<?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：Kubernetes</title>
  <link>https://aiwulin.itsmygo.uk/tools/kubernetes/</link>
  <description>AI 武林收錄的內容裡，最新提到 Kubernetes 的 30 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/kubernetes/rss.xml" rel="self" type="application/rss+xml"/>
  <item>
    <title>Building resilient systems with Sam Newman</title>
    <link>https://aiwulin.itsmygo.uk/c/c2e9218036/</link>
    <guid isPermaLink="false">aiwulin-c2e9218036</guid>
    <pubDate>Thu, 08 Oct 2026 00:00:00 +0800</pubDate>
    <dc:creator>The Pragmatic Engineer</dc:creator>
    <description>&lt;p&gt;Sam Newman 探討微服務架構的誤區與新書《建構高韌性分散式系統》中的三大原則，強調資訊傳遞時延、資源有限與服務獨立部署的重要性。&lt;/p&gt;&lt;p&gt;原站：&lt;a href=&quot;https://newsletter.pragmaticengineer.com/p/building-resilient-systems-with-sam&quot;&gt;https://newsletter.pragmaticengineer.com/p/building-resilient-systems-with-sam&lt;/a&gt;&lt;/p&gt;</description>
    <category>Podcast</category>
    <category>AI 輔助軟體工程</category>
  </item>
  <item>
    <title>Production Monitoring with Codex: Grafana, Kubernetes, &amp; Security</title>
    <link>https://aiwulin.itsmygo.uk/c/7bfd3a5299/</link>
    <guid isPermaLink="false">aiwulin-7bfd3a5299</guid>
    <pubDate>Wed, 07 Oct 2026 00:00:00 +0800</pubDate>
    <dc:creator>OpenAI</dc:creator>
    <description>&lt;p&gt;演示如何利用 Codex 輔助在 Grafana 與 Kubernetes 環境進行生產監控與故障排查，涵蓋記憶體溢位重啟與安全發現的調查流程。觀眾可學習工程師如何結合遙測資料與程式碼支援來驗證變更結果。&lt;/p&gt;&lt;p&gt;原站：&lt;a href=&quot;https://www.youtube.com/watch?v=DmCaBWv5yG8&quot;&gt;https://www.youtube.com/watch?v=DmCaBWv5yG8&lt;/a&gt;&lt;/p&gt;</description>
    <category>影片</category>
    <category>可觀測性與 LLMOps</category>
    <category>Coding Agent</category>
  </item>
  <item>
    <title>Can a Cloud-Native Harness Make Agents Reliable Beyond the Desktop?</title>
    <link>https://aiwulin.itsmygo.uk/c/f2c199a6dd/</link>
    <guid isPermaLink="false">aiwulin-f2c199a6dd</guid>
    <pubDate>Wed, 07 Oct 2026 00:00:00 +0800</pubDate>
    <dc:creator>Latent Space</dc:creator>
    <description>&lt;p&gt;探討如何將程式設計代理（coding agents）從桌面端移至雲端，以解決管理與安全性問題。作者介紹了由 Kubernetes 創始人 Craig McLuckie 和 Joe Beda 開發的 Stacklok 及其產品 Mecatl…&lt;/p&gt;&lt;p&gt;原站：&lt;a href=&quot;https://www.latent.space/p/stacklok&quot;&gt;https://www.latent.space/p/stacklok&lt;/a&gt;&lt;/p&gt;</description>
    <category>文章</category>
    <category>Coding Agent</category>
    <category>AI Agent 基礎</category>
    <category>Agent Skills 與 Harness</category>
  </item>
  <item>
    <title>GPU Died. Training Didn't: Self-Healing Training at Scale — Crusoe</title>
    <link>https://aiwulin.itsmygo.uk/c/08cf6e6515/</link>
    <guid isPermaLink="false">aiwulin-08cf6e6515</guid>
    <pubDate>Sun, 04 Oct 2026 00:00:00 +0800</pubDate>
    <dc:creator>AI Engineer</dc:creator>
