、改作、散布、公開傳輸或商業利用。侵權者將依法追訴。\n","date":"August 15, 2025","permalink":"https://chunghaolee.com/posts/2025-08-15_ai_real_power_lies_in_solving_problems_not_just_dazzling_with_tech/","section":"Posts","summary":"\u003ch3 id=\"ais-purpose-isnt-to-impress-with-flashy-technologyits-to-solve-real-business-problems-\" class=\"relative group\"\u003eAI’s purpose isn’t to impress with flashy technology—it’s to solve real business problems 🎯 \u003cspan class=\"absolute top-0 w-6 transition-opacity opacity-0 -start-6 not-prose group-hover:opacity-100\"\u003e\u003ca class=\"group-hover:text-primary-300 dark:group-hover:text-neutral-700\" style=\"text-decoration-line: none !important;\" href=\"#ais-purpose-isnt-to-impress-with-flashy-technologyits-to-solve-real-business-problems-\" aria-label=\"Anchor\"\u003e#\u003c/a\u003e\u003c/span\u003e\u003c/h3\u003e\u003cp\u003eIn today’s AI-driven landscape, I’ve learned that the most impactful implementations aren’t always the most sophisticated ones. The key lies in understanding genuine business pain points and selecting the most appropriate solution, rather than chasing the latest trends that may look impressive but fail to deliver results.\u003c/p\u003e","title":"Beyond the Numbers: AI’s Real Power Lies in Solving Problems, Not Just Dazzling with Tech🚀"},{"content":"In today’s fast‑moving business environment, AI-driven data analytics isn’t about replacing your expertise—it’s about amplifying it. I treat AI as a high‑performance co‑pilot: I set the destination, define the route, and ensure we stay on course, while AI accelerates every step of the journey.\n在當今快速變動的商業環境中，AI驅動的數據分析並非要取代你的專業，而是要放大它。我將AI視為一位高效能的副駕駛：由我設定目的地、規劃路線，並確保航向正確，而AI則加速推進旅程中的每一步。\nHere’s how we collaborate to drive meaningful insights and valuable outputs: # Topic Selection: When I have a broad objective but need to pinpoint the right analysis, I brief AI on the direction I’m aiming for. It suggests potential angles—and then I apply my judgment to choose the most promising one.\nData Gathering: Once the topic is locked in, I craft precise prompts and leverage AI’s research capabilities to surface relevant datasets, reports, and statistics.\nReport Drafting: With a clear outline in hand, I direct AI to generate a first draft of the analysis, ensuring structure and flow match my goals.\n我們如何協作以產出有意義的洞見與高價值成果： # 議題篩選：當我有宏觀目標但尚需聚焦分析切入點時，我會向AI說明方向；它會提出潛在角度，再由我運用專業判斷，選出最具潛力的路徑。\n資料蒐集：議題確定後，我設計精準提示（prompt），善用AI的研究能力，快速挖掘相關數據集、報告與統計資料。\n報告草擬：手握清晰大綱後，我指示AI產出初步分析草稿，確保內容架構與邏輯流暢度符合我的策略目標。\nWhere human expertise takes the helm: # Prompt Engineering: I translate objectives into crystal‑clear instructions, iterating until AI fully grasps the goal.\nValidation \u0026amp; Quality Control: Every chart, figure, and data source is vetted by me to guarantee reliability and integrity.\nDomain Insight: When AI’s output drifts off course, I apply my domain knowledge to realign the analysis and refine the narrative.\n關鍵決策仍由人類主導： # 提示工程：我將商業目標轉化為清晰明確的指令，並反覆調整，直到AI完全掌握任務核心。\n驗證與品質控管：每張圖表、每個數字、每項資料來源，都經我親自審核，確保其可靠性與資料誠信。\n領域洞察：當AI產出偏離現實脈絡時，我運用專業知識修正分析方向，並打磨敘事邏輯。\nThe Captain and the Crew #I navigate; AI executes. This partnership delivers insights faster and with greater depth—yet always under a human safeguard. By combining strategic vision, rigorous validation, and AI’s computational power, we chart a path to data‑driven decisions that move business forward.\n船長與船員 #我掌舵，AI執行。這種夥伴關係讓洞見產出更快、更深，卻始終在人類的把關之下。透過融合策略視野、嚴謹驗證與AI的運算能力，我們共同開拓一條通往真正驅動商業前進的數據決策之路。\n\u0026ldquo;Traditional data analysts may find themselves left behind—but those who master AI collaboration and bring deep domain expertise will thrive in tomorrow’s market.\u0026rdquo;\n「傳統的數據分析師可能面臨被淘汰的風險——但那些能掌握AI協作、並帶有深厚領域專業的人，將在明日市場中脫穎而出。」\n© Chung-Hao Lee. All Rights Reserved. All content on this webpage—including but not limited to text, images, design, code, and multimedia materials—is protected under the international copyright treaties. Unauthorized reproduction, modification, distribution, public transmission, or commercial use is strictly prohibited. Legal action will be taken against infringement. © 李崇豪。保留所有權利。 本網頁之內容（包括但不限於文字、圖片、設計、程式碼及多媒體素材）均受國際著作權條約保護。未經書面授權，嚴禁任何形式之複製、改作、散布、公開傳輸或商業利用。侵權者將依法追訴。\n","date":"August 1, 2025","permalink":"https://chunghaolee.com/posts/2025-08-01_ai_serves_as_my_analytics_copilot_steering_data_not_just_processing_it/","section":"Posts","summary":"\u003cp\u003eIn today’s fast‑moving business environment, AI-driven data analytics isn’t about replacing your expertise—it’s about amplifying it. I treat AI as a high‑performance co‑pilot: I set the destination, define the route, and ensure we stay on course, while AI accelerates every step of the journey.\u003c/p\u003e\n\u003cbr\u003e\n\u003cp\u003e在當今快速變動的商業環境中，AI驅動的數據分析並非要取代你的專業，而是要放大它。我將AI視為一位高效能的副駕駛：由我設定目的地、規劃路線，並確保航向正確，而AI則加速推進旅程中的每一步。\u003c/p\u003e\n\u003chr\u003e\n\u003ch3 id=\"heres-how-we-collaborate-to-drive-meaningful-insights-and-valuable-outputs\" class=\"relative group\"\u003eHere’s how we collaborate to drive meaningful insights and valuable outputs: \u003cspan class=\"absolute top-0 w-6 transition-opacity opacity-0 -start-6 not-prose group-hover:opacity-100\"\u003e\u003ca class=\"group-hover:text-primary-300 dark:group-hover:text-neutral-700\" style=\"text-decoration-line: none !important;\" href=\"#heres-how-we-collaborate-to-drive-meaningful-insights-and-valuable-outputs\" aria-label=\"Anchor\"\u003e#\u003c/a\u003e\u003c/span\u003e\u003c/h3\u003e\u003cul\u003e\n\u003cli\u003e\n\u003cp\u003e\u003cstrong\u003eTopic Selection:\u003c/strong\u003e When I have a broad objective but need to pinpoint the right analysis, I brief AI on the direction I’m aiming for. It suggests potential angles—and then I apply my judgment to choose the most promising one.\u003c/p\u003e","title":"Beyond the Numbers: AI serves as my analytics co‑pilot—steering data, not just processing it 🚢"},{"content":" 📥 Download Portfolio PDF 📄 Full Portfolio Document # Your browser doesn't support PDF viewing. Please download the PDF to view it.\nKey Highlights #🚨 As the countdown to the Brisbane 2032 Olympic Games intensifies, so does the critical question: how will Queensland\u0026rsquo;s construction industry meet the unprecedented workforce demands of building world-class venues? As a Business Data Analyst, my latest analysis dives deep into this challenge, revealing the significant workforce requirements that lie ahead and offering data-driven insights to inform strategic business decisions.\n🏗️ Workforce Analysis: Brisbane 2032 Olympic Venue Construction 🏗️ #My recent analysis highlights the significant workforce demands for the Brisbane 2032 Olympic Games, focusing specifically on the construction of its 20 major venues. My findings indicate a projected demand of approximately 43,000 full-time equivalent workers over the construction timeline. This translates to an average annual requirement of 6,100 additional workers specifically for the Olympic venues.\nIntensifying Labor Shortfall: Projected need of ~43,000 full‑time equivalent (FTE) workers to complete both new builds and upgrades across these 20 sites. An average requirement of ~6,100 additional workers per year, intensifying competition in an already stretched labor market\nCascading Risk Profile: Labor shortages create compounding effects—schedule delays, cost overruns from wage inflation, potential quality compromises, and heightened safety concerns across all 20 venues.\nMitigation Imperatives: Addressing this challenge requires a multi-pronged approach, including accelerated vocational training, streamlined skilled migration, mandated adoption of productivity-enhancing technologies (e.g., BIM, prefabrication), and improved industry coordination.\nAI as My Analytical Co-Pilot 🤖 #As AI’s transformative power sweeps through every industry, including construction and energy. This study also harnesses AI capabilities to boost research efficiency. Yet true insight comes not from AI alone but from marrying advanced algorithms with rigorous personal insights and deep domain expertise. In my next article, I will outline how to partner with AI effectively to deliver analyses that are both faster and more robust.