<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:dc="http://purl.org/dc/elements/1.1/">
<channel>
  <title>AI 武林：LangGraph</title>
  <link>https://aiwulin.itsmygo.uk/tools/langgraph/</link>
  <description>AI 武林收錄的內容裡，最新提到 LangGraph 的 28 筆，每筆附中文摘要。</description>
  <language>zh-TW</language>
  <lastBuildDate>Thu, 08 Oct 2026 17:22:41 GMT</lastBuildDate>
  <atom:link href="https://aiwulin.itsmygo.uk/tools/langgraph/rss.xml" rel="self" type="application/rss+xml"/>
  <item>
    <title>Why Your AI Agents Can't Talk to Each Other (Yet) — Vlad Luzin, BAND</title>
    <link>https://aiwulin.itsmygo.uk/c/24f09b180e/</link>
    <guid isPermaLink="false">aiwulin-24f09b180e</guid>
    <pubDate>Thu, 08 Oct 2026 00:00:00 +0800</pubDate>
    <dc:creator>AI Engineer</dc:creator>
    <description>&lt;p&gt;Vlad Luzin 指出人類目前像路由器一樣在多個 AI 之間傳遞訊息，他認為建立 AI 與 AI 間的通訊層能解決單一代理的偏見與注意力分散問題。BAND 提供包含身份、登錄檔與即時對話的互動層，讓代理能跨組織協作並保留人類參與。&lt;/p&gt;&lt;p&gt;原站：&lt;a href=&quot;https://www.youtube.com/watch?v=toq-jyGLZDk&quot;&gt;https://www.youtube.com/watch?v=toq-jyGLZDk&lt;/a&gt;&lt;/p&gt;</description>
    <category>影片</category>
    <category>AI Agent 基礎</category>
  </item>
  <item>
    <title>How Zip Uses LangSmith and LangGraph to Ship New Features Faster</title>
    <link>https://aiwulin.itsmygo.uk/c/3d126cb7fd/</link>
    <guid isPermaLink="false">aiwulin-3d126cb7fd</guid>
    <pubDate>Tue, 06 Oct 2026 00:00:00 +0800</pubDate>
    <dc:creator>LangChain</dc:creator>
    <description>&lt;p&gt;Zip 團隊分享如何透過 LangGraph 與 LangSmith 加速 AI 功能開發與追蹤。透過這些工具，團隊無需額外編碼即可獲得節點互動與 LLM 呼叫的追蹤資料，並建立評估機制防止回退問題。&lt;/p&gt;&lt;p&gt;原站：&lt;a href=&quot;https://www.youtube.com/watch?v=sGrD9a8dh_0&quot;&gt;https://www.youtube.com/watch?v=sGrD9a8dh_0&lt;/a&gt;&lt;/p&gt;</description>
    <category>影片</category>
    <category>開發框架</category>
    <category>可觀測性與 LLMOps</category>
  </item>
  <item>
    <title>The Human Is an Async API — Melanie Warrick, Temporal</title>
    <link>https://aiwulin.itsmygo.uk/c/8e675b4784/</link>
    <guid isPermaLink="false">aiwulin-8e675b4784</guid>
    <pubDate>Mon, 05 Oct 2026 00:00:00 +0800</pubDate>
    <dc:creator>AI Engineer</dc:creator>
    <description>&lt;p&gt;Melanie Warrick 透過冰淇淋配送示範，說明如何建立不阻塞且能持久化狀態的人機迴圈代理系統。她展示如何利用 Temporal 的機制讓代理在等待人類決策時不中斷其他工作，並從失敗中恢復。&lt;/p&gt;&lt;p&gt;原站：&lt;a href=&quot;https://www.youtube.com/watch?v=jc3kbZkuHTo&quot;&gt;https://www.youtube.com/watch?v=jc3kbZkuHTo&lt;/a&gt;&lt;/p&gt;</description>
    <category>影片</category>
    <category>AI Agent 基礎</category>
  </item>
  <item>
    <title>Interrupt NYC: Opening Keynote</title>
    <link>https://aiwulin.itsmygo.uk/c/6f7fa0b31d/</link>
    <guid isPermaLink="false">aiwulin-6f7fa0b31d</guid>