    <description>&lt;p&gt;Crusoe 團隊介紹如何將 Slurm 執行於 Kubernetes 之上，以自動化處理大規模 GPU 訓練中的硬體故障。透過 AutoClusters 系統，可在分鐘內自動更換故障節點並從檢查點恢復訓練，無需人工介入。&lt;/p&gt;&lt;p&gt;原站：&lt;a href=&quot;https://www.youtube.com/watch?v=bRGyYaE0lxI&quot;&gt;https://www.youtube.com/watch?v=bRGyYaE0lxI&lt;/a&gt;&lt;/p&gt;</description>
    <category>影片</category>
    <category>AI 晶片與硬體</category>
  </item>
  <item>
    <title>Lessons from Generating 12 Trillion Synthetic Tokens — Bogdan Gaza, DatologyAI</title>
    <link>https://aiwulin.itsmygo.uk/c/0da6961f4e/</link>
    <guid isPermaLink="false">aiwulin-0da6961f4e</guid>
    <pubDate>Sat, 03 Oct 2026 00:00:00 +0800</pubDate>
    <dc:creator>AI Engineer</dc:creator>
    <description>&lt;p&gt;Bogdan Gaza 分享 DatologyAI 如何透過「種子重寫」技術，將合成資料規模從 300 億擴展至約 1200 億 token。&lt;/p&gt;&lt;p&gt;原站：&lt;a href=&quot;https://www.youtube.com/watch?v=FQwTqUmcbRg&quot;&gt;https://www.youtube.com/watch?v=FQwTqUmcbRg&lt;/a&gt;&lt;/p&gt;</description>
    <category>影片</category>
    <category>訓練資料與合成資料</category>
  </item>
  <item>
    <title>AI 代理來襲！台灣 74% 企業推 AI 僅 35% 導入 Kubernetes，Nutanix 揭 IT 雙重負擔</title>
    <link>https://aiwulin.itsmygo.uk/c/d12f348e42/</link>
    <guid isPermaLink="false">aiwulin-d12f348e42</guid>
    <pubDate>Thu, 01 Oct 2026 00:00:00 +0800</pubDate>
    <dc:creator>INSIDE</dc:creator>
    <description>&lt;p&gt;台灣企業推動 AI 意願高，但 Kubernetes 導入率僅 35%，導致 IT 團隊面臨虛擬機器與容器架構並存的雙重維運壓力。&lt;/p&gt;&lt;p&gt;原站：&lt;a href=&quot;https://www.inside.com.tw/article/42539-nutanix-taiwan-kubernetes-ai-survey&quot;&gt;https://www.inside.com.tw/article/42539-nutanix-taiwan-kubernetes-ai-survey&lt;/a&gt;&lt;/p&gt;</description>
    <category>文章</category>
    <category>AI Agent 基礎</category>
    <category>AI 晶片與硬體</category>
    <category>推論與部署</category>
  </item>
  <item>
    <title>SUNK Anywhere: Unified AI training across Kubernetes (GKE, EKS, multi-cloud)</title>
    <link>https://aiwulin.itsmygo.uk/c/dabcb8cfe3/</link>
    <guid isPermaLink="false">aiwulin-dabcb8cfe3</guid>
    <pubDate>Tue, 29 Sep 2026 00:00:00 +0800</pubDate>
    <dc:creator>CoreWeave</dc:creator>
    <description>&lt;p&gt;SUNK Anywhere 讓使用者能在 GKE、EKS 等多個 Kubernetes 環境中統一執行 AI 訓練，解決工具碎片化問題。影片展示如何透過標準化操作模型，在單一叢集執行訓練與推論，並提供簡化的使用者授權與可觀測性。&lt;/p&gt;&lt;p&gt;原站：&lt;a href=&quot;https://www.youtube.com/watch?v=o4fG8HblUFQ&quot;&gt;https://www.youtube.com/watch?v=o4fG8HblUFQ&lt;/a&gt;&lt;/p&gt;</description>
    <category>影片</category>
    <category>推論與部署</category>
  </item>
  <item>
    <title>How to Secure &amp; Run AI Agents with NVIDIA OpenShell</title>
    <link>https://aiwulin.itsmygo.uk/c/7d4b527fa9/</link>
    <guid isPermaLink="false">aiwulin-7d4b527fa9</guid>
    <pubDate>Mon, 28 Sep 2026 00:00:00 +0800</pubDate>
    <dc:creator>NVIDIA Developer</dc:creator>
    <description>&lt;p&gt;演示如何在 NVIDIA OpenShell 安全執行現有 AI 代理，無需修改代理本身。觀眾將學習設定檔案系統、網路與程序策略，並透過自動批准機制管理權限，防止代理洩漏敏感資料。&lt;/p&gt;&lt;p&gt;原站：&lt;a href=&quot;https://www.youtube.com/watch?v=GYYP-eW58ug&quot;&gt;https://www.youtube.com/watch?v=GYYP-eW58ug&lt;/a&gt;&lt;/p&gt;</description>