\n🔗 Let’s connect and explore how data-driven strategies can transform business outcomes!\n© Chung-Hao Lee. All Rights Reserved. All content on this webpage—including but not limited to text, images, design, code, and multimedia materials—is protected under the international copyright treaties. Unauthorized reproduction, modification, distribution, public transmission, or commercial use is strictly prohibited. Legal action will be taken against infringement. © 李崇豪。保留所有權利。 本網頁之內容（包括但不限於文字、圖片、設計、程式碼及多媒體素材）均受國際著作權條約保護。未經書面授權，嚴禁任何形式之複製、改作、散布、公開傳輸或商業利用。侵權者將依法追訴。\n","date":"July 20, 2025","permalink":"https://chunghaolee.com/portfolio/brisbane-2032-olympics-workforce-demand-analysis/","section":"Portfolios","summary":"\u003cdiv style=\"text-align: center; margin: 20px 0;\"\u003e\n    \u003ca href=\"Brisbane 2032 Olympics Workforce Demand Analysis.pdf\" download class=\"download-button\" style=\"display: inline-block; padding: 12px 24px; background-color: #007bff; color: white; text-decoration: none; border-radius: 5px; font-weight: bold; transition: background-color 0.3s;\"\u003e\n        📥 Download Portfolio PDF\n    \u003c/a\u003e\n\u003c/div\u003e\n\u003cstyle\u003e\n.download-button:hover {\n    background-color: #0056b3 !important;\n}\n\n.pdf-container {\n    width: 100%;\n    max-width: 900px;\n    margin: 0 auto;\n    padding: 20px 0;\n}\n\n.pdf-embed {\n    width: 100%;\n    height: 600px;\n    border: 1px solid #ddd;\n    border-radius: 8px;\n    box-shadow: 0 2px 10px rgba(0,0,0,0.1);\n    margin-bottom: 30px;\n}\n\n.pdf-page-image {\n    width: 100%;\n    margin-bottom: 20px;\n    border: 1px solid #e0e0e0;\n    border-radius: 8px;\n    box-shadow: 0 2px 8px rgba(0,0,0,0.08);\n}\n\n@media (max-width: 768px) {\n    .pdf-embed {\n        height: 600px;\n    }\n}\n\u003c/style\u003e\n\u003cdiv class=\"pdf-container\"\u003e\n\u003ch3 id=\"-full-portfolio-document\" class=\"relative group\"\u003e📄 Full Portfolio Document \u003cspan class=\"absolute top-0 w-6 transition-opacity opacity-0 -start-6 not-prose group-hover:opacity-100\"\u003e\u003ca class=\"group-hover:text-primary-300 dark:group-hover:text-neutral-700\" style=\"text-decoration-line: none !important;\" href=\"#-full-portfolio-document\" aria-label=\"Anchor\"\u003e#\u003c/a\u003e\u003c/span\u003e\u003c/h3\u003e\u003c!-- Option 1: Using iframe embed (works if PDF is in static folder) --\u003e\n\u003ciframe src=\"Brisbane 2032 Olympics Workforce Demand Analysis.pdf\" class=\"pdf-embed\" type=\"application/pdf\"\u003e\n    \u003cp\u003eYour browser doesn't support PDF viewing. Please \u003ca href=\"Brisbane 2032 Olympics Workforce Demand Analysis.pdf\"\u003edownload the PDF\u003c/a\u003e to view it.\u003c/p\u003e","title":"Brisbane 2032 Olympics Workforce Demand Analysis"},{"content":"In my journey as a PM within AI and technology sectors, I’ve consistently operated at the nexus of execution and strategy—delivering on objectives, aligning multidisciplinary teams, and driving tangible business impact. Yet, behind every milestone lies the same fundamental question:\u2028What hidden insights does the data hold—and how can they inform smarter decisions?\nI’ve come to realize that a good business decision isn’t propelled by data alone, but by the insights uncovering the “why.” While mastering SQL, statistics, and dashboards is indispensable, the real differentiator in business data analytics is the ability to translate raw numbers into strategic insights.\n在我擔任人工智慧與科技領域產品經理的歷程中，我始終處於執行與策略的交會點——不僅達成目標、協調跨領域團隊，更推動具體的商業成果。然而，每一個里程碑背後，總存在同一個根本問題：\n數據究竟藏著哪些未被察覺的洞見？我們又該如何運用這些洞見做出更聰明的決策？\n我逐漸體悟到：一項優質的商業決策，從不只是靠數據本身驅動，而是靠揭開「為什麼」的深度洞察。熟練 SQL、統計學與儀表板固然不可或缺，但真正讓商業數據分析脫穎而出的關鍵，在於能否將原始數字轉化為具策略價值的見解。\n🔍 Here’s an example that cemented this perspective: #During my MBA, I conducted a case study comparing China’s two bike‑sharing giants: Mobike and ofo.\nOn paper, both were burning VC capital at breakneck speed, deploying massive fleets, and chasing similar KPIs. Yet Mobike thrived while ofo exited the market. What explained their divergent destinies?\nAlthough data analytics can dissect supply‑chain efficiency, marketing ROI, and operational performance, none of those metrics ultimately decided the outcome. The real inflection point lay not in the spreadsheets but in user‑experience design.\nSpecifically, Mobike’s strategic move to integrate directly with WeChat gave it a game-changing edge:\n📲 Channel: In China, WeChat is the digital backbone of daily life. Embedding Mobike directly into WeChat opened the door to its 1.4 billion users.\n⏳ Seamless Onboarding: No separate app download or lengthy registration—users could unlock a bike with a single tap inside WeChat.\n☝ One‑Click Payments: Leveraging WeChat Pay eliminated the friction of bank transfers and credit‑card sign‑ups that bogged down ofo.\n↗️ Outcome: Lower adoption barriers, stronger retention rates, and rapid market penetration.\n🔍 以下案例讓我確信這個觀點： #在攻讀MBA期間，我曾深入研究中國兩大共享單車巨頭——摩拜（Mobike）與 ofo 的競爭局勢。\n表面上看，兩家公司都以驚人速度消耗創投資金、大量投放車輛、追逐相似的關鍵績效指標。然而，摩拜最終蓬勃發展，ofo 卻黯然退場。究竟是什麼造就了截然不同的命運？\n雖然數據分析能拆解供應鏈效率、行銷投資報酬率與營運表現，但這些指標最終並未決定成敗。真正的轉折點不在試算表裡，而在使用者體驗設計上。\n尤其關鍵的是，摩拜選擇直接整合進微信的策略性舉動，為其帶來了顛覆性的優勢：\n📲 通路優勢：在中國，微信是數位生活的基礎建設。摩拜內嵌至微信，等於直接觸及14億用戶。\n⏳ 無縫開通體驗：無需額外下載App或冗長註冊流程——使用者在微信內單點即能解鎖單車。\n☝ 一鍵支付：透過微信支付，省去了銀行轉帳或信用卡綁定的繁瑣步驟，而這正是 ofo 用戶流失的痛點。\n↗️ 成效：大幅降低使用門檻、提升用戶留存率，並加速市場滲透。\nThis case reminded me that:\n👉 Data tells us what happened.\n👉 Insight—rooted in strategic vision and thinking—reveals why it mattered.\n這個案例讓我深刻體會到：\n👉 數據告訴我們「發生了什麼」。\n👉 洞察——奠基於策略視野與深度思考——才揭示「為何這件事如此重要」。\nToday, I combine a PM’s execution mindset with deep data‑analysis expertise to unearth actionable insights that drive smarter data-backed business decisions. From optimizing operations to pinpointing customer pain points or validating product‑market fit - I’m excited to bridge the gap between data and business strategy.\n如今，我融合產品經理的執行思維與扎實的數據分析能力，挖掘可付諸行動的洞察，推動更聰明、有數據支撐的商業決策。無論是優化營運、精準定位客戶痛點，或是驗證產品市場契合度——我都熱切期待填補數據與商業策略之間的鴻溝。\n© Chung-Hao Lee. All Rights Reserved. All content on this webpage—including but not limited to text, images, design, code, and multimedia materials—is protected under the international copyright treaties. Unauthorized reproduction, modification, distribution, public transmission, or commercial use is strictly prohibited. Legal action will be taken against infringement. © 李崇豪。保留所有權利。 本網頁之內容（包括但不限於文字、圖片、設計、程式碼及多媒體素材）均受國際著作權條約保護。未經書面授權，嚴禁任何形式之複製、改作、散布、公開傳輸或商業利用。侵權者將依法追訴。\n","date":"June 29, 2025","permalink":"https://chunghaolee.com/posts/2025-06-29_turning_data_into_insights_for_smarter_business_decisions/","section":"Posts","summary":"\u003cp\u003eIn my journey as a PM within AI and technology sectors, I’ve consistently operated at the nexus of execution and strategy—delivering on objectives, aligning multidisciplinary teams, and driving tangible business impact. Yet, behind every milestone lies the same fundamental question:\u2028What hidden insights does the data hold—and how can they inform smarter decisions?\u003c/p\u003e\n\u003cp\u003eI’ve come to realize that a good business decision isn’t propelled by data alone, but by the insights uncovering the “why.” While mastering SQL, statistics, and dashboards is indispensable, the real differentiator in business data analytics is the ability to translate raw numbers into strategic insights.\u003c/p\u003e","title":"Beyond the Numbers: Turning Data into Insights for Smarter Business Decisions"},{"content":"As a PM, I’ve navigated countless crossroads where data seemed to hold all the answers. But here’s the reality I’ve learned:\n📊 Data points are signposts, not destinations.\nWhile metrics powerfully guide product direction and feature prioritization, over-reliance on raw data can lead to costly missteps. True business impact emerges when quantitative insights fuse with domain expertise and contextual wisdom.\n身為產品經理，我曾歷經無數次十字路口，當時數據看似掌握所有答案。但這是我學到的真相：\n📊 數據只是路標，而非終點。\n雖然指標能有力指引產品方向與功能優先順序，但過度仰賴原始數據，反而會導致代價高昂的誤判。唯有當量化洞察與領域專業和情境智慧融合時，才能真正創造商業價值。\nA case study from my previous project #When Data Tells Half the Story: How Expert Judgment Transformed Our Predictive Model\nIn a project analyzing laptop logs to predict hardware failures, my team identified a key power-related log pattern strongly correlated with malfunctions. Initial models flagged this as a critical failure signal—until false positives skyrocketed.