    <pubDate>Mon, 05 Oct 2026 00:00:00 +0800</pubDate>
    <dc:creator>LangChain</dc:creator>
    <description>&lt;p&gt;Harrison Chase 分享如何透過 LangGraph 與 LangSmith 建構領域專屬智慧，並介紹 Managed Deep Agents、LangSmith Engine v2 等最新功能與觀察工具。&lt;/p&gt;&lt;p&gt;原站：&lt;a href=&quot;https://www.youtube.com/watch?v=950byF7njfw&quot;&gt;https://www.youtube.com/watch?v=950byF7njfw&lt;/a&gt;&lt;/p&gt;</description>
    <category>影片</category>
    <category>開發框架</category>
  </item>
  <item>
    <title>Stellantis: How a Global Automaker Deployed AI Agents at Scale</title>
    <link>https://aiwulin.itsmygo.uk/c/c1e18a4e3d/</link>
    <guid isPermaLink="false">aiwulin-c1e18a4e3d</guid>
    <pubDate>Fri, 02 Oct 2026 00:00:00 +0800</pubDate>
    <dc:creator>Databricks</dc:creator>
    <description>&lt;p&gt;Stellantis 透過 Databricks 架構部署 AI 代理，將供應鏈規劃時間縮短至一天並節省成本。影片介紹了 Optima 系統的生產架構與大型企業部署多代理系統的經驗。&lt;/p&gt;&lt;p&gt;原站：&lt;a href=&quot;https://www.youtube.com/watch?v=HEv_AYi1QAc&quot;&gt;https://www.youtube.com/watch?v=HEv_AYi1QAc&lt;/a&gt;&lt;/p&gt;</description>
    <category>影片</category>
    <category>AI Agent 基礎</category>
  </item>
  <item>
    <title>Why Your AI Agent Fails in Production (And How to Catch It)</title>
    <link>https://aiwulin.itsmygo.uk/c/c80507865d/</link>
    <guid isPermaLink="false">aiwulin-c80507865d</guid>
    <pubDate>Wed, 30 Sep 2026 00:00:00 +0800</pubDate>
    <dc:creator>Google Cloud Tech</dc:creator>
    <description>&lt;p&gt;展示如何建立端到端的 AI 代理評估管道，將本地演示中的問題轉化為生產環境的實際測試。觀眾能學習使用 Antigravity、OpenTelemetry 等工具標準化追蹤並設定自動化評分標準。&lt;/p&gt;&lt;p&gt;原站：&lt;a href=&quot;https://www.youtube.com/watch?v=wPdoZRbvaF4&quot;&gt;https://www.youtube.com/watch?v=wPdoZRbvaF4&lt;/a&gt;&lt;/p&gt;</description>
    <category>影片</category>
    <category>AI Agent 基礎</category>
    <category>AI 評測</category>
  </item>
  <item>
    <title>Your Agents Are in Solitary Confinement: Why MCP &amp; A2A Aren't Enough — Vlad Luzin, Band</title>
    <link>https://aiwulin.itsmygo.uk/c/d97ab48b8d/</link>
    <guid isPermaLink="false">aiwulin-d97ab48b8d</guid>
    <pubDate>Wed, 30 Sep 2026 00:00:00 +0800</pubDate>
    <dc:creator>AI Engineer</dc:creator>
    <description>&lt;p&gt;Vlad Luzin 指出當前 MCP 與 A2A 協議無法解決多代理協作的狀態保持與發現問題，導致代理陷入「數位孤立監禁」。&lt;/p&gt;&lt;p&gt;原站：&lt;a href=&quot;https://www.youtube.com/watch?v=UOcHfR3_tys&quot;&gt;https://www.youtube.com/watch?v=UOcHfR3_tys&lt;/a&gt;&lt;/p&gt;</description>
    <category>影片</category>
    <category>AI Agent 基礎</category>
    <category>MCP 模型上下文協定</category>
    <category>多代理系統</category>
  </item>
  <item>
    <title>How to Build AI Agents in Python - 3 Ways</title>