    <category>影片</category>
    <category>AI Agent 基礎</category>
    <category>AI 晶片與硬體</category>
  </item>
  <item>
    <title>An AI Research Agent That Runs Your Experiments — Tim Sweeney, Weights &amp; Biases</title>
    <link>https://aiwulin.itsmygo.uk/c/d204b03626/</link>
    <guid isPermaLink="false">aiwulin-d204b03626</guid>
    <pubDate>Sat, 26 Sep 2026 00:00:00 +0800</pubDate>
    <dc:creator>AI Engineer</dc:creator>
    <description>&lt;p&gt;Tim Sweeney 示範 Weights &amp;amp; Biases 新推出的 ARIA 研究代理，可自動執行實驗、分析資料並生成報告。觀眾能了解如何利用 ARIA 加速機器學習工作流，以及建立可觀察性與評估套件來生產化 AI 系統。&lt;/p&gt;&lt;p&gt;原站：&lt;a href=&quot;https://www.youtube.com/watch?v=hd7TOvmyAxU&quot;&gt;https://www.youtube.com/watch?v=hd7TOvmyAxU&lt;/a&gt;&lt;/p&gt;</description>
    <category>影片</category>
    <category>AI Agent 基礎</category>
  </item>
  <item>
    <title>Walking Tokyo Talking Agent Protocols</title>
    <link>https://aiwulin.itsmygo.uk/c/8278f93a56/</link>
    <guid isPermaLink="false">aiwulin-8278f93a56</guid>
    <pubDate>Fri, 18 Sep 2026 00:00:00 +0800</pubDate>
    <dc:creator>Agentic Conversations（原 MLOps.community）</dc:creator>
    <description>&lt;p&gt;訪談 Angie Jones 探討 AI 代理協議，重點介紹 WebMCP 如何讓代理直接在登入狀態下操作網頁，解決現有電腦操作代理耗時且需截圖的問題。&lt;/p&gt;&lt;p&gt;原站：&lt;a href=&quot;https://podcasters.spotify.com/pod/show/mlops/episodes/Walking-Tokyo-Talking-Agent-Protocols-e3p1m8o&quot;&gt;https://podcasters.spotify.com/pod/show/mlops/episodes/Walking-Tokyo-Talking-Agent-Protocols-e3p1m8o&lt;/a&gt;&lt;/p&gt;</description>
    <category>Podcast</category>
    <category>AI Agent 基礎</category>
    <category>瀏覽器與電腦操作</category>
    <category>Coding Agent</category>
  </item>
  <item>
    <title>Scaling Agent Workloads at Vercel</title>
    <link>https://aiwulin.itsmygo.uk/c/050b068a05/</link>
    <guid isPermaLink="false">aiwulin-050b068a05</guid>
    <pubDate>Thu, 17 Sep 2026 00:00:00 +0800</pubDate>
    <dc:creator>Software Engineering Daily</dc:creator>
    <description>&lt;p&gt;Vercel 工程師與產品負責人分享 Eve 框架，解決 AI 代理在企業級環境中擴展、隔離與恢復的挑戰。透過宣告式配置檔案將代理定義轉為基礎設施程式碼，讓平台自動提供所需資源，無需手動編寫複雜基礎設施。&lt;/p&gt;&lt;p&gt;原站：&lt;a href=&quot;https://softwareengineeringdaily.com/podcasts/scaling-agent-workloads-at-vercel&quot;&gt;https://softwareengineeringdaily.com/podcasts/scaling-agent-workloads-at-vercel&lt;/a&gt;&lt;/p&gt;</description>
    <category>Podcast</category>
    <category>AI Agent 基礎</category>
    <category>Coding Agent</category>
  </item>
  <item>
    <title>Docker, Adobe &amp; tldraw: Where Should Your Agent Run?</title>
    <link>https://aiwulin.itsmygo.uk/c/04ca2e6db3/</link>
    <guid isPermaLink="false">aiwulin-04ca2e6db3</guid>
    <pubDate>Wed, 16 Sep 2026 00:00:00 +0800</pubDate>
    <dc:creator>AI Native Dev</dc:creator>