\n前一個專案的實例 #當數據只說了一半的故事：專家判斷如何徹底轉變我們的預測模型\n在一個分析筆電日誌以預測硬體故障的專案中，我們團隊發現某種與電源相關的日誌模式，與故障高度相關。初期模型將此視為關鍵失效信號——直到誤報率暴增才發現事有蹊蹺。\n🔍 The breakthrough came through collaboration: #Domain experts revealed this log triggered not just during failures, but also in normal low-battery states while users didn’t plug in chargers. The causality flip:\n→ Faulty devices led users to abandon charging (expecting imminent failure).\n→ This behavior massively amplified log generation during low-battery states.\n→ Raw data misinterpreted correlation as causation.\nBlind data reliance missed this nuance; human experience reframed the logic.\nResult: Revised models reduced false positives by contextualizing data with behavioral reality.\n🔍 突破來自跨領域協作： #領域專家指出，這類日誌不僅出現在故障時，也常在電量偏低但使用者未插電的正常情境下觸發。\n因果關係其實是反過來的：\n→ 有問題的設備讓使用者預期即將故障，乾脆放棄充電。\n→ 此行為大幅增加了低電量狀態下的日誌產出量。\n→ 原始數據誤將相關性當成因果關係。\n盲目信賴數據忽略了這層細微差異；人的經驗才重新釐清了背後邏輯。\n結果：修正後的模型結合行為脈絡解讀數據，大幅降低了誤報率。\nThe AI Era’s Unspoken Truth: #As AI democratizes analytics, polished dashboards and automated reports are table stakes. What separates strategic leaders?\n✅ Business acumen to ask why behind the numbers\n✅ Domain fluency to separate signals from noise\n✅ Judgment to balance data with real-world complexity\n人工智慧時代未被明說的真相： #當AI讓數據分析日益普及，精緻的儀表板與自動化報表已成基本配備。真正區隔戰略型領導者的關鍵在於：\n✅ 能洞察數字背後「為什麼」的商業直覺\n✅ 能在雜訊中辨識真正訊號的領域敏銳度\n✅ 能在數據與現實複雜性之間取得平衡的判斷力\n✨ Key Takeaway: #Data illuminate paths, but experience navigates them. The future belongs to professionals who wield analytics not just as technicians, but as translators of insights into actionable strategy.\n✨ 核心啟示： #數據照亮路徑，經驗引領方向。未來屬於那些不只是技術操作員、更能將洞察轉譯為具體策略的專業人才。\n© Chung-Hao Lee. All Rights Reserved. All content on this webpage—including but not limited to text, images, design, code, and multimedia materials—is protected under the international copyright treaties. Unauthorized reproduction, modification, distribution, public transmission, or commercial use is strictly prohibited. Legal action will be taken against infringement. © 李崇豪。保留所有權利。 本網頁之內容（包括但不限於文字、圖片、設計、程式碼及多媒體素材）均受國際著作權條約保護。未經書面授權，嚴禁任何形式之複製、改作、散布、公開傳輸或商業利用。侵權者將依法追訴。\n","date":"June 1, 2025","permalink":"https://chunghaolee.com/posts/2025-06-01_the_art_of_data-informed_decision_making/","section":"Posts","summary":"\u003cp\u003eAs a PM, I’ve navigated countless crossroads where data seemed to hold all the answers. But here’s the reality I’ve learned:\u003cbr\u003e\n📊 Data points are signposts, not destinations.\u003cbr\u003e\nWhile metrics powerfully guide product direction and feature prioritization, over-reliance on raw data can lead to costly missteps. True business impact emerges when quantitative insights fuse with domain expertise and contextual wisdom.\u003c/p\u003e\n\u003cbr\u003e\n\u003cp\u003e身為產品經理，我曾歷經無數次十字路口，當時數據看似掌握所有答案。但這是我學到的真相：\u003cbr\u003e\n📊 數據只是路標，而非終點。\u003cbr\u003e\n雖然指標能有力指引產品方向與功能優先順序，但過度仰賴原始數據，反而會導致代價高昂的誤判。唯有當量化洞察與領域專業和情境智慧融合時，才能真正創造商業價值。\u003c/p\u003e\n\u003chr\u003e\n\u003ch2 id=\"a-case-study-from-my-previous-project\" class=\"relative group\"\u003eA case study from my previous project \u003cspan class=\"absolute top-0 w-6 transition-opacity opacity-0 -start-6 not-prose group-hover:opacity-100\"\u003e\u003ca class=\"group-hover:text-primary-300 dark:group-hover:text-neutral-700\" style=\"text-decoration-line: none !important;\" href=\"#a-case-study-from-my-previous-project\" aria-label=\"Anchor\"\u003e#\u003c/a\u003e\u003c/span\u003e\u003c/h2\u003e\u003cp\u003eWhen Data Tells Half the Story: How Expert Judgment Transformed Our Predictive Model\u003cbr\u003e\nIn a project analyzing laptop logs to predict hardware failures, my team identified a key power-related log pattern strongly correlated with malfunctions. Initial models flagged this as a critical failure signal—until false positives skyrocketed.\u003c/p\u003e","title":"Beyond the Numbers: The Art of Data-Informed Decision Making"},{"content":"Queensland Aluminium Industry Emissions: 2024 Status, 2030 Reduction Targets and Decarbonization Roadmap\n📥 Download Portfolio PDF 📄 Full Portfolio Document # Your browser doesn't support PDF viewing. Please download the PDF to view it.\nKey Highlights #🇦🇺 Excited to have recently relocated from Taiwan to Australia, I bring over five years of business data analysis experience in tech, supply chain and academia, where I transformed complex datasets into actionable insights. Now in Australia, I remain passionate about leveraging data analysis to uncover actionable insights and translate findings into impactful business solutions. I am channeling this expertise to understand the dynamics of key sectors, with a particular interest in the mining and construction industries. Through my analyses, I aim to deepen my understanding of these two key sectors and welcome guidance from industry experts.\n🏭 Recent Analysis Highlight: Queensland’s Aluminium Industry 🏭 #Bridging a 3.8 Mt CO₂e gap by 2030 to hit a 50% reduction target is not a small feat. Key insights:\nTargeted Impact: Alumina refining (49%) and aluminium smelting (48%) dominate emissions—innovative decarbonization here offers the greatest leverage. Renewables as Foundation: Large‑scale renewable energy deployment is the essential enabler for all decarbonization technologies. Portfolio Strategy: No single silver bullet—MVR, electric boilers, hydrogen calcination and inert anodes must run in parallel to meet the goal. (Acknowledgements to Rio Tinto and Queensland Government for their data transparency, which formed the basis of this research.) In my analysis process, I’ve embraced AI as a powerful co‑pilot. 🤖 How? # Accelerated Data Gathering: AI expedites research and enriches brainstorming with diverse analytical perspectives. Human‑in‑the‑Loop Validation: I’ve learned that AI outputs demand meticulous human oversight—those numbers can be tricky! My Approach #I architect robust analytical frameworks, apply precise prompt engineering to guide AI models, then rigorously validate every data point. This synergy of AI’s computational power and strategic human analysis ensures insights that are accurate, relevant and immediately actionable.\n🔗 Let’s connect and explore how data‑driven strategies can transform business outcomes!\n© Chung-Hao Lee. All Rights Reserved. All content on this webpage—including but not limited to text, images, design, code, and multimedia materials—is protected under the international copyright treaties. Unauthorized reproduction, modification, distribution, public transmission, or commercial use is strictly prohibited. Legal action will be taken against infringement. © 李崇豪。保留所有權利。 本網頁之內容（包括但不限於文字、圖片、設計、程式碼及多媒體素材）均受國際著作權條約保護。未經書面授權，嚴禁任何形式之複製、改作、散布、公開傳輸或商業利用。侵權者將依法追訴。\n","date":"May 27, 2025","permalink":"https://chunghaolee.com/portfolio/queensland_aluminum_industry_emissions/","section":"Portfolios","summary":"\u003cp\u003e\u003cstrong\u003eQueensland Aluminium Industry Emissions:\u003c/strong\u003e 2024 Status, 2030 Reduction Targets and Decarbonization Roadmap\u003c/p\u003e\n\u003cdiv style=\"text-align: center; margin: 20px 0;\"\u003e\n    \u003ca href=\"Queensland Aluminium Industry Emissions.pdf\" download class=\"download-button\" style=\"display: inline-block; padding: 12px 24px; background-color: #007bff; color: white; text-decoration: none; border-radius: 5px; font-weight: bold; transition: background-color 0.3s;\"\u003e\n        📥 Download Portfolio PDF\n    \u003c/a\u003e\n\u003c/div\u003e\n\u003cstyle\u003e\n.download-button:hover {\n    background-color: #0056b3 !important;\n}\n\n.pdf-container {\n    width: 100%;\n    max-width: 900px;\n    margin: 0 auto;\n    padding: 20px 0;\n}\n\n.pdf-embed {\n    width: 100%;\n    height: 1200px;\n    border: 1px solid #ddd;\n    border-radius: 8px;\n    box-shadow: 0 2px 10px rgba(0,0,0,0.1);\n    margin-bottom: 30px;\n}\n\n.pdf-page-image {\n    width: 100%;\n    margin-bottom: 20px;\n    border: 1px solid #e0e0e0;\n    border-radius: 8px;\n    box-shadow: 0 2px 8px rgba(0,0,0,0.08);\n}\n\n@media (max-width: 768px) {\n    .pdf-embed {\n        height: 600px;\n    }\n}\n\u003c/style\u003e\n\u003cdiv class=\"pdf-container\"\u003e\n\u003ch3 id=\"-full-portfolio-document\" class=\"relative group\"\u003e📄 Full Portfolio Document \u003cspan class=\"absolute top-0 w-6 transition-opacity opacity-0 -start-6 not-prose group-hover:opacity-100\"\u003e\u003ca class=\"group-hover:text-primary-300 dark:group-hover:text-neutral-700\" style=\"text-decoration-line: none !important;\" href=\"#-full-portfolio-document\" aria-label=\"Anchor\"\u003e#\u003c/a\u003e\u003c/span\u003e\u003c/h3\u003e\u003c!