<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
  <channel>
    <title>Stefan Horne — Writing</title>
    <link>https://www.stefanhorne.com/blog</link>
    <description>Writing by Stefan Horne on software engineering, AI-native workflows, reliability, and product delivery.</description>
    <language>en</language>
    <lastBuildDate>Thu, 11 Jun 2026 00:00:00 GMT</lastBuildDate>
    <atom:link href="https://www.stefanhorne.com/rss.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>An Agent Takes the Shape of Its Context</title>
      <link>https://www.stefanhorne.com/post/an-agent-takes-the-shape-of-its-context</link>
      <guid>https://www.stefanhorne.com/post/an-agent-takes-the-shape-of-its-context</guid>
      <pubDate>Thu, 11 Jun 2026 00:00:00 GMT</pubDate>
      <description>SCOPED is a working vocabulary for structuring context so agents can navigate a workspace instead of a fuzzy pile.</description>
    </item>
    <item>
      <title>AI-native engineering is a workflow, not a vibe</title>
      <link>https://www.stefanhorne.com/post/ai-native-engineering</link>
      <guid>https://www.stefanhorne.com/post/ai-native-engineering</guid>
      <pubDate>Fri, 05 Jun 2026 00:00:00 GMT</pubDate>
      <description>Using AI well in engineering means designing the loop around it, not lowering the bar.</description>
    </item>
    <item>
      <title>Agents Need Memory, Not Just More Context</title>
      <link>https://www.stefanhorne.com/post/engineering-memory-infrastructure</link>
      <guid>https://www.stefanhorne.com/post/engineering-memory-infrastructure</guid>
      <pubDate>Sat, 30 May 2026 00:00:00 GMT</pubDate>
      <description>Why coding agents need project history with provenance, ownership, and freshness, not just larger context windows.</description>
    </item>
  </channel>
</rss>