    <link>https://aiwulin.itsmygo.uk/c/029191a38a/</link>
    <guid isPermaLink="false">aiwulin-029191a38a</guid>
    <pubDate>Wed, 23 Sep 2026 00:00:00 +0800</pubDate>
    <dc:creator>Tech With Tim</dc:creator>
    <description>&lt;p&gt;介紹如何用 CrewAI、OpenAI Agents SDK 與 LangGraph 三種方式在 Python 建構 AI Agent。透過實際程式碼示範，比較各框架在簡單原型、功能完整性與生產級複雜度上的差異，幫助讀者選擇合適工具。&lt;/p&gt;&lt;p&gt;原站：&lt;a href=&quot;https://www.youtube.com/watch?v=-RTgK6qX6A8&quot;&gt;https://www.youtube.com/watch?v=-RTgK6qX6A8&lt;/a&gt;&lt;/p&gt;</description>
    <category>影片</category>
    <category>開發框架</category>
    <category>AI Agent 基礎</category>
    <category>Coding Agent</category>
  </item>
  <item>
    <title>Middleware for Managed Deep Agents</title>
    <link>https://aiwulin.itsmygo.uk/c/ea03aae3f3/</link>
    <guid isPermaLink="false">aiwulin-ea03aae3f3</guid>
    <pubDate>Thu, 17 Sep 2026 00:00:00 +0800</pubDate>
    <dc:creator>LangChain</dc:creator>
    <description>&lt;p&gt;LangChain 產品經理 Nathan Drezner 介紹管理式深度代理的中介層，示範如何加入預設的 PII 中介層來遮蔽客戶郵件，並從零編寫自訂的日誌中介層。觀眾能學習如何為代理擴展生命週期，加入政策執行、容錯與速率限制等自訂行為。&lt;/p&gt;&lt;p&gt;原站：&lt;a href=&quot;https://www.youtube.com/watch?v=d6aNbE-3dxo&quot;&gt;https://www.youtube.com/watch?v=d6aNbE-3dxo&lt;/a&gt;&lt;/p&gt;</description>
    <category>影片</category>
    <category>AI Agent 基礎</category>
    <category>開發框架</category>
  </item>
  <item>
    <title>How Lyft Increased Its Agent Resolution Rate by 16% with LangSmith and LangGraph</title>
    <link>https://aiwulin.itsmygo.uk/c/3876f1f3dd/</link>
    <guid isPermaLink="false">aiwulin-3876f1f3dd</guid>
    <pubDate>Wed, 16 Sep 2026 00:00:00 +0800</pubDate>
    <dc:creator>LangChain</dc:creator>
    <description>&lt;p&gt;Lyft 工程師分享如何利用 LangGraph 與 LangSmith 重建客服代理架構，將開發時間從六個月縮短至兩週，並將解決率提升 16%。透過元代理架構與可觀測性工具，讓產品經理與營運人員無需寫程式即可快速部署新代理。&lt;/p&gt;&lt;p&gt;原站：&lt;a href=&quot;https://www.youtube.com/watch?v=M9BMTC8o9-w&quot;&gt;https://www.youtube.com/watch?v=M9BMTC8o9-w&lt;/a&gt;&lt;/p&gt;</description>
    <category>影片</category>
    <category>開發框架</category>
    <category>可觀測性與 LLMOps</category>
  </item>
  <item>
    <title>Share your Managed Deep Agent with your team using Slack</title>
    <link>https://aiwulin.itsmygo.uk/c/c15cad5a83/</link>
    <guid isPermaLink="false">aiwulin-c15cad5a83</guid>
    <pubDate>Tue, 15 Sep 2026 00:00:00 +0800</pubDate>
    <dc:creator>LangChain</dc:creator>
    <description>&lt;p&gt;示範如何將 LangSmith 中的研究助理代理部署到 Slack 團隊環境，包含設定連線、授權、監控追蹤及自訂外觀。看完即可掌握將 AI 代理整合至 Slack 並讓團隊使用的具體步驟。&lt;/p&gt;&lt;p&gt;原站：&lt;a href=&quot;https://www.youtube.com/watch?v=a5Yie-Bgx7A&quot;&gt;https://www.youtube.com/watch?v=a5Yie-Bgx7A&lt;/a&gt;&lt;/p&gt;</description>
    <category>影片</category>