-- Option 1: Using iframe embed (works if PDF is in static folder) --\u003e\n\u003ciframe src=\"Queensland Aluminium Industry Emissions.pdf\" class=\"pdf-embed\" type=\"application/pdf\"\u003e\n    \u003cp\u003eYour browser doesn't support PDF viewing. Please \u003ca href=\"Queensland Aluminium Industry Emissions.pdf\"\u003edownload the PDF\u003c/a\u003e to view it.\u003c/p\u003e","title":"Queensland Aluminium Industry Emissions"},{"content":"Product Portfolio - Telemetry AI Analyzer # 📥 Download Portfolio PDF 📄 Full Portfolio Document # Your browser doesn't support PDF viewing. Please download the PDF to view it.\nKey Highlights #🚀 Product: Telemetry AI Analyzer #An AI system trained on 40+ billion data records from 30+ million notebooks that automatically identifies, analyzes, and predicts potential laptop issues.\n📊 Impact \u0026amp; Results # $1.2M+ cost saving per year 70% debugging efficiency increase 28 engineering symptoms predicted Received Outstanding Leadership and Execution Award from Dell Received Excellent Digital Transformation Project Award from Wistron 🎯 Core Features # One-Stop Platform: Integrated all data, logs, and information for engineers AI Models: Predict potential engineering risks for laptops Issue Library \u0026amp; Search Engine: Knowledge graph with LLM for finding similar cases Dashboards: Real-time monitoring of risks, trends, and statistics 💡 Problem Solved # Reduced laptop repair time from 70+ days to immediate action Eliminated \u0026ldquo;cannot duplicate\u0026rdquo; issues through predictive analytics Enabled proactive issue resolution before failures occur © Chung-Hao Lee. All Rights Reserved. All content on this webpage—including but not limited to text, images, design, code, and multimedia materials—is protected under the international copyright treaties. Unauthorized reproduction, modification, distribution, public transmission, or commercial use is strictly prohibited. Legal action will be taken against infringement. © 李崇豪。保留所有權利。 本網頁之內容（包括但不限於文字、圖片、設計、程式碼及多媒體素材）均受國際著作權條約保護。未經書面授權，嚴禁任何形式之複製、改作、散布、公開傳輸或商業利用。侵權者將依法追訴。\n","date":"December 31, 2024","permalink":"https://chunghaolee.com/portfolio/2025-08-24_chs_portfolio/","section":"Portfolios","summary":"\u003ch2 id=\"product-portfolio---telemetry-ai-analyzer\" class=\"relative group\"\u003eProduct Portfolio - Telemetry AI Analyzer \u003cspan class=\"absolute top-0 w-6 transition-opacity opacity-0 -start-6 not-prose group-hover:opacity-100\"\u003e\u003ca class=\"group-hover:text-primary-300 dark:group-hover:text-neutral-700\" style=\"text-decoration-line: none !important;\" href=\"#product-portfolio---telemetry-ai-analyzer\" aria-label=\"Anchor\"\u003e#\u003c/a\u003e\u003c/span\u003e\u003c/h2\u003e\u003cdiv style=\"text-align: center; margin: 20px 0;\"\u003e\n    \u003ca href=\"CHLee_Portfolio.pdf\" download class=\"download-button\" style=\"display: inline-block; padding: 12px 24px; background-color: #007bff; color: white; text-decoration: none; border-radius: 5px; font-weight: bold; transition: background-color 0.3s;\"\u003e\n        📥 Download Portfolio PDF\n    \u003c/a\u003e\n\u003c/div\u003e\n\u003cstyle\u003e\n.download-button:hover {\n    background-color: #0056b3 !important;\n}\n\n.pdf-container {\n    width: 100%;\n    max-width: 900px;\n    margin: 0 auto;\n    padding: 20px 0;\n}\n\n.pdf-embed {\n    width: 100%;\n    height: 1200px;\n    border: 1px solid #ddd;\n    border-radius: 8px;\n    box-shadow: 0 2px 10px rgba(0,0,0,0.1);\n    margin-bottom: 30px;\n}\n\n.pdf-page-image {\n    width: 100%;\n    margin-bottom: 20px;\n    border: 1px solid #e0e0e0;\n    border-radius: 8px;\n    box-shadow: 0 2px 8px rgba(0,0,0,0.08);\n}\n\n@media (max-width: 768px) {\n    .pdf-embed {\n        height: 600px;\n    }\n}\n\u003c/style\u003e\n\u003cdiv class=\"pdf-container\"\u003e\n\u003ch3 id=\"-full-portfolio-document\" class=\"relative group\"\u003e📄 Full Portfolio Document \u003cspan class=\"absolute top-0 w-6 transition-opacity opacity-0 -start-6 not-prose group-hover:opacity-100\"\u003e\u003ca class=\"group-hover:text-primary-300 dark:group-hover:text-neutral-700\" style=\"text-decoration-line: none !important;\" href=\"#-full-portfolio-document\" aria-label=\"Anchor\"\u003e#\u003c/a\u003e\u003c/span\u003e\u003c/h3\u003e\u003c!-- Option 1: Using iframe embed (works if PDF is in static folder) --\u003e\n\u003ciframe src=\"CHLee_Portfolio.pdf\" class=\"pdf-embed\" type=\"application/pdf\"\u003e\n    \u003cp\u003eYour browser doesn't support PDF viewing. Please \u003ca href=\"CHLee_Portfolio.pdf\"\u003edownload the PDF\u003c/a\u003e to view it.\u003c/p\u003e","title":"My product portfolio at Wiston"},{"content":"I am deeply grateful to my team for their support and dedicated efforts in product development over the past year. It is also a great honor to be recognized by Dell with the 2023 Dell Outstanding Leadership \u0026amp; Execution Award.\n© Chung-Hao Lee. All Rights Reserved. All content on this webpage—including but not limited to text, images, design, code, and multimedia materials—is protected under the international copyright treaties. Unauthorized reproduction, modification, distribution, public transmission, or commercial use is strictly prohibited. Legal action will be taken against infringement. © 李崇豪。保留所有權利。 本網頁之內容（包括但不限於文字、圖片、設計、程式碼及多媒體素材）均受國際著作權條約保護。未經書面授權，嚴禁任何形式之複製、改作、散布、公開傳輸或商業利用。侵權者將依法追訴。\n","date":"October 31, 2023","permalink":"https://chunghaolee.com/posts/2023-10-31-dell-outstanding-award/","section":"Posts","summary":"\u003cp\u003eI am deeply grateful to my team for their support and dedicated efforts in product development over the past year. It is also a great honor to be recognized by Dell with the 2023 Dell Outstanding Leadership \u0026amp; Execution Award.\u003c/p\u003e\n\u003cp\u003e\n\n\n\n\n\n\u003cfigure\u003e\n    \n    \n\n\n\n\n\n\n\n\n  \n    \u003cpicture\n      class=\"mx-auto my-0 rounded-md\"\n      \n    \u003e\n      \n      \n      \n      \n        \u003csource\n          \n            srcset=\"https://chunghaolee.com/posts/2023-10-31-dell-outstanding-award/pic_hu_bffd855fbbd2350a.webp 330w,https://chunghaolee.com/posts/2023-10-31-dell-outstanding-award/pic_hu_3a7ed46a22827df0.webp 660w\n            \n              ,https://chunghaolee.com/posts/2023-10-31-dell-outstanding-award/pic_hu_cfff975159ca1318.webp 1024w\n            \n            \n              ,https://chunghaolee.com/posts/2023-10-31-dell-outstanding-award/pic_hu_4af46b2b8bc5dcb0.webp 1320w\n            \"\n          \n          sizes=\"100vw\"\n          type=\"image/webp\"\n        /\u003e\n      \n      \u003cimg\n        width=\"2048\"\n        height=\"1130\"\n        class=\"mx-auto my-0 rounded-md\"\n        alt=\"Award\"\n        loading=\"lazy\" decoding=\"async\"\n        \n          src=\"https://chunghaolee.com/posts/2023-10-31-dell-outstanding-award/pic_hu_87143b25e0064152.jpg\" srcset=\"https://chunghaolee.com/posts/2023-10-31-dell-outstanding-award/pic_hu_43744be66693c61e.jpg 330w,https://chunghaolee.com/posts/2023-10-31-dell-outstanding-award/pic_hu_87143b25e0064152.jpg 660w\n          \n            ,https://chunghaolee.com/posts/2023-10-31-dell-outstanding-award/pic_hu_c47b2b79ebc15619.jpg 1024w\n          \n          \n            ,https://chunghaolee.com/posts/2023-10-31-dell-outstanding-award/pic_hu_a571e2897c9fc985.jpg 1320w\n          \"\n          sizes=\"100vw\"\n        \n      /\u003e\n    \u003c/picture\u003e\n  \n\n\n\u003c/figure\u003e\n\u003c/p\u003e\n\u003chr\u003e\n\u003cp\u003e© Chung-Hao Lee. All Rights Reserved. All content on this webpage—including but not limited to text, images, design, code, and multimedia materials—is protected under the international copyright treaties. Unauthorized reproduction, modification, distribution, public transmission, or commercial use is strictly prohibited. Legal action will be taken against infringement. \u003cbr\u003e\n© 李崇豪。