    <description>&lt;p&gt;四位開發者探討程式設計代理應在哪種環境執行：Docker 使用微虛擬機器提供安全隔離，Helix ML 主張每個代理專屬獨立電腦以利協作，Adobe 讓代理在瀏覽器標籤內執行，tldraw 則將代理置於無限畫布上作為可互動角色。&lt;/p&gt;&lt;p&gt;原站：&lt;a href=&quot;https://www.youtube.com/watch?v=SLh-wN45ZIM&quot;&gt;https://www.youtube.com/watch?v=SLh-wN45ZIM&lt;/a&gt;&lt;/p&gt;</description>
    <category>影片</category>
    <category>AI Agent 基礎</category>
    <category>Coding Agent</category>
  </item>
  <item>
    <title>Essential Skills for Becoming an AI Engineer: RAG, AI Agents, &amp; More</title>
    <link>https://aiwulin.itsmygo.uk/c/a03fc34c23/</link>
    <guid isPermaLink="false">aiwulin-a03fc34c23</guid>
    <pubDate>Sun, 13 Sep 2026 00:00:00 +0800</pubDate>
    <dc:creator>IBM Technology</dc:creator>
    <description>&lt;p&gt;介紹 AI 工程師的職涯路徑與所需技能，區分機器學習研究員與 AI 工程師的差異。內容涵蓋從 Python、Git、Linux 等基礎，到 RAG、向量搜尋與 AI Agent，最後是 Kubernetes 部署與可觀測性，並提供實作專案…&lt;/p&gt;&lt;p&gt;原站：&lt;a href=&quot;https://www.youtube.com/watch?v=XN3xNJvWXsc&quot;&gt;https://www.youtube.com/watch?v=XN3xNJvWXsc&lt;/a&gt;&lt;/p&gt;</description>
    <category>影片</category>
    <category>AI 輔助軟體工程</category>
    <category>AI Agent 基礎</category>
    <category>Agent Skills 與 Harness</category>
  </item>
  <item>
    <title>The Five-Layer Cake Approach to Scaling AI Without Wasting Money</title>
    <link>https://aiwulin.itsmygo.uk/c/f7b1dd60ed/</link>
    <guid isPermaLink="false">aiwulin-f7b1dd60ed</guid>
    <pubDate>Fri, 04 Sep 2026 00:00:00 +0800</pubDate>
    <dc:creator>Agentic Conversations（原 MLOps.community）</dc:creator>
    <description>&lt;p&gt;Ambud Sharma 分享「五層蛋糕」架構，從硬體採購到治理分層最佳化 AI 效率。透過不同商業場景（如 Cursor 或 YouTube）示範如何根據用例選擇模型與架構，並強調實驗與基準測試的重要性。&lt;/p&gt;&lt;p&gt;原站：&lt;a href=&quot;https://podcasters.spotify.com/pod/show/mlops/episodes/The-Five-Layer-Cake-Approach-to-Scaling-AI-Without-Wasting-Money-e3ob95o&quot;&gt;https://podcasters.spotify.com/pod/show/mlops/episodes/The-Five-Layer-Cake-Approach-to-Scaling-AI-Without-Wasting-Money-e3ob95o&lt;/a&gt;&lt;/p&gt;</description>
    <category>Podcast</category>
    <category>推論與部署</category>
    <category>AI 評測</category>
    <category>AI 晶片與硬體</category>
  </item>
  <item>
    <title>Can agents keep a secret? We asked 1Password’s CTO Nancy Wang</title>
    <link>https://aiwulin.itsmygo.uk/c/7c69729b42/</link>
    <guid isPermaLink="false">aiwulin-7c69729b42</guid>
    <pubDate>Tue, 25 Aug 2026 00:00:00 +0800</pubDate>
    <dc:creator>Dev Interrupted</dc:creator>
    <description>&lt;p&gt;訪談 OnePassword CTO Nancy Wang，探討如何讓自主編碼代理在安全與效率間取得平衡。她提出「即時存取」策略，將秘密管理嵌入運算流程而非靜態儲存，並分享將客服人員轉為程式碼開發者的實戰案例。&lt;/p&gt;&lt;p&gt;原站：&lt;a href=&quot;https://www.buzzsprout.com/1422892/episodes/19697932-can-agents-keep-a-secret-we-asked-1password-s-cto-nancy-wang&quot;&gt;https://www.buzzsprout.com/1422892/episodes/19697932-can-agents-keep-a-secret-we-asked-1password-s-cto-nancy-wang&lt;/a&gt;&lt;/p&gt;</description>
    <category>Podcast</category>
    <category>Coding Agent</category>