    <category>AI Agent 基礎</category>
    <category>可觀測性與 LLMOps</category>
  </item>
  <item>
    <title>Build fences not sandboxes, earn the currency of trust, and share the cognitive burden of agents across engineers</title>
    <link>https://aiwulin.itsmygo.uk/c/3ac49dc628/</link>
    <guid isPermaLink="false">aiwulin-3ac49dc628</guid>
    <pubDate>Sat, 29 Aug 2026 00:00:00 +0800</pubDate>
    <dc:creator>Dev Interrupted</dc:creator>
    <description>&lt;p&gt;探討 Ox Alpha 等黑箱模型的風險、工程信任的貨幣價值、多代理圖形架構以及資深工程師如何轉型為 Staff Engineer。透過實例與專家對談，提供建立安全邊界、最佳化認知負荷及跨團隊協作的具體策略。&lt;/p&gt;&lt;p&gt;原站：&lt;a href=&quot;https://www.buzzsprout.com/1422892/episodes/19717515-build-fences-not-sandboxes-earn-the-currency-of-trust-and-share-the-cognitive-burden-of-agents-across-engineers&quot;&gt;https://www.buzzsprout.com/1422892/episodes/19717515-build-fences-not-sandboxes-earn-the-currency-of-trust-and-share-the-cognitive-burden-of-agents-across-engineers&lt;/a&gt;&lt;/p&gt;</description>
    <category>Podcast</category>
    <category>AI Agent 基礎</category>
    <category>Coding Agent</category>
    <category>提示工程</category>
  </item>
  <item>
    <title>System Design for AI Agents – Building a Multi-Agent PR Reviewer</title>
    <link>https://aiwulin.itsmygo.uk/c/be50944285/</link>
    <guid isPermaLink="false">aiwulin-be50944285</guid>
    <pubDate>Fri, 14 Aug 2026 00:00:00 +0800</pubDate>
    <dc:creator>freeCodeCamp.org</dc:creator>
    <description>&lt;p&gt;詳細教學如何設計並構建一個生產級的多代理 AI 系統，用於自動審查程式碼合併請求。內容涵蓋系統架構、資料庫選擇、故障容錯機制與驗證流程，讓讀者掌握將 AI 整合進工程實作的完整方法。&lt;/p&gt;&lt;p&gt;原站：&lt;a href=&quot;https://www.youtube.com/watch?v=iqRcGCah0Kw&quot;&gt;https://www.youtube.com/watch?v=iqRcGCah0Kw&lt;/a&gt;&lt;/p&gt;</description>
    <category>影片</category>
    <category>多代理系統</category>
    <category>AI Agent 基礎</category>
  </item>
  <item>
    <title>Why Graph Engineering will 10x your Claude/Codex</title>
    <link>https://aiwulin.itsmygo.uk/c/48ecbcdcb3/</link>
    <guid isPermaLink="false">aiwulin-48ecbcdcb3</guid>
    <pubDate>Tue, 04 Aug 2026 00:00:00 +0800</pubDate>
    <dc:creator>Greg Isenberg</dc:creator>
    <description>&lt;p&gt;深入解析「圖工程」概念，說明如何將單一大對話轉換為結構化的工作流圖，透過規劃、並行研究、懷疑者檢查與合併等步驟提升 AI 產出品質。看完後讀者能學會如何將現有任務拆解成可管理的步驟，並加入人工審核機制以減少錯誤。&lt;/p&gt;&lt;p&gt;原站：&lt;a href=&quot;https://www.youtube.com/watch?v=JWhICz1QR8M&quot;&gt;https://www.youtube.com/watch?v=JWhICz1QR8M&lt;/a&gt;&lt;/p&gt;</description>
    <category>影片</category>
    <category>工作流自動化</category>
    <category>Coding Agent</category>
    <category>開發框架</category>
  </item>
  <item>
    <title>not much happened today</title>
    <link>https://aiwulin.itsmygo.uk/c/566e9a667b/</link>