保留所有權利。 本網頁之內容（包括但不限於文字、圖片、設計、程式碼及多媒體素材）均受國際著作權條約保護。未經書面授權，嚴禁任何形式之複製、改作、散布、公開傳輸或商業利用。侵權者將依法追訴。\u003c/p\u003e","title":"2023 Dell Outstanding Leadership \u0026 Execution Award"},{"content":"This is my first time co-hosting a workshop. I presented a demo of our product to participants and led a collaborative discussion on strategies for resolving challenging engineering issues in notebooks.\nPresented a demo of a product that my team had been developing for a year Explained features of the product © Chung-Hao Lee. All Rights Reserved. All content on this webpage—including but not limited to text, images, design, code, and multimedia materials—is protected under the international copyright treaties. Unauthorized reproduction, modification, distribution, public transmission, or commercial use is strictly prohibited. Legal action will be taken against infringement. © 李崇豪。保留所有權利。 本網頁之內容（包括但不限於文字、圖片、設計、程式碼及多媒體素材）均受國際著作權條約保護。未經書面授權，嚴禁任何形式之複製、改作、散布、公開傳輸或商業利用。侵權者將依法追訴。\n","date":"October 24, 2023","permalink":"https://chunghaolee.com/posts/2023-10-24-co-host-a-workshop/","section":"Posts","summary":"\u003cp\u003eThis is my first time co-hosting a workshop. I presented a demo of our product to participants and led a collaborative discussion on strategies for resolving challenging engineering issues in notebooks.\u003c/p\u003e\n\u003cp\u003e\n\n\n\n\n\n\u003cfigure\u003e\n    \n    \n\n\n\n\n\n\n\n\n  \n    \u003cpicture\n      class=\"mx-auto my-0 rounded-md\"\n      \n    \u003e\n      \n      \n      \n      \n        \u003csource\n          \n            srcset=\"https://chunghaolee.com/posts/2023-10-24-co-host-a-workshop/pic2_hu_6fafeb0d30afaeca.webp 330w,https://chunghaolee.com/posts/2023-10-24-co-host-a-workshop/pic2_hu_57e96bd5873b345b.webp 660w\n            \n              ,https://chunghaolee.com/posts/2023-10-24-co-host-a-workshop/pic2_hu_d73e34eec78d6d6e.webp 1024w\n            \n            \n              \n                ,https://chunghaolee.com/posts/2023-10-24-co-host-a-workshop/pic2_hu_18856814342789a3.webp 1225w\n              \n            \"\n          \n          sizes=\"100vw\"\n          type=\"image/webp\"\n        /\u003e\n      \n      \u003cimg\n        width=\"1225\"\n        height=\"1304\"\n        class=\"mx-auto my-0 rounded-md\"\n        alt=\"pic2\"\n        loading=\"lazy\" decoding=\"async\"\n        \n          src=\"https://chunghaolee.com/posts/2023-10-24-co-host-a-workshop/pic2_hu_654d538b8c3c724e.jpeg\" srcset=\"https://chunghaolee.com/posts/2023-10-24-co-host-a-workshop/pic2_hu_3a62cda691fdbf45.jpeg 330w,https://chunghaolee.com/posts/2023-10-24-co-host-a-workshop/pic2_hu_654d538b8c3c724e.jpeg 660w\n          \n            ,https://chunghaolee.com/posts/2023-10-24-co-host-a-workshop/pic2_hu_88b358ae77e23e4c.jpeg 1024w\n          \n          \n            ,https://chunghaolee.com/posts/2023-10-24-co-host-a-workshop/pic2.jpeg 1225w\n          \"\n          sizes=\"100vw\"\n        \n      /\u003e\n    \u003c/picture\u003e\n  \n\n\u003cfigcaption class=\"text-center\"\u003ePresented a demo of a product that my team had been developing for a year\u003c/figcaption\u003e\n\u003c/figure\u003e\n\u003c/p\u003e\n\u003cp\u003e\n\n\n\n\n\n\u003cfigure\u003e\n    \n    \n\n\n\n\n\n\n\n\n  \n    \u003cpicture\n      class=\"mx-auto my-0 rounded-md\"\n      \n    \u003e\n      \n      \n      \n      \n        \u003csource\n          \n            srcset=\"https://chunghaolee.com/posts/2023-10-24-co-host-a-workshop/pic1_hu_35b847224797b3e1.webp 330w,https://chunghaolee.com/posts/2023-10-24-co-host-a-workshop/pic1_hu_36b7e6c67697e6b7.webp 660w\n            \n              ,https://chunghaolee.com/posts/2023-10-24-co-host-a-workshop/pic1_hu_b91b9e4ea8f4fb6.webp 1024w\n            \n            \n              ,https://chunghaolee.com/posts/2023-10-24-co-host-a-workshop/pic1_hu_74e57c06dd3cfecc.webp 1320w\n            \"\n          \n          sizes=\"100vw\"\n          type=\"image/webp\"\n        /\u003e\n      \n      \u003cimg\n        width=\"3024\"\n        height=\"4032\"\n        class=\"mx-auto my-0 rounded-md\"\n        alt=\"pic1\"\n        loading=\"lazy\" decoding=\"async\"\n        \n          src=\"https://chunghaolee.com/posts/2023-10-24-co-host-a-workshop/pic1_hu_cfe60d68402c748b.jpeg\" srcset=\"https://chunghaolee.com/posts/2023-10-24-co-host-a-workshop/pic1_hu_28c591b643fabc29.jpeg 330w,https://chunghaolee.com/posts/2023-10-24-co-host-a-workshop/pic1_hu_cfe60d68402c748b.jpeg 660w\n          \n            ,https://chunghaolee.com/posts/2023-10-24-co-host-a-workshop/pic1_hu_6e0dfddc5095299f.jpeg 1024w\n          \n          \n            ,https://chunghaolee.com/posts/2023-10-24-co-host-a-workshop/pic1_hu_e4249a6de7c0623a.jpeg 1320w\n          \"\n          sizes=\"100vw\"\n        \n      /\u003e\n    \u003c/picture\u003e\n  \n\n\u003cfigcaption class=\"text-center\"\u003eExplained features of the product\u003c/figcaption\u003e\n\u003c/figure\u003e\n\u003c/p\u003e","title":"First time to co-host a workshop"},{"content":"","date":"September 21, 2023","permalink":"https://chunghaolee.com/posts/external-article-umd-research-could-help-minor-leaguers-swing-for-the-fences/","section":"Posts","summary":"","title":"[External Article] - UMD Research Could Help Minor Leaguers Swing for the Fences"},{"content":"Media Coverage #Honored to be featured in Maryland Today for my research work!\nUMD Research Could Help Minor Leaguers Swing for the Fences New AI-Powered Analytical Model Tracks Baseball Players' Path to the Big Show. Using machine learning to analyze baseball performance data, helping minor league players identify skills that need improvement to increase their chances of reaching the majors. Research found that batting average is more decisive than on-base percentage for player call-ups.\ntoday.umd.edu • September 21, 2023 Click the link for reading the report !\n© Chung-Hao Lee. All Rights Reserved. All content on this webpage—including but not limited to text, images, design, code, and multimedia materials—is protected under the international copyright treaties. Unauthorized reproduction, modification, distribution, public transmission, or commercial use is strictly prohibited. Legal action will be taken against infringement. © 李崇豪。保留所有權利。 本網頁之內容（包括但不限於文字、圖片、設計、程式碼及多媒體素材）均受國際著作權條約保護。未經書面授權，嚴禁任何形式之複製、改作、散布、公開傳輸或商業利用。侵權者將依法追訴。\n","date":"September 21, 2023","permalink":"https://chunghaolee.com/posts/2023-09-21-umd-research-could-help-minor-leaguers-swing-for-the-fences/","section":"Posts","summary":"\u003ch2 id=\"media-coverage\" class=\"relative group\"\u003eMedia Coverage \u003cspan class=\"absolute top-0 w-6 transition-opacity opacity-0 -start-6 not-prose group-hover:opacity-100\"\u003e\u003ca class=\"group-hover:text-primary-300 dark:group-hover:text-neutral-700\" style=\"text-decoration-line: none !important;\" href=\"#media-coverage\" aria-label=\"Anchor\"\u003e#\u003c/a\u003e\u003c/span\u003e\u003c/h2\u003e\u003cp\u003e\u003cstrong\u003eHonored to be featured in \u003ca href=\"https://today.umd.edu/\" target=\"_blank\" rel=\"noreferrer\"\u003eMaryland Today\u003c/a\u003e for my research work!\u003c/strong\u003e\u003c/p\u003e\n\u003chr\u003e\n\u003cdiv class=\"link-preview\"\u003e\n  \u003ca href=\"https://today.umd.edu/umd-research-could-help-minor-leaguers-swing-for-the-fences\" target=\"_blank\" rel=\"noopener\"\u003e\n    \u003cdiv class=\"preview-card\"\u003e\n      \u003cdiv class=\"preview-image\"\u003e\n        \u003cimg src=\"pic1.jpg\" alt=\"Baseball player batting\" /\u003e\n      \u003c/div\u003e\n      \u003cdiv class=\"preview-content\"\u003e\n        \u003ch3\u003eUMD Research Could Help Minor Leaguers Swing for the Fences\u003c/h3\u003e\n        \u003cp\u003eNew AI-Powered Analytical Model Tracks Baseball Players' Path to the Big Show. Using machine learning to analyze baseball performance data, helping minor league players identify skills that need improvement to increase their chances of reaching the majors. Research found that batting average is more decisive than on-base percentage for player call-ups.