
  <rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
    <channel>
      <title>Wayne Blog</title>
      <link>https://waynes.gv.uy/blog</link>
      <description>个人技术主页与写作空间</description>
      <language>zh-CN</language>
      <managingEditor> (Wayne)</managingEditor>
      <webMaster> (Wayne)</webMaster>
      <lastBuildDate>Sun, 05 Apr 2026 00:00:00 GMT</lastBuildDate>
      <atom:link href="https://waynes.gv.uy/tags/ai/feed.xml" rel="self" type="application/rss+xml"/>
      
  <item>
    <guid>https://waynes.gv.uy/blog/Learn-Agent</guid>
    <title>Learn Agent</title>
    <link>https://waynes.gv.uy/blog/Learn-Agent</link>
    <description>Agent 持续学习中</description>
    <pubDate>Sun, 05 Apr 2026 00:00:00 GMT</pubDate>
    <author> (Wayne)</author>
    <category>Agent</category><category>Ai</category>
  </item>

  <item>
    <guid>https://waynes.gv.uy/blog/agent-mcp-skills-intro</guid>
    <title>Agent、MCP、Skills 的简单介绍</title>
    <link>https://waynes.gv.uy/blog/agent-mcp-skills-intro</link>
    <description>用工程化视角解释 Agent、MCP 和 Skills 的关系，梳理从闭环系统、协议标准到技能单元的基本概念和实践方式。</description>
    <pubDate>Tue, 03 Feb 2026 00:00:00 GMT</pubDate>
    <author> (Wayne)</author>
    <category>AI</category><category>Agent</category><category>MCP</category><category>Skills</category>
  </item>

  <item>
    <guid>https://waynes.gv.uy/blog/qwen3-4b-local-cpu-deploy</guid>
    <title>在 AMD 笔记本上本地跑 Qwen3-4B：一套不依赖 CUDA 的轻量部署方案</title>
    <link>https://waynes.gv.uy/blog/qwen3-4b-local-cpu-deploy</link>
    <description>在一台没有 NVIDIA GPU 的 AMD 笔记本上，用 GGUF + llama.cpp 跑通 Qwen3-4B。本篇重点不是“把模型跑起来”，而是把选型、下载、校验和运行方式整理成一套可复现、可维护的最小闭环。</description>
    <pubDate>Tue, 28 Apr 2026 00:00:00 GMT</pubDate>
    <author> (Wayne)</author>
    <category>AI</category><category>LLM</category><category>Qwen</category><category>llama.cpp</category><category>本地部署</category>
  </item>

  <item>
    <guid>https://waynes.gv.uy/blog/rag-knowledge-base-learning-plan</guid>
    <title>RAG 不只是接个向量库：从最小系统到工程判断</title>
    <link>https://waynes.gv.uy/blog/rag-knowledge-base-learning-plan</link>
    <description>RAG 真正难的不是把检索接到模型前面，而是把文档边界、召回质量、上下文约束和评估体系做对。本文按工程链路拆开 RAG，讲清楚第一版系统该怎么做，问题通常死在哪，以及什么时候才值得上更复杂的优化。</description>
    <pubDate>Thu, 09 Apr 2026 00:00:00 GMT</pubDate>
    <author> (Wayne)</author>
    <category>AI</category><category>RAG</category><category>知识库</category><category>工程实践</category>
  </item>

    </channel>
  </rss>