    <guid isPermaLink="false">aiwulin-566e9a667b</guid>
    <pubDate>Fri, 31 Jul 2026 00:00:00 +0800</pubDate>
    <dc:creator>AINews（Latent Space／smol.ai）</dc:creator>
    <description>&lt;p&gt;DeepSeek 推出 V4-Flash API，在架構與參數規模不變的情況下，透過後訓練最佳化大幅提升效能，Terminal-Bench 分數達 82.7，成本比 GPT-5.6 Luna 低約 60%，並立即以 MIT 授權開放權重。&lt;/p&gt;&lt;p&gt;原站：&lt;a href=&quot;https://news.smol.ai/issues/26-07-31-not-much&quot;&gt;https://news.smol.ai/issues/26-07-31-not-much&lt;/a&gt;&lt;/p&gt;</description>
    <category>文章</category>
    <category>開源模型</category>
    <category>模型發布與實測</category>
  </item>
  <item>
    <title>Agentic AI – Complete Course for Beginners</title>
    <link>https://aiwulin.itsmygo.uk/c/eeda2c0cc9/</link>
    <guid isPermaLink="false">aiwulin-eeda2c0cc9</guid>
    <pubDate>Thu, 30 Jul 2026 00:00:00 +0800</pubDate>
    <dc:creator>freeCodeCamp.org</dc:creator>
    <description>&lt;p&gt;教導如何從零開始構建生產級多代理系統，掌握 LangChain 與 LangGraph 的核心概念與實作技巧。學員將學習如何設計序列、並行與條件工作流，整合 RAG、記憶機制與人工介入控制，並透過實際專案將應用程式部署至 AWS 與…&lt;/p&gt;&lt;p&gt;原站：&lt;a href=&quot;https://www.youtube.com/watch?v=Zy7EXDONlTY&quot;&gt;https://www.youtube.com/watch?v=Zy7EXDONlTY&lt;/a&gt;&lt;/p&gt;</description>
    <category>影片</category>
    <category>AI Agent 基礎</category>
    <category>開發框架</category>
    <category>多代理系統</category>
  </item>
  <item>
    <title>not much happened today</title>
    <link>https://aiwulin.itsmygo.uk/c/30ed204412/</link>
    <guid isPermaLink="false">aiwulin-30ed204412</guid>
    <pubDate>Mon, 20 Jul 2026 00:00:00 +0800</pubDate>
    <dc:creator>AINews（Latent Space／smol.ai）</dc:creator>
    <description>&lt;p&gt;匯總了 2026 年 7 月 AI 領域的最新動態，涵蓋中美 AI 政策博弈、開源模型競爭及基礎設施發展。內容重點包括美國可能限制中國開源模型、Hugging Face 因安全事件轉向自部署 GLM-5.2、Kimi K3 與 Qwen…&lt;/p&gt;&lt;p&gt;原站：&lt;a href=&quot;https://news.smol.ai/issues/26-07-20-not-much&quot;&gt;https://news.smol.ai/issues/26-07-20-not-much&lt;/a&gt;&lt;/p&gt;</description>
    <category>文章</category>
    <category>開源模型</category>
    <category>AI 安全與治理</category>
  </item>
  <item>
    <title>The Control-vs-Magic Spectrum Building Agents</title>
    <link>https://aiwulin.itsmygo.uk/c/54cb118c0d/</link>
    <guid isPermaLink="false">aiwulin-54cb118c0d</guid>
    <pubDate>Fri, 05 Jun 2026 00:00:00 +0800</pubDate>
    <dc:creator>Agentic Conversations（原 MLOps.community）</dc:creator>
    <description>&lt;p&gt;訪談 iFood 資料與 AI 總監 Thiago Cardoso，分享如何將 AI 代理應用於大規模生產環境。內容涵蓋從控制到自動化的譜系、多代理架構設計、語境工程的重要性，以及利用 WhatsApp 作為介面提供個人化金融服務。&lt;/p&gt;&lt;p&gt;原站：&lt;a href=&quot;https://podcasters.spotify.com/pod/show/mlops/episodes/The-Control-vs-Magic-Spectrum-Building-Agents-e3k9n0g&quot;&gt;https://podcasters.spotify.com/pod/show/mlops/episodes/The-Control-vs-Magic-Spectrum-Building-Agents-e3k9n0g&lt;/a&gt;&lt;/p&gt;</description>
