<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
  <channel>
    <title>AI AX 노트</title>
    <link>https://aiaxnote.com</link>
    <description>중소기업 AI 전환(AI AX) 방법론과 AI 기술을 실제로 만들어 운영하며 측정한 데이터로 설명합니다. 소버린 AI, LLM 라우팅과 비용 구조, 에이전트 운영을 다룹니다.</description>
    <language>ko</language>
    <atom:link href="https://aiaxnote.com/rss.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>AI 자율 실행, 승인은 어디까지 걸어야 하는가</title>
      <link>https://aiaxnote.com/posts/where-to-draw-the-approval-line</link>
      <guid isPermaLink="true">https://aiaxnote.com/posts/where-to-draw-the-approval-line</guid>
      <description>AI 에이전트에게 실행 권한을 주면 승인 절차를 어디까지 남길지가 곧 설계 문제가 된다. 되돌릴 수 있는 일과 없는 일을 가르는 기준, 그리고 실행됐다는 보고를 믿어도 되는지 검증하는 방법을 실제 운영 경험으로 정리했다.</description>
      <pubDate>Thu, 13 Aug 2026 23:30:00 GMT</pubDate>
      <category>ai-ax</category>
    </item>
    <item>
      <title>소버린 AI는 국산 모델을 만드는 일이 아니다</title>
      <link>https://aiaxnote.com/posts/sovereign-ai-is-not-just-a-domestic-model</link>
      <guid isPermaLink="true">https://aiaxnote.com/posts/sovereign-ai-is-not-just-a-domestic-model</guid>
      <description>소버린 AI를 국산 LLM 개발로만 이해하면 정작 중요한 것을 놓친다. 주권은 모델 소유가 아니라 인프라·데이터·프로세스·운영권을 스스로 쥐고 있느냐의 문제다.</description>
      <pubDate>Thu, 13 Aug 2026 20:10:00 GMT</pubDate>
      <category>ai-ax</category>
    </item>
  </channel>
</rss>