把追蹤全球總體經濟的判斷邏輯整理成一套框架庫——不是在猜「央行下次會不會降息」，而是在追問「這則消息真正該問的問題是什麼」。20 個框架涵蓋央行政策與獨立性風險、黃金與避險資產順位、石油地緣政治風險溢價、美元與去美元化敘事，每一個都對應到真實案例：2024 年日圓套利交易崩盤衝擊全球股匯市、2025 年黃金取代公債成為市場首選避險資產、日本主權債務結構為何未爆發危機等。
Global Macro Economy
這個知識圖譜是一套用來解讀全球總體經濟新聞的分析框架庫，由資深總體經濟分析師建立。內容涵蓋央行政策與獨立性、黃金與避險資產、石油與地緣政治風險、美元與貨幣體系等 20 個判斷法則，每一個都對應到真實、有時間點的案例（2024 年日圓套利交易崩盤、2025 年黃金取代公債成為首選避險資產、日本主權債務結構等）。當使用者看到一則央行決議、油價暴漲或地緣政治衝突的新聞，想知道這則消息背後真正該追問的問題是什麼時，這個圖譜最有幫助。
英特爾 18A 良率卡關、蘋果自研晶片持續蠶食供應商訂單、新台幣升值壓縮毛利——這些新聞大家都看得到，但多數人只停在「發生了什麼」，沒有往下追問「這對護城河跟訂價權意味著什麼」。 我把這幾年看台灣科技與半導體供應鏈的判斷邏輯整理成 20 個框架：晶圓代工護城河怎麼算、先進封裝瓶頸卡在哪一段、地緣政治風險怎麼定價、人才磁吸效應多強、資本支出周期跟匯率曝險又是怎麼互相牽動的。每個框架都對應一個真實案例，包括台積電 CoWoS 產能吃緊、英特爾 18A 良率困境、蘋果自研晶片、新台幣急升衝擊毛利率——不是憑空的理論。
Taiwan Tech Stack
這個知識圖譜是一套用來解讀台灣科技與半導體產業新聞的分析框架庫，由專注台灣供應鏈的產業分析師建立。內容涵蓋晶圓代工護城河、先進封裝瓶頸、地緣政治風險、人才流動、資本支出與匯率曝險等 20 個判斷法則，每一個都對應到真實、有時間點的案例（台積電 CoWoS 產能、英特爾良率困境、蘋果自研晶片等）。當使用者看到一則台灣科技股相關新聞，想知道這則消息背後真正該追問的問題是什麼時，這個圖譜最有幫助。
Defense investment
Defense investment
This knowledge graph is an investment research workspace focused on autonomous defense systems — specifically AI pilots, unmanned aerial vehicles, and multi-agent teaming software. It is most useful for answering questions about the defense autonomy investment landscape: which technologies (edge AI, GNSS-denied operations, mission autonomy) are driving procurement, which countries and programs are accelerating spending, and how companies like Shield AI position their hardware/software stacks against defense customer requirements. The graph contains structured program-level deal data (customer, platform, technology), competitive benchmarking against public-market peers, demand-driver analysis tied to real budget events, and explicit investment thesis frameworks including risk catalysts and valuation considerations.
Defense technology stack
Defense
Spent the last while building out a framework library for reading US software industry news — not "what happened," but "what should I ask to know if this actually matters." 20 interlinked models covering moat durability, pricing signals, M&A/antitrust patterns, talent signals, and AI capex risk, each anchored to a real, dated case (Adobe/Figma, the Slack-Teams antitrust fight, the MongoDB→Redis relicensing wave, OpenAI's board crisis, and more). Wired it up as an MCP server, so any AI agent can query it directly instead of me re-explaining the same frameworks every time a new headline drops. Public graph + connect URL below.
US Software & SaaS
This knowledge graph is a framework library for reasoning about the US software industry, built by a financial analyst covering software and SaaS. Rather than tracking company news or facts, it encodes 20 decision-rule frameworks — moat durability, pricing-model signals, M&A/antitrust patterns, talent/org signals, AI capex and valuation dynamics — each anchored to a real, dated case study. It's most useful when you have a piece of tech news and want to know what questions to ask about its downstream impact, not just what happened.
Mapped out Vincent Cheng-Wen Yu's investment writing into a knowledge graph — moat durability, De Beers, capital allocation risk at Meta, commitment premium, and the "market whale" myth all connected. Ran it against a baseline agent with no context: graph-augmented won 5/5 comparisons, scoring 5.0/5.0/5.0 vs. baseline's 3.0/3.0/2.8 on groundedness, framework-consistency, and specificity. Connect your research agent to it via MCP and see for yourself.
Vincent Cheng-Wen Yu's Investment Frameworks
This knowledge graph captures a Chinese-language analyst's (Vincent Cheng-Wen Yu) original investment and business strategy frameworks, built around case studies like Meta's AI capex cycle, De Beers' collapsing moat, and the Hunt Brothers silver corner. It is most useful for answering questions about moat durability and definitional power, how markets price capital allocation risk in founder-led companies, what makes a commitment credible vs. a sunk cost, how to diagnose whether a "market whale/manipulator" narrative is real or a psychological artifact, and how companies can reduce investor uncertainty by making opaque internal bets externally verifiable.
Have your own knowledge graph? Upload one and share it.
@mimi
@annie
@derek
@reggiehsu111