    <category>Podcast</category>
    <category>AI Agent 基礎</category>
  </item>
  <item>
    <title>The Age of Async Agents — Cognition's Walden Yan &amp; OpenInspect's Cole Murray</title>
    <link>https://aiwulin.itsmygo.uk/c/cf994e29e8/</link>
    <guid isPermaLink="false">aiwulin-cf994e29e8</guid>
    <pubDate>Fri, 29 May 2026 00:00:00 +0800</pubDate>
    <dc:creator>Latent Space</dc:creator>
    <description>&lt;p&gt;訪談 Cognition 創辦人 Walden Yan 與 OpenInspect 創辦人 Cole Murray，探討背景代理（Background Agents）如何從概念走向實用，以及 Devin 等工具在 2025 年底後的演變。&lt;/p&gt;&lt;p&gt;原站：&lt;a href=&quot;https://www.latent.space/p/cognition&quot;&gt;https://www.latent.space/p/cognition&lt;/a&gt;&lt;/p&gt;</description>
    <category>Podcast</category>
    <category>AI Agent 基礎</category>
    <category>Coding Agent</category>
  </item>
  <item>
    <title>深度拆解：AI Agent Harness 的构造</title>
    <link>https://aiwulin.itsmygo.uk/c/08dc63f309/</link>
    <guid isPermaLink="false">aiwulin-08dc63f309</guid>
    <pubDate>Sun, 10 May 2026 00:00:00 +0800</pubDate>
    <dc:creator>寶玉</dc:creator>
    <description>&lt;p&gt;深入解析將大型語言模型轉化為生產級 AI Agent 所需的「Agent Harness」架構，涵蓋編排迴圈、工具、記憶與上下文管理等核心元件。&lt;/p&gt;&lt;p&gt;原站：&lt;a href=&quot;https://baoyu.io/translations/2026-05-10/akshay-pachaar-2041146899319971922&quot;&gt;https://baoyu.io/translations/2026-05-10/akshay-pachaar-2041146899319971922&lt;/a&gt;&lt;/p&gt;</description>
    <category>文章</category>
    <category>深度學習</category>
  </item>
  <item>
    <title>【十字路口】探秘 Claude Code，搞懂 Agent Harness｜对谈来新璐【视频播客】</title>
    <link>https://aiwulin.itsmygo.uk/c/56a225c0f7/</link>
    <guid isPermaLink="false">aiwulin-56a225c0f7</guid>
    <pubDate>Wed, 06 May 2026 00:00:00 +0800</pubDate>
    <dc:creator>Koji杨远骋</dc:creator>
    <description>&lt;p&gt;訪談拆解 Agent Harness 的執行、狀態與治理三層架構，並透過 Claude Code 原始碼分析其上下文壓縮與記憶機制。內容涵蓋從 LangChain 到原生 Agent 的開發趨勢、沙箱技術演進，以及未來「零人公司」的構想。&lt;/p&gt;&lt;p&gt;原站：&lt;a href=&quot;https://www.youtube.com/watch?v=FZO0eLl5jqo&quot;&gt;https://www.youtube.com/watch?v=FZO0eLl5jqo&lt;/a&gt;&lt;/p&gt;</description>
    <category>影片</category>
    <category>Claude Code</category>
    <category>Agent Skills 與 Harness</category>
    <category>AI Agent 基礎</category>
  </item>
  <item>
    <title>Ep 85: Has AI Infra Stabilized, FM Vibe Shift, &amp; What's Next for Coding Agents</title>
    <link>https://aiwulin.itsmygo.uk/c/43468d9e28/</link>
    <guid isPermaLink="false">aiwulin-43468d9e28</guid>
    <pubDate>Thu, 23 Apr 2026 00:00:00 +0800</pubDate>
    <dc:creator>Unsupervised Learning</dc:creator>
