<?xml version="1.0" encoding="utf-8" standalone="yes"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
  <channel>
    <title>Working Set on Programmer.ie: Modern AI programming</title>
    <link>http://programmer.ie/tags/working-set/</link>
    <description>Recent content in Working Set on Programmer.ie: Modern AI programming</description>
    <generator>Hugo</generator>
    <language>en-us</language>
    <lastBuildDate>Wed, 23 Sep 2026 05:00:13 +0100</lastBuildDate>
    <atom:link href="http://programmer.ie/tags/working-set/index.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>Externalize the Working Set</title>
      <link>http://programmer.ie/books/context/13-chapter/</link>
      <pubDate>Wed, 23 Sep 2026 05:00:13 +0100</pubDate>
      <guid>http://programmer.ie/books/context/13-chapter/</guid>
      <description>&lt;p&gt;Chapter 12 ended with an irreducible remainder: pinned exact spans, active dense episodes, compact views of the recent past, anchors of the deep past. Still resident, still costing capacity and cache stability. Consider what that remainder looks like in a concrete session. An agent halfway through a dependency migration holds five working objects: an architecture investigation at roughly 18,000 tokens, a benchmark report at 12,000, a compiler trace at 25,000, a design-alternatives note at 15,000, and a vendored API reference at 30,000. Progressive fidelity has done its honest work and the resident views are compact. The sum is still near 100,000 tokens of occupancy, and on most turns the model consults at most one of the five. The tiers answered how little of each object must be shown. None of them asked whether the object needs to be shown at all.&lt;/p&gt;</description>
    </item>
  </channel>
</rss>
