<?xml version="1.0" encoding="utf-8" standalone="yes"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
  <channel>
    <title>Perturbation Testing on Programmer.ie: Modern AI programming</title>
    <link>http://programmer.ie/tags/perturbation-testing/</link>
    <description>Recent content in Perturbation Testing on Programmer.ie: Modern AI programming</description>
    <generator>Hugo</generator>
    <language>en-us</language>
    <lastBuildDate>Wed, 22 Apr 2026 10:35:46 +0100</lastBuildDate>
    <atom:link href="http://programmer.ie/tags/perturbation-testing/index.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>Beyond Hallucination Energy: A Three-Dimensional Framework for Reliable AI Outputs</title>
      <link>http://programmer.ie/post/trendslop/</link>
      <pubDate>Wed, 22 Apr 2026 10:35:46 +0100</pubDate>
      <guid>http://programmer.ie/post/trendslop/</guid>
      <description>&lt;h2 id=&#34;-1--tldr&#34;&gt;🧩 1.  TLDR&lt;/h2&gt;&#xA;&lt;blockquote&gt;&#xA;&lt;p&gt;&lt;strong&gt;AI doesn&amp;rsquo;t just hallucinate.&#xA;Sometimes it gives answers that are fluent, safe… and completely useless.&lt;/strong&gt;&lt;/p&gt;&lt;/blockquote&gt;&#xA;&lt;p&gt;Most discussions about AI failure focus on hallucination:&lt;/p&gt;&#xA;&lt;ul&gt;&#xA;&lt;li&gt;making things up&lt;/li&gt;&#xA;&lt;li&gt;getting facts wrong&lt;/li&gt;&#xA;&lt;li&gt;fabricating sources&lt;/li&gt;&#xA;&lt;/ul&gt;&#xA;&lt;p&gt;That&amp;rsquo;s real. It matters.&lt;/p&gt;&#xA;&lt;p&gt;But it&amp;rsquo;s not the most dangerous failure mode in production systems.&lt;/p&gt;&#xA;&lt;p&gt;There is a quieter one.&lt;/p&gt;&#xA;&lt;p&gt;A more subtle one.&lt;/p&gt;&#xA;&lt;p&gt;And in practice a more &lt;em&gt;pervasive&lt;/em&gt; one.&lt;/p&gt;&#xA;&lt;blockquote&gt;&#xA;&lt;p&gt;&lt;strong&gt;AI systems often fail not by being wrong,&#xA;but by failing to think at all.&lt;/strong&gt;&lt;/p&gt;</description>
    </item>
  </channel>
</rss>
