<?xml version="1.0" encoding="utf-8" standalone="yes"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:content="http://purl.org/rss/1.0/modules/content/">
  <channel>
    <title>Context-Window on Loop &amp; Retry</title>
    <link>https://loopandretry.github.io/tags/context-window/</link>
    <description>Recent content in Context-Window on Loop &amp; Retry</description>
    <image>
      <title>Loop &amp; Retry</title>
      <url>https://loopandretry.github.io/images/og-default.png</url>
      <link>https://loopandretry.github.io/images/og-default.png</link>
    </image>
    <generator>Hugo</generator>
    <language>en-us</language>
    <lastBuildDate>Tue, 07 Jul 2026 20:00:00 -0400</lastBuildDate>
    <atom:link href="https://loopandretry.github.io/tags/context-window/index.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>The context window is a cache, not a memory</title>
      <link>https://loopandretry.github.io/posts/context-window-is-a-cache/</link>
      <pubDate>Tue, 07 Jul 2026 20:00:00 -0400</pubDate>
      <guid>https://loopandretry.github.io/posts/context-window-is-a-cache/</guid>
      <description>Treating the context window as append-only memory is how agents get slow, expensive, and quietly wrong. The fix is to run it like a cache with a budget and an eviction policy: decide what earns its tokens every turn. Here&amp;rsquo;s the cost math, the accuracy failure mode, and a working context manager.</description>
    </item>
  </channel>
</rss>