    <description>&lt;p&gt;訪談探討 AI 基礎設施是否已穩定、程式碼代理戰爭的現狀與未來，以及企業自建模型與強化學習的價值。兩人深入分析從 LangChain 到 DeepAgents 的架構演變，討論 Claude Code 與 Codex 的競爭，並展望世界模…&lt;/p&gt;&lt;p&gt;原站：&lt;a href=&quot;https://unsupervised-learning.simplecast.com/episodes/ep-85-has-ai-infra-stabilized-fm-vibe-shift-whats-next-for-coding-agents-IjTaA0mQ&quot;&gt;https://unsupervised-learning.simplecast.com/episodes/ep-85-has-ai-infra-stabilized-fm-vibe-shift-whats-next-for-coding-agents-IjTaA0mQ&lt;/a&gt;&lt;/p&gt;</description>
    <category>Podcast</category>
    <category>Coding Agent</category>
    <category>AI Agent 基礎</category>
  </item>
  <item>
    <title>Everything Gets Rebuilt: The New AI Agent Stack | Harrison Chase, LangChain</title>
    <link>https://aiwulin.itsmygo.uk/c/75b9327745/</link>
    <guid isPermaLink="false">aiwulin-75b9327745</guid>
    <pubDate>Thu, 12 Mar 2026 00:00:00 +0800</pubDate>
    <dc:creator>The MAD Podcast</dc:creator>
    <description>&lt;p&gt;Harrison Chase 與 LangChain 共同創辦人深入探討 AI 代理（Agent）如何從簡單提示演變為具備規劃、編碼與工具使用能力的複雜系統。&lt;/p&gt;&lt;p&gt;原站：&lt;a href=&quot;https://podcasters.spotify.com/pod/show/firstmark/episodes/Everything-Gets-Rebuilt-The-New-AI-Agent-Stack--Harrison-Chase--LangChain-e3gapv9&quot;&gt;https://podcasters.spotify.com/pod/show/firstmark/episodes/Everything-Gets-Rebuilt-The-New-AI-Agent-Stack--Harrison-Chase--LangChain-e3gapv9&lt;/a&gt;&lt;/p&gt;</description>
    <category>Podcast</category>
    <category>AI Agent 基礎</category>
    <category>開發框架</category>
  </item>
  <item>
    <title>Episode 71: Durable Agents - How to Build AI Systems That Survive a Crash with Samuel Colvin</title>
    <link>https://aiwulin.itsmygo.uk/c/506d109689/</link>
    <guid isPermaLink="false">aiwulin-506d109689</guid>
    <pubDate>Wed, 18 Feb 2026 00:00:00 +0800</pubDate>
    <dc:creator>Vanishing Gradients</dc:creator>
    <description>&lt;p&gt;探討如何將軟體工程原則應用於 AI 代理，以建立能從失敗中恢復的「耐損代理」。講者 Samuel Colvin 強調生產環境需要高可靠性，並介紹了將代理拆分為確定性工作流與隨機活動的架構，利用 Temporal 等工具實現狀態持久化與重試…&lt;/p&gt;&lt;p&gt;原站：&lt;a href=&quot;https://hugobowne.substack.com/p/episode-71-durable-agents-how-to&quot;&gt;https://hugobowne.substack.com/p/episode-71-durable-agents-how-to&lt;/a&gt;&lt;/p&gt;</description>
    <category>Podcast</category>
    <category>AI Agent 基礎</category>
    <category>Coding Agent</category>
  </item>
  <item>
    <title>How to write a good spec for AI agents</title>
    <link>https://aiwulin.itsmygo.uk/c/659ca50d65/</link>
    <guid isPermaLink="false">aiwulin-659ca50d65</guid>
    <pubDate>Mon, 19 Jan 2026 00:00:00 +0800</pubDate>
    <dc:creator>Addy Osmani（部落格）</dc:creator>