    <category>AI Agent 基礎</category>
  </item>
  <item>
    <title>Rebuilding the Cloud for AI Agent Code</title>
    <link>https://aiwulin.itsmygo.uk/c/156b7ad6df/</link>
    <guid isPermaLink="false">aiwulin-156b7ad6df</guid>
    <pubDate>Thu, 13 Aug 2026 00:00:00 +0800</pubDate>
    <dc:creator>Software Engineering Daily</dc:creator>
    <description>&lt;p&gt;訪談 Render 創辦人 Anurag Goel，探討 AI 生成程式碼與 AI Agent 直接操作基礎設施後，雲端架構如何演變。&lt;/p&gt;&lt;p&gt;原站：&lt;a href=&quot;https://softwareengineeringdaily.com/podcasts/rebuilding-the-cloud-for-ai-agent-code&quot;&gt;https://softwareengineeringdaily.com/podcasts/rebuilding-the-cloud-for-ai-agent-code&lt;/a&gt;&lt;/p&gt;</description>
    <category>Podcast</category>
    <category>AI Agent 基礎</category>
    <category>Coding Agent</category>
  </item>
  <item>
    <title>AI-Powered Threats to the Software Supply Chain</title>
    <link>https://aiwulin.itsmygo.uk/c/afc197d5b0/</link>
    <guid isPermaLink="false">aiwulin-afc197d5b0</guid>
    <pubDate>Tue, 04 Aug 2026 00:00:00 +0800</pubDate>
    <dc:creator>Software Engineering Daily</dc:creator>
    <description>&lt;p&gt;Chainguard 創辦人 Matt Moore 訪談，探討軟體供應鏈攻擊現狀、CI/CD 管道風險與自動化修復策略。內容涵蓋從容器到語言庫的防護演進，以及 Anthropic Mythos 模型如何加速漏洞發現與修復的緊迫性。&lt;/p&gt;&lt;p&gt;原站：&lt;a href=&quot;https://softwareengineeringdaily.com/podcasts/ai-powered-threats-to-the-software-supply-chain&quot;&gt;https://softwareengineeringdaily.com/podcasts/ai-powered-threats-to-the-software-supply-chain&lt;/a&gt;&lt;/p&gt;</description>
    <category>Podcast</category>
    <category>AI 資安</category>
    <category>工作流自動化</category>
  </item>
  <item>
    <title>Building Durable AI Agents</title>
    <link>https://aiwulin.itsmygo.uk/c/77116e3faf/</link>
    <guid isPermaLink="false">aiwulin-77116e3faf</guid>
    <pubDate>Thu, 09 Jul 2026 00:00:00 +0800</pubDate>
    <dc:creator>Practical AI</dc:creator>
    <description>&lt;p&gt;探討如何將 AI 代理從實驗性演示轉化為可靠的大規模生產系統。Hamza Tahir 分享 MLOps 原則如何應用於代理開發，涵蓋工作流、代理框架及基礎設施。&lt;/p&gt;&lt;p&gt;原站：&lt;a href=&quot;https://share.transistor.fm/s/facb92e2&quot;&gt;https://share.transistor.fm/s/facb92e2&lt;/a&gt;&lt;/p&gt;</description>
    <category>Podcast</category>
    <category>AI Agent 基礎</category>
    <category>MLOps</category>
  </item>
  <item>
    <title>Why AI Infrastructure must evolve for Agent Experience — Akshat Bubna, Modal CTO</title>
    <link>https://aiwulin.itsmygo.uk/c/a8ed1f4deb/</link>
    <guid isPermaLink="false">aiwulin-a8ed1f4deb</guid>
    <pubDate>Thu, 09 Jul 2026 00:00:00 +0800</pubDate>
    <dc:creator>Latent Space</dc:creator>
    <description>&lt;p&gt;Modal CTO Akshat Bubna 探討 AI Agent 如何改變雲端基礎設施需求，強調傳統 Kubernetes 架構對開發者不友好，而 Agent 需要更緊密整合、快速迭代與隔離環境。&lt;/p&gt;&lt;p&gt;原站：&lt;a href=&quot;https://www.latent.space/p/modal2026&quot;&gt;https://www.latent.space/p/modal2026&lt;/a&gt;&lt;/p&gt;</description>
    <category>Podcast</category>