\u003c/p\u003e","title":"My research project was featured in Maryland Today !"},{"content":"","date":"September 15, 2023","permalink":"https://chunghaolee.com/posts/external-article-smith-school-research-helps-minor-league-baseball-players-swing-for-the-fences/","section":"Posts","summary":"","title":"[External Article] - Smith School Research Helps Minor League Baseball Players Swing for the Fences"},{"content":" Please find the Paper Link\nSummary: #The study analyzes data from 4,321 fielders (non-pitchers) drafted by MLB teams between 2001 and 2010, using their aggregated MiLB career statistics and draft information. The goal is to build predictive models that can accurately determine whether a player will reach the MLB, a feat achieved by less than 20% of drafted prospects. The authors employ exploratory data analysis (EDA), variable selection techniques (including Lasso regression), and four different ML models (eXtreme Gradient Boosting - XGB, Random Forest - RF, Decision Tree - DT, and Support Vector Machine - SVM) to achieve this. The models are trained on 75% of the data and validated on the remaining 25%. The best-performing model is then tested on an out-of-sample dataset of players drafted in 2013.\nKey Highlights: # High Prediction Accuracy: All 16 model/variable selection combinations achieved validation accuracies above 90% and Area Under the Curve (AUC) scores above 0.88, indicating strong predictive power. The models were also successfully validated on 2013 draft data, with the best model maintaining an accuracy above 89%.\nBest Performing Model: The eXtreme Gradient Boosting (XGB) model consistently outperformed the other ML techniques (RF, DT, SVM) across key metrics like F-score and AUC, demonstrating its effectiveness for this prediction task. Effective Variable Selection: The study found that using Lasso regression for variable selection generally improved model performance. The most effective combination was the XGB model using the \u0026ldquo;Lasso4lch\u0026rdquo; variable set, which contained only 11 key variables.\nMost Important Predictive Factors: Overall Draft Pick: This non-baseball variable was consistently identified as one of the most important predictors. Being drafted earlier significantly increases the likelihood of an MLB debut.\nOn-Base Plus Slugging (OPS): This baseball statistic, which combines a player\u0026rsquo;s ability to get on base (OBP) and hit for power (SLG), was found to be the most representative and impactful performance metric for predicting success.\nBatting Average (AVG) and At-Bats (AB): AVG was also consistently ranked as a top predictor. A high number of ABs was strongly associated with making the MLB, suggesting that longevity and consistent playing time in the minors are crucial factors.\nSurprising Finding on OBP: Contrary to the \u0026ldquo;Moneyball\u0026rdquo; philosophy that popularized On-Base Percentage (OBP), the models often found AVG to be a more important predictor than OBP alone. The authors suggest this is because OPS, which incorporates both OBP and SLG, provides a more comprehensive measure, and AVG is a fundamental component of OBP.\nPractical Application: The research provides a framework for MLB teams to make more informed decisions about player development and for MiLB players to understand the key factors (staying in the system longer, improving AVG and SLG/OPS) that can increase their chances of reaching the majors.\n© Chung-Hao Lee. All Rights Reserved. All content on this webpage—including but not limited to text, images, design, code, and multimedia materials—is protected under the international copyright treaties. Unauthorized reproduction, modification, distribution, public transmission, or commercial use is strictly prohibited. Legal action will be taken against infringement. © 李崇豪。保留所有權利。 本網頁之內容（包括但不限於文字、圖片、設計、程式碼及多媒體素材）均受國際著作權條約保護。未經書面授權，嚴禁任何形式之複製、改作、散布、公開傳輸或商業利用。侵權者將依法追訴。\n","date":"May 24, 2023","permalink":"https://chunghaolee.com/portfolio/baseball-informaticsfrom-milb-to-mlb-debut/","section":"Portfolios","summary":"\u003chr\u003e\n\u003cp\u003ePlease find the \u003ca href=\"https://link.springer.com/chapter/10.1007/978-981-19-9658-0_5\" target=\"_blank\" rel=\"noreferrer\"\u003ePaper Link\u003c/a\u003e\u003c/p\u003e\n\u003ch3 id=\"summary\" class=\"relative group\"\u003eSummary: \u003cspan class=\"absolute top-0 w-6 transition-opacity opacity-0 -start-6 not-prose group-hover:opacity-100\"\u003e\u003ca class=\"group-hover:text-primary-300 dark:group-hover:text-neutral-700\" style=\"text-decoration-line: none !important;\" href=\"#summary\" aria-label=\"Anchor\"\u003e#\u003c/a\u003e\u003c/span\u003e\u003c/h3\u003e\u003cp\u003eThe study analyzes data from 4,321 fielders (non-pitchers) drafted by MLB teams between 2001 and 2010, using their aggregated MiLB career statistics and draft information. The goal is to build predictive models that can accurately determine whether a player will reach the MLB, a feat achieved by less than 20% of drafted prospects. The authors employ exploratory data analysis (EDA), variable selection techniques (including Lasso regression), and four different ML models (eXtreme Gradient Boosting - XGB, Random Forest - RF, Decision Tree - DT, and Support Vector Machine - SVM) to achieve this. The models are trained on 75% of the data and validated on the remaining 25%. The best-performing model is then tested on an out-of-sample dataset of players drafted in 2013.\u003c/p\u003e","title":"Baseball Informatics—From MiLB to MLB Debut"},{"content":" 📥 Download Portfolio PDF 📄 Full Portfolio Document # Your browser doesn't support PDF viewing. Please download the PDF to view it.\nSummary: #This paper introduces ADDCAT (Anomaly Detection in Driving by Cluster Analysis Twice), a novel method for detecting driving anomalies (like aggressive driving or potholes) using vehicle sensor data (G-sensor and gyroscope). Unlike popular deep learning methods that require extensive training and computational power, ADDCAT leverages the HDBSCAN clustering algorithm applied twice to identify anomalies based on the principle that normal driving events form the largest, most cohesive cluster, while anomalies are outliers.\nThe method involves: # Data Processing: Sensor data is aggregated into 2-second intervals (\u0026ldquo;events\u0026rdquo;) and 20 aggregate statistics (AS) are calculated (e.g., mean speed, std of acceleration).\nCalibration: To remove the confounding effect of vehicle speed on sensor readings, a linear calibration is performed, making the AS speed-independent.\nDouble Clustering:\nFirst Clustering: All 20 calibrated and normalized AS are fed into HDBSCAN to separate events primarily by speed and identify a major \u0026ldquo;normal\u0026rdquo; cluster and outliers.\nSecond Clustering: HDBSCAN is applied again, separately for the X, Y, and Z axes, using only specific subsets of the AS (mainly standard deviations) from the largest cluster and the outlier group identified in the first step. Anomaly Identification: An event is classified as an anomaly if it does not belong to the largest cluster in all three axis-specific second clusterings.\nValidated on an open dataset containing pothole information, ADDCAT achieved an accuracy of 0.62, successfully identifying 31 out of 79 potholes.\nHighlights: # No Training Required: A major advantage of ADDCAT is that it is an unsupervised method, eliminating the need for labeled training data, which is costly and time-consuming to acquire, especially for autonomous driving development.\nLow Computational Cost: Compared to deep learning models, ADDCAT is designed to be computationally lighter, making it potentially more suitable for real-time applications or resource-constrained environments.\nAdaptive Clustering: The use of HDBSCAN is key, as it can automatically determine the number of clusters and handle clusters of varying densities, which is ideal for unpredictable real-world driving data.\nSpeed Calibration: The paper identifies vehicle speed as a dominant factor and implements a calibration step to isolate its effect, allowing other sensor features to better distinguish between normal and anomalous events.