<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Rag on Unbound Force</title><link>https://unboundforce.dev/tags/rag/</link><description>Recent content in Rag on Unbound Force</description><generator>Hugo</generator><language>en-US</language><copyright>Copyright (c) 2025-2026 Unbound Force</copyright><lastBuildDate>Tue, 07 Apr 2026 00:00:00 +0000</lastBuildDate><atom:link href="https://unboundforce.dev/tags/rag/index.xml" rel="self" type="application/rss+xml"/><item><title>The Librarian vs The Index: Two Ways to Give AI Agents a Memory</title><link>https://unboundforce.dev/blog/dewey-vs-karpathy/</link><pubDate>Tue, 07 Apr 2026 00:00:00 +0000</pubDate><guid>https://unboundforce.dev/blog/dewey-vs-karpathy/</guid><description>&lt;h2 id="the-problem-both-solve"&gt;The Problem Both Solve&lt;/h2&gt;
&lt;p&gt;Every developer using AI coding tools knows the feeling: you spend an hour building context with an agent — explaining the architecture, the naming conventions, the reasons behind a particular design decision — and then the session ends. The next session starts blank. The agent has no memory of what you discussed, no access to the decisions you made, no awareness of the other repositories in your organization.&lt;/p&gt;</description></item></channel></rss>