    <category>AI Agent 基礎</category>
  </item>
  <item>
    <title>The Next Programming Language Is English</title>
    <link>https://aiwulin.itsmygo.uk/c/718d9a1020/</link>
    <guid isPermaLink="false">aiwulin-718d9a1020</guid>
    <pubDate>Mon, 06 Jul 2026 00:00:00 +0800</pubDate>
    <dc:creator>Agentic Conversations（原 MLOps.community）</dc:creator>
    <description>&lt;p&gt;探討自然語言如何成為新的程式開發抽象層，並討論如何透過可持續執行（durable execution）等機制來補償 AI 的不確定性。內容涵蓋從組碼語言到雲端架構的抽象演進，以及 AI 代理在長期任務與人類介入流程中的應用。&lt;/p&gt;&lt;p&gt;原站：&lt;a href=&quot;https://podcasters.spotify.com/pod/show/mlops/episodes/The-Next-Programming-Language-Is-English-e3lnahg&quot;&gt;https://podcasters.spotify.com/pod/show/mlops/episodes/The-Next-Programming-Language-Is-English-e3lnahg&lt;/a&gt;&lt;/p&gt;</description>
    <category>Podcast</category>
    <category>Coding Agent</category>
    <category>AI Agent 基礎</category>
    <category>大型語言模型入門</category>
  </item>
  <item>
    <title>How Spotify runs agents across 20M+ lines of code, with Niklas Gustavsson</title>
    <link>https://aiwulin.itsmygo.uk/c/aaf8b09e68/</link>
    <guid isPermaLink="false">aiwulin-aaf8b09e68</guid>
    <pubDate>Tue, 30 Jun 2026 00:00:00 +0800</pubDate>
    <dc:creator>Claude</dc:creator>
    <description>&lt;p&gt;Spotify 工程總監 Niklas Gustavsson 分享如何利用 Claude 與自研工具 Honk，在 2000 萬行程式碼的單體倉庫中自動生成原型。&lt;/p&gt;&lt;p&gt;原站：&lt;a href=&quot;https://www.youtube.com/watch?v=9DHZLw5653E&quot;&gt;https://www.youtube.com/watch?v=9DHZLw5653E&lt;/a&gt;&lt;/p&gt;</description>
    <category>影片</category>
    <category>AI Agent 基礎</category>
    <category>Coding Agent</category>
    <category>Claude Code</category>
  </item>
  <item>
    <title>Zero Trust for AI Agents</title>
    <link>https://aiwulin.itsmygo.uk/c/839ce1d6ff/</link>
    <guid isPermaLink="false">aiwulin-839ce1d6ff</guid>
    <pubDate>Thu, 11 Jun 2026 00:00:00 +0800</pubDate>
    <dc:creator>Practical AI</dc:creator>
    <description>&lt;p&gt;深入解析 Anthropic 提出的「AI 代理零信任」安全架構，探討自主 AI 代理帶來的新型威脅與風險。內容涵蓋身份權限濫用、供應鏈依賴風險及動態攻擊防禦，並提供具體的實施階段與控制措施，幫助組織在部署 AI 代理時建立安全邊界。&lt;/p&gt;&lt;p&gt;原站：&lt;a href=&quot;https://share.transistor.fm/s/5c1a087d&quot;&gt;https://share.transistor.fm/s/5c1a087d&lt;/a&gt;&lt;/p&gt;</description>
    <category>Podcast</category>
    <category>AI Agent 基礎</category>
    <category>AI 資安</category>
  </item>
  <item>
    <title>Kubernetes and retiring at the top with Kelsey Hightower</title>
    <link>https://aiwulin.itsmygo.uk/c/53cd438707/</link>
    <guid isPermaLink="false">aiwulin-53cd438707</guid>
    <pubDate>Thu, 04 Jun 2026 00:00:00 +0800</pubDate>
    <dc:creator>The Pragmatic Engineer</dc:creator>
    <description>&lt;p&gt;Kelsey Hightower 分享從麥當勞員工到 Google 首席工程師的非凡生涯，探討容器化、Kubernetes 的興起與演進。&lt;/p&gt;&lt;p&gt;原站：&lt;a href=&quot;https://newsletter.pragmaticengineer.com/p/kubernetes-and-retiring-at-the-top&quot;&gt;https://newsletter.pragmaticengineer.com/p/kubernetes-and-retiring-at-the-top&lt;/a&gt;&lt;/p&gt;</description>
