<?xml version="1.0" encoding="utf-8" standalone="yes"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
  <channel>
    <title>Compliance on Jeff Geiser</title>
    <link>https://jeffgeiser.dev/tags/compliance/</link>
    <description>Recent content in Compliance on Jeff Geiser</description>
    <generator>Hugo</generator>
    <language>en-US</language>
    <lastBuildDate>Mon, 12 Oct 2026 00:00:00 +0000</lastBuildDate>
    <atom:link href="https://jeffgeiser.dev/tags/compliance/index.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>Test the Stack, Not the Diagram</title>
      <link>https://jeffgeiser.dev/posts/test-the-stack-not-the-diagram/</link>
      <pubDate>Mon, 12 Oct 2026 00:00:00 +0000</pubDate>
      <guid>https://jeffgeiser.dev/posts/test-the-stack-not-the-diagram/</guid>
      <description>&lt;p&gt;&lt;strong&gt;Data in an AI stack doesn&amp;rsquo;t live in one place. It sprawls across caches, logs, vector stores, and traces. Residency is a surface, and you can&amp;rsquo;t prove a surface from a diagram.&lt;/strong&gt;&lt;/p&gt;&#xA;&lt;p&gt;&lt;em&gt;By Jeff Geiser — October 2026&lt;/em&gt;&lt;/p&gt;&#xA;&lt;hr&gt;&#xA;&lt;p&gt;If you&amp;rsquo;ve built an AI stack, you already know the uncomfortable version of this: your data does not live in one place.&lt;/p&gt;&#xA;&lt;p&gt;You designed a request path. Prompt comes in, hits the gateway, goes to a model, answer goes back. That path is what&amp;rsquo;s on the architecture diagram, and it&amp;rsquo;s real. It&amp;rsquo;s also a small fraction of where the data actually goes.&lt;/p&gt;</description>
    </item>
  </channel>
</rss>