\nTwo-Stage Process: The \u0026ldquo;Cluster Analysis Twice\u0026rdquo; approach refines the detection. The first pass groups events broadly, and the second pass performs a more focused analysis on specific physical axes to pinpoint anomalies.\nPractical Application: The method is positioned as highly valuable for the Transportation and Logistics (T\u0026amp;L) industry to monitor driver behavior, road conditions, and overall delivery efficiency. It\u0026rsquo;s also suggested as a tool to help label data for training more complex autonomous driving systems.\nPerformance Insight: While achieving moderate accuracy (0.62), the results show ADDCAT is effective at detecting anomalies across different speeds. However, it struggles with events that have very high variability in speed or whose sensor signatures become too similar to normal events after calibration.\n© Chung-Hao Lee. All Rights Reserved. All content on this webpage—including but not limited to text, images, design, code, and multimedia materials—is protected under the international copyright treaties. Unauthorized reproduction, modification, distribution, public transmission, or commercial use is strictly prohibited. Legal action will be taken against infringement. © 李崇豪。保留所有權利。 本網頁之內容（包括但不限於文字、圖片、設計、程式碼及多媒體素材）均受國際著作權條約保護。未經書面授權，嚴禁任何形式之複製、改作、散布、公開傳輸或商業利用。侵權者將依法追訴。\n","date":"December 15, 2022","permalink":"https://chunghaolee.com/portfolio/2022-12-15-anomaly-detection-in-driving-by-cluster-analysis-twice/","section":"Portfolios","summary":"\u003cdiv style=\"text-align: center; margin: 20px 0;\"\u003e\n    \u003ca href=\"Anomaly Detection in Driving by Cluster Analysis Twice.pdf\" download class=\"download-button\" style=\"display: inline-block; padding: 12px 24px; background-color: #007bff; color: white; text-decoration: none; border-radius: 5px; font-weight: bold; transition: background-color 0.3s;\"\u003e\n        📥 Download Portfolio PDF\n    \u003c/a\u003e\n\u003c/div\u003e\n\u003cstyle\u003e\n.download-button:hover {\n    background-color: #0056b3 !important;\n}\n\n.pdf-container {\n    width: 100%;\n    max-width: 900px;\n    margin: 0 auto;\n    padding: 20px 0;\n}\n\n.pdf-embed {\n    width: 100%;\n    height: 1200px;\n    border: 1px solid #ddd;\n    border-radius: 8px;\n    box-shadow: 0 2px 10px rgba(0,0,0,0.1);\n    margin-bottom: 30px;\n}\n\n.pdf-page-image {\n    width: 100%;\n    margin-bottom: 20px;\n    border: 1px solid #e0e0e0;\n    border-radius: 8px;\n    box-shadow: 0 2px 8px rgba(0,0,0,0.08);\n}\n\n@media (max-width: 768px) {\n    .pdf-embed {\n        height: 600px;\n    }\n}\n\u003c/style\u003e\n\u003cdiv class=\"pdf-container\"\u003e\n\u003ch3 id=\"-full-portfolio-document\" class=\"relative group\"\u003e📄 Full Portfolio Document \u003cspan class=\"absolute top-0 w-6 transition-opacity opacity-0 -start-6 not-prose group-hover:opacity-100\"\u003e\u003ca class=\"group-hover:text-primary-300 dark:group-hover:text-neutral-700\" style=\"text-decoration-line: none !important;\" href=\"#-full-portfolio-document\" aria-label=\"Anchor\"\u003e#\u003c/a\u003e\u003c/span\u003e\u003c/h3\u003e\u003c!-- Option 1: Using iframe embed (works if PDF is in static folder) --\u003e\n\u003ciframe src=\"Anomaly Detection in Driving by Cluster Analysis Twice.pdf\" class=\"pdf-embed\" type=\"application/pdf\"\u003e\n    \u003cp\u003eYour browser doesn't support PDF viewing. Please \u003ca href=\"Anomaly Detection in Driving by Cluster Analysis Twice.pdf\"\u003edownload the PDF\u003c/a\u003e to view it.\u003c/p\u003e","title":"Anomaly Detection in Driving by Cluster Analysis Twice"},{"content":"Honored to share my research in UMD GRAD 2022 and recognized as 2nd place\nGraduate Research Appreciation Day (GRAD) provides an on-campus conference venue in which graduate students from all academic backgrounds can share their research or work, obtain valuable faculty and peer feedback, and hone their communication skills within the comfort of their own campus environment.\nMe take photo with my poster Me take photo with Award winner list! © Chung-Hao Lee. All Rights Reserved. All content on this webpage—including but not limited to text, images, design, code, and multimedia materials—is protected under the international copyright treaties. Unauthorized reproduction, modification, distribution, public transmission, or commercial use is strictly prohibited. Legal action will be taken against infringement. © 李崇豪。保留所有權利。 本網頁之內容（包括但不限於文字、圖片、設計、程式碼及多媒體素材）均受國際著作權條約保護。未經書面授權，嚴禁任何形式之複製、改作、散布、公開傳輸或商業利用。侵權者將依法追訴。\n","date":"May 11, 2022","permalink":"https://chunghaolee.com/posts/2022-05-11-grad-2nd-place/","section":"Posts","summary":"\u003cp\u003e\u003cstrong\u003eHonored to share my research in UMD GRAD 2022 and recognized as 2nd place\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eGraduate Research Appreciation Day (GRAD) provides an on-campus conference venue in which graduate students from all academic backgrounds can share their research or work, obtain valuable faculty and peer feedback, and hone their communication skills within the comfort of their own campus environment.\u003c/p\u003e\n\u003cp\u003e\n\n\n\n\n\n\u003cfigure\u003e\n    \n    \n\n\n\n\n\n\n\n\n  \n    \u003cpicture\n      class=\"mx-auto my-0 rounded-md\"\n      \n    \u003e\n      \n      \n      \n      \n        \u003csource\n          \n            srcset=\"https://chunghaolee.com/posts/2022-05-11-grad-2nd-place/pic1_hu_76e42abb0aacf28.webp 330w,https://chunghaolee.com/posts/2022-05-11-grad-2nd-place/pic1_hu_9184370d585c88db.webp 660w\n            \n              ,https://chunghaolee.com/posts/2022-05-11-grad-2nd-place/pic1_hu_8a1435513172591d.webp 1024w\n            \n            \n              ,https://chunghaolee.com/posts/2022-05-11-grad-2nd-place/pic1_hu_6399a1d48cb37240.webp 1320w\n            \"\n          \n          sizes=\"100vw\"\n          type=\"image/webp\"\n        /\u003e\n      \n      \u003cimg\n        width=\"2726\"\n        height=\"3407\"\n        class=\"mx-auto my-0 rounded-md\"\n        alt=\"pic1\"\n        loading=\"lazy\" decoding=\"async\"\n        \n          src=\"https://chunghaolee.com/posts/2022-05-11-grad-2nd-place/pic1_hu_5346b4824cff7f3f.jpeg\" srcset=\"https://chunghaolee.com/posts/2022-05-11-grad-2nd-place/pic1_hu_e8042e7e2a2624f8.jpeg 330w,https://chunghaolee.com/posts/2022-05-11-grad-2nd-place/pic1_hu_5346b4824cff7f3f.jpeg 660w\n          \n            ,https://chunghaolee.com/posts/2022-05-11-grad-2nd-place/pic1_hu_81581f44267c3a0a.jpeg 1024w\n          \n          \n            ,https://chunghaolee.com/posts/2022-05-11-grad-2nd-place/pic1_hu_a700becb2d91a03f.jpeg 1320w\n          \"\n          sizes=\"100vw\"\n        \n      /\u003e\n    \u003c/picture\u003e\n  \n\n\u003cfigcaption class=\"text-center\"\u003eMe take photo with my poster\u003c/figcaption\u003e\n\u003c/figure\u003e\n\u003c/p\u003e","title":"2nd Place in UMD GRAD 2022"},{"content":"","date":"October 24, 2018","permalink":"https://chunghaolee.com/posts/external-article-%E6%BC%94%E7%B9%B9%E8%88%87%E6%AD%B8%E7%B4%8D/","section":"Posts","summary":"","title":"[External Article] - 演繹與歸納"},{"content":" 📥 Download Portfolio PDF 📄 Full Portfolio Document # Your browser doesn't support PDF viewing. Please download the PDF to view it.\nThis study successfully developed a method for the long-term cryopreservation of recalcitrant Blue Waterlily (Nymphaea caerulea) seeds by creating an enhanced cryoprotectant solution.\nSummary:\nThe research first confirmed that Nymphaea caerulea seeds are recalcitrant, meaning they cannot tolerate desiccation (drying) or freezing temperatures. Traditional seed banking methods (drying and storing at -20°C) were ineffective, with seeds losing viability within 6 months, and direct exposure to -20°C or liquid nitrogen (LN) killing them outright.\nThe standard cryoprotectant, Plant Vitrification Solution 3 (PVS3), was also insufficient, yielding a maximum survival rate of only 23% after cryopreservation in LN. The study identified oxidative stress from the accumulation of Reactive Oxygen Species (ROS), particularly hydrogen peroxide (H₂O₂), as the primary cause of this failure. The cryopreservation process itself induced a massive surge in H₂O₂ levels within the seeds.