    <category>Podcast</category>
    <category>職涯與學習路線</category>
  </item>
  <item>
    <title>Rebooting Enterprise AI with MCP and Kubernetes</title>
    <link>https://aiwulin.itsmygo.uk/c/1bbcf826a7/</link>
    <guid isPermaLink="false">aiwulin-1bbcf826a7</guid>
    <pubDate>Thu, 28 May 2026 00:00:00 +0800</pubDate>
    <dc:creator>Practical AI</dc:creator>
    <description>&lt;p&gt;探討 MCP 協議如何讓 AI 代理像員工一樣安全地存取企業系統，解決身份驗證與權限控制問題。透過 Stacklok 與 ToolHive 的實作案例，說明如何建立代理並行運作的架構，大幅提升開發效率與生產力。&lt;/p&gt;&lt;p&gt;原站：&lt;a href=&quot;https://share.transistor.fm/s/d76e02d5&quot;&gt;https://share.transistor.fm/s/d76e02d5&lt;/a&gt;&lt;/p&gt;</description>
    <category>Podcast</category>
    <category>MCP 模型上下文協定</category>
  </item>
  <item>
    <title>Railway: The Agent-Native Cloud — Jake Cooper</title>
    <link>https://aiwulin.itsmygo.uk/c/c72c362505/</link>
    <guid isPermaLink="false">aiwulin-c72c362505</guid>
    <pubDate>Thu, 21 May 2026 00:00:00 +0800</pubDate>
    <dc:creator>Latent Space</dc:creator>
    <description>&lt;p&gt;Railway 創辦人 Jake Cooper 分享公司從手動拉客到快速成長的歷程，強調透過裸金屬資料中心與自研基礎設施降低部署門檻。內容探討如何讓 AI 代理安全地進行生產環境迭代，並提出未來將完全自有的資料中心取代公有雲的願景。&lt;/p&gt;&lt;p&gt;原站：&lt;a href=&quot;https://www.latent.space/p/railway&quot;&gt;https://www.latent.space/p/railway&lt;/a&gt;&lt;/p&gt;</description>
    <category>Podcast</category>
    <category>AI Agent 基礎</category>
    <category>Coding Agent</category>
  </item>
  <item>
    <title>The Agent Harness: Building Secure Sandboxes for Autonomous AI Workloads</title>
    <link>https://aiwulin.itsmygo.uk/c/fe7c2dad73/</link>
    <guid isPermaLink="false">aiwulin-fe7c2dad73</guid>
    <pubDate>Thu, 14 May 2026 00:00:00 +0800</pubDate>
    <dc:creator>The MAD Podcast</dc:creator>
    <description>&lt;p&gt;訪談探討 AI 代理（Agent）需要獨立安全沙盒執行的必要性，解釋為何現有無狀態雲端架構無法滿足需求，並分析微虛擬機器與自訂排程器的技術細節。聽眾可了解如何建立安全代理環境，以及未來可能面臨的 CPU 短缺挑戰。&lt;/p&gt;&lt;p&gt;原站：&lt;a href=&quot;https://podcasters.spotify.com/pod/show/firstmark/episodes/The-Agent-Harness-Building-Secure-Sandboxes-for-Autonomous-AI-Workloads-e3japf7&quot;&gt;https://podcasters.spotify.com/pod/show/firstmark/episodes/The-Agent-Harness-Building-Secure-Sandboxes-for-Autonomous-AI-Workloads-e3japf7&lt;/a&gt;&lt;/p&gt;</description>
    <category>Podcast</category>
    <category>AI Agent 基礎</category>
    <category>Agent Skills 與 Harness</category>
  </item>
  <item>
    <title>Automation at the speed of Swamp (Friends)</title>
    <link>https://aiwulin.itsmygo.uk/c/f6c9e5f172/</link>
    <guid isPermaLink="false">aiwulin-f6c9e5f172</guid>
    <pubDate>Thu, 14 May 2026 00:00:00 +0800</pubDate>
    <dc:creator>The Changelog</dc:creator>