    <description>&lt;p&gt;提供撰寫高績效 AI 程式碼代理規格書的實戰框架，涵蓋從高階願景到執行細節的規劃步驟。讀者可學習如何將大型任務拆解、設定執行命令與測試規範，並利用 Plan Mode 避免過度工程化，確保 AI 專注且高效產出。&lt;/p&gt;&lt;p&gt;原站：&lt;a href=&quot;https://addyo.substack.com/p/how-to-write-a-good-spec-for-ai-agents&quot;&gt;https://addyo.substack.com/p/how-to-write-a-good-spec-for-ai-agents&lt;/a&gt;&lt;/p&gt;</description>
    <category>文章</category>
    <category>AI Agent 基礎</category>
    <category>Coding Agent</category>
  </item>
  <item>
    <title>How New Libraries Saw a 50% Improvement | Maria Gorinova</title>
    <link>https://aiwulin.itsmygo.uk/c/ac0d6585bc/</link>
    <guid isPermaLink="false">aiwulin-ac0d6585bc</guid>
    <pubDate>Tue, 09 Dec 2025 00:00:00 +0800</pubDate>
    <dc:creator>The AI Native Dev</dc:creator>
    <description>&lt;p&gt;訪談 Tesla 工程師 Maria Gorinova，探討 AI 代理在程式開發中對函式庫抽象的誤用問題。透過 Tesla Registry 與 Tiles 技術，提供上下文讓代理更精準理解函式庫，測試顯示新函式庫使用率提升 50%。&lt;/p&gt;&lt;p&gt;原站：&lt;a href=&quot;https://www.buzzsprout.com/2375985/episodes/18324020-how-new-libraries-saw-a-50-improvement-maria-gorinova&quot;&gt;https://www.buzzsprout.com/2375985/episodes/18324020-how-new-libraries-saw-a-50-improvement-maria-gorinova&lt;/a&gt;&lt;/p&gt;</description>
    <category>Podcast</category>
    <category>Coding Agent</category>
    <category>AI Agent 基礎</category>
    <category>AI 輔助軟體工程</category>
  </item>
  <item>
    <title>Episode 61: The AI Agent Reliability Cliff: What Happens When Tools Fail in Production</title>
    <link>https://aiwulin.itsmygo.uk/c/85e6eee69f/</link>
    <guid isPermaLink="false">aiwulin-85e6eee69f</guid>
    <pubDate>Thu, 16 Oct 2025 00:00:00 +0800</pubDate>
    <dc:creator>Vanishing Gradients</dc:creator>
    <description>&lt;p&gt;Alex Strick Van Linschoten 分享從 LLM Ops Database 累積的真實案例，指出多代理系統若過度複雜會陷入混亂。&lt;/p&gt;&lt;p&gt;原站：&lt;a href=&quot;https://hugobowne.substack.com/p/episode-61-the-ai-agent-reliability-2a3&quot;&gt;https://hugobowne.substack.com/p/episode-61-the-ai-agent-reliability-2a3&lt;/a&gt;&lt;/p&gt;</description>
    <category>Podcast</category>
    <category>AI Agent 基礎</category>
    <category>MLOps</category>
    <category>多代理系統</category>
  </item>
  <item>
    <title>[Livestream] CrewAI vs LangGraph vs AutoGen</title>
    <link>https://aiwulin.itsmygo.uk/c/9e3f75c2fc/</link>
    <guid isPermaLink="false">aiwulin-9e3f75c2fc</guid>
    <pubDate>Sat, 17 Feb 2024 00:00:00 +0800</pubDate>
    <dc:creator>Shawn Wang (swyx)</dc:creator>
    <description>&lt;p&gt;比較 CrewAI、LangGraph 與 AutoGen 這三個 AI 程式設計框架的差異與應用。&lt;/p&gt;&lt;p&gt;原站：&lt;a href=&quot;https://www.youtube.com/watch?v=S6MtNDIm3oc&quot;&gt;https://www.youtube.com/watch?v=S6MtNDIm3oc&lt;/a&gt;&lt;/p&gt;</description>
    <category>影片</category>
    <category>開發框架</category>
  </item>
</channel>
</rss>