\nTo address this, the researchers developed a modified solution called PVS+ by adding the antioxidant glutathione (GSH) to PVS3. The optimal concentration was found to be 0.08 mM GSH. PVS+ dramatically reduced ROS accumulation during cryopreservation, keeping H₂O₂ levels near normal. This intervention rescued seed viability, increasing the survival rate after cryopreservation in LN to an impressive 97%.\nKey Highlights:\nRecalcitrant Seeds: Nymphaea caerulea seeds were definitively classified as recalcitrant, being highly sensitive to drying and cold, making traditional storage methods unsuitable for long-term conservation. PVS3 Inefficacy: The commonly used cryoprotectant PVS3 failed to adequately protect the seeds, achieving only a 23% survival rate after cryopreservation. ROS as the Culprit: The primary reason for PVS3\u0026rsquo;s failure was identified as oxidative stress caused by a significant buildup of ROS (specifically H₂O₂) during the cryopreservation process. PVS+ Solution: A novel vitrification solution, PVS+, was created by adding 0.08 mM glutathione (GSH), a potent antioxidant, to PVS3. High Success Rate: PVS+ effectively mitigated oxidative stress by scavenging ROS, resulting in a dramatically improved survival rate of 97% for cryopreserved seeds. New Strategy: The study demonstrates that suppressing ROS-induced oxidative stress using antioxidants like glutathione is a highly effective and promising strategy for improving the cryopreservation of recalcitrant seeds. © Chung-Hao Lee. All Rights Reserved. All content on this webpage—including but not limited to text, images, design, code, and multimedia materials—is protected under the international copyright treaties. Unauthorized reproduction, modification, distribution, public transmission, or commercial use is strictly prohibited. Legal action will be taken against infringement. © 李崇豪。保留所有權利。 本網頁之內容（包括但不限於文字、圖片、設計、程式碼及多媒體素材）均受國際著作權條約保護。未經書面授權，嚴禁任何形式之複製、改作、散布、公開傳輸或商業利用。侵權者將依法追訴。\n","date":"June 29, 2016","permalink":"https://chunghaolee.com/portfolio/cryopreservation-of-seeds-of-blue-waterlily-nymphaea-caerulea-using-glutathione-adding-plant-vitrification-solution-pvs+/","section":"Portfolios","summary":"\u003cdiv style=\"text-align: center; margin: 20px 0;\"\u003e\n    \u003ca href=\"Cryopreservation of seeds of blue waterlily -Nymphaea caerulea- using glutathione adding plant vitrification solution, PVS+.pdf\" download class=\"download-button\" style=\"display: inline-block; padding: 12px 24px; background-color: #007bff; color: white; text-decoration: none; border-radius: 5px; font-weight: bold; transition: background-color 0.3s;\"\u003e\n        📥 Download Portfolio PDF\n    \u003c/a\u003e\n\u003c/div\u003e\n\u003cstyle\u003e\n.download-button:hover {\n    background-color: #0056b3 !important;\n}\n\n.pdf-container {\n    width: 100%;\n    max-width: 900px;\n    margin: 0 auto;\n    padding: 20px 0;\n}\n\n.pdf-embed {\n    width: 100%;\n    height: 1200px;\n    border: 1px solid #ddd;\n    border-radius: 8px;\n    box-shadow: 0 2px 10px rgba(0,0,0,0.1);\n    margin-bottom: 30px;\n}\n\n.pdf-page-image {\n    width: 100%;\n    margin-bottom: 20px;\n    border: 1px solid #e0e0e0;\n    border-radius: 8px;\n    box-shadow: 0 2px 8px rgba(0,0,0,0.08);\n}\n\n@media (max-width: 768px) {\n    .pdf-embed {\n        height: 600px;\n    }\n}\n\u003c/style\u003e\n\u003cdiv class=\"pdf-container\"\u003e\n\u003ch3 id=\"-full-portfolio-document\" class=\"relative group\"\u003e📄 Full Portfolio Document \u003cspan class=\"absolute top-0 w-6 transition-opacity opacity-0 -start-6 not-prose group-hover:opacity-100\"\u003e\u003ca class=\"group-hover:text-primary-300 dark:group-hover:text-neutral-700\" style=\"text-decoration-line: none !important;\" href=\"#-full-portfolio-document\" aria-label=\"Anchor\"\u003e#\u003c/a\u003e\u003c/span\u003e\u003c/h3\u003e\u003c!-- Option 1: Using iframe embed (works if PDF is in static folder) --\u003e\n\u003ciframe src=\"Cryopreservation of seeds of blue waterlily -Nymphaea caerulea- using glutathione adding plant vitrification solution, PVS+.pdf\" class=\"pdf-embed\" type=\"application/pdf\"\u003e\n    \u003cp\u003eYour browser doesn't support PDF viewing. Please \u003ca href=\"Cryopreservation of seeds of blue waterlily -Nymphaea caerulea- using glutathione adding plant vitrification solution, PVS+.pdf\"\u003edownload the PDF\u003c/a\u003e to view it.\u003c/p\u003e","title":"Cryopreservation of seeds of blue waterlily (Nymphaea caerulea) using glutathione adding plant vitrification solution, PVS+"},{"content":"\u0026quot; Product = Analysis + Process \u0026ldquo;\nHi, I am Chung-Hao Lee 李崇豪\nI am a Product Operation Manager managing AI/ML products at TikTok. Prior to TikTok, I have 6+ years experiences at Wistron as AI PM and at Pegatron as Factory PM.\nI received MBA and Master’s in Information Systems dual degrees from University of Maryland, College Park and Bachelor’s degree from Tsing Hua University.\nI bring comprehensive expertise in technology product development, integrating data-driven decision-making to optimize Product Development, Supply Chain Management, and Manufacturing.\nI like being an analyst who enjoys exploring diverse areas such as product design, business models, baseball data, Supply Chain, and geopolitics. I’m passionate about discovering the essence of these subjects and understanding the underlying \u0026lsquo;why\u0026rsquo;, \u0026lsquo;what\u0026rsquo;, and \u0026lsquo;how\u0026rsquo;.\nIn my spare time, I am an avid LA Dodgers’ fan and I am a photographer who loves to record beauty of daily life.\nActivity Summaries # [2025 - Now] Product Operation Manager, Model Operation at TikTok [2022 - 2024] AI Product Manager at Wistron [2020 - 2022] Research Assistent at University of Maryland - College Park [2018 - 2020] Factory PM at Pegatron [2016 - 2017] SMS at International Taoyuan Airport Education # MBA \u0026amp; MSIS, University of Maryland - College Park (2022) Bachelor, Tsinghau University (2014) [Achievements] #TikTok: # Spot Bonus Award, 2025 Q4 Wistron: # Dell Outstanding Leadership \u0026amp; Execution Award, 2023 Co-hosted a Dell-Wistron Joint Engineering Workshop, 2023 Scrum Master Certification, 2023 Published Anomaly Detection in Driving by Cluster Analysis Twice University of Maryland, College Park: # Smith Fellow scholarship, Recipient UMD Outstanding Graduate Student Award, 2022, Nomination Reported by Maryland Today Published Baseball Informatics—From MiLB to MLB Debut Impact Consulting Fellowship, Summer 2021, 1st Place ISM Global Case Competition, January 2022 UMD Info Challenge, March 2022 Bitcamp, April 2022 Graduate Research Appreciation Day, April 2022, 2nd Place UMD Smith Analytics Consortium Deloitte datathon, May 2022 UMD IS \u0026amp; Hearst Communications Inc. Data Competition, May 2022, 3rd Place Selected as VP for Smith Masters Supply Chain Management Association \u0026amp; APICS Management Substitute Millatary Service: # Honorably discharged, summa cum laude Received a certificate of merit from the Ministry of the Interior (Taiwan) Tsing Hua University: # Dean’s list, 2012 Spring Dean’s list, 2011 Fall Dean’s list, 2011 Spring ","date":null,"permalink":"https://chunghaolee.com/about/","section":"","summary":"\u003cp\u003e\u003cstrong\u003e\u0026quot; Product = Analysis + Process \u0026ldquo;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eHi, I am Chung-Hao Lee 李崇豪\u003c/p\u003e\n\u003cp\u003eI am a Product Operation Manager managing AI/ML products at TikTok. Prior to TikTok, I have 6+ years experiences at Wistron as AI PM and at Pegatron as Factory PM.\u003c/p\u003e\n\u003cp\u003eI received MBA and Master’s in Information Systems dual degrees from University of Maryland, College Park and Bachelor’s degree from Tsing Hua University.\u003c/p\u003e\n\u003cp\u003eI bring comprehensive expertise in technology product development, integrating data-driven decision-making to optimize Product Development, Supply Chain Management, and Manufacturing.\u003c/p\u003e","title":"About"}]