    <description>&lt;p&gt;Adam Jacob 與 Adam Stacoviak 訪談 Swamp 創辦人，探討 AI 代理如何重塑軟體開發。透過現場示範，展示 Swamp 如何自動管理 Proxmox 虛擬機器、撰寫工作流並最佳化使用者體驗，強調架構與領域驅動設…&lt;/p&gt;&lt;p&gt;原站：&lt;a href=&quot;https://changelog.com/friends/130&quot;&gt;https://changelog.com/friends/130&lt;/a&gt;&lt;/p&gt;</description>
    <category>Podcast</category>
    <category>工作流自動化</category>
    <category>AI Agent 基礎</category>
    <category>Coding Agent</category>
  </item>
  <item>
    <title>NVIDIA's AI Engineers: Agent Inference at Planetary Scale and &quot;Speed of Light&quot; — Nader Khalil (Brev), Kyle Kranen (Dynamo)</title>
    <link>https://aiwulin.itsmygo.uk/c/35dfdbec2a/</link>
    <guid isPermaLink="false">aiwulin-35dfdbec2a</guid>
    <pubDate>Tue, 10 Mar 2026 00:00:00 +0800</pubDate>
    <dc:creator>Latent Space</dc:creator>
    <description>&lt;p&gt;訪談 NVIDIA 工程師 Nader Khalil 與 Kyle Kranen，探討 Brev 如何降低開發者使用高階 GPU 的門檻，以及 Dynamo 框架如何透過預填充與解碼分離技術，實現大規模推理加速與成本最佳化。&lt;/p&gt;&lt;p&gt;原站：&lt;a href=&quot;https://www.latent.space/p/nvidia-brev-dynamo&quot;&gt;https://www.latent.space/p/nvidia-brev-dynamo&lt;/a&gt;&lt;/p&gt;</description>
    <category>Podcast</category>
    <category>推論與部署</category>
    <category>AI 晶片與硬體</category>
    <category>AI 輔助軟體工程</category>
  </item>
  <item>
    <title>Kubernetes, Compliance, and Control: The Operational Backbone of AI Sovereignty</title>
    <link>https://aiwulin.itsmygo.uk/c/57848968c9/</link>
    <guid isPermaLink="false">aiwulin-57848968c9</guid>
    <pubDate>Wed, 25 Feb 2026 00:00:00 +0800</pubDate>
    <dc:creator>AI Engineering Podcast</dc:creator>
    <description>&lt;p&gt;探討企業與國家如何透過自管基礎設施與合規策略實現 AI 主權，涵蓋 Kubernetes、PyTorch 等工具在 LLM 服務中的應用。內容涵蓋從 GPU 存取難題到代理身份認證等實務挑戰，幫助讀者理解 AI 主權的技術與營運路徑。&lt;/p&gt;&lt;p&gt;原站：&lt;a href=&quot;https://www.aiengineeringpodcast.com/ai-sovereignty-infrastructure-adaptation-episode-78&quot;&gt;https://www.aiengineeringpodcast.com/ai-sovereignty-infrastructure-adaptation-episode-78&lt;/a&gt;&lt;/p&gt;</description>
    <category>Podcast</category>
    <category>AI 晶片與硬體</category>
    <category>大型語言模型入門</category>
    <category>推論與部署</category>
  </item>
  <item>
    <title>Mistral AI vs. Silicon Valley: The Rise of Sovereign AI</title>
    <link>https://aiwulin.itsmygo.uk/c/4f2227a03d/</link>
    <guid isPermaLink="false">aiwulin-4f2227a03d</guid>
    <pubDate>Thu, 12 Feb 2026 00:00:00 +0800</pubDate>
    <dc:creator>The MAD Podcast</dc:creator>
    <description>&lt;p&gt;Timothée Lacroix 分享 Mistral AI 如何從開源實驗室轉型為具備完整堆疊的「主權 AI」供應商，透過自營資料中心與 Mistral Compute 架構，讓企業掌握模型與工作流的控制權。&lt;/p&gt;&lt;p&gt;原站：&lt;a href=&quot;https://podcasters.spotify.com/pod/show/firstmark/episodes/Mistral-AI-vs--Silicon-Valley-The-Rise-of-Sovereign-AI-e3f003u&quot;&gt;https://podcasters.spotify.com/pod/show/firstmark/episodes/Mistral-AI-vs--Silicon-Valley-The-Rise-of-Sovereign-AI-e3f003u&lt;/a&gt;&lt;/p&gt;</description>
    <category>Podcast</category>
    <category>開源模型</category>
    <category>AI 晶片與硬體</category>
  </item>
</channel>
</rss>
