<?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>Developer Culture on AI Pulse</title>
    <link>https://cuigh.com/tags/developer-culture/</link>
    <description>Recent content in Developer Culture on AI Pulse</description>
    <generator>Hugo</generator>
    <language>en-US</language>
    <lastBuildDate>Fri, 26 Jun 2026 11:12:19 +0800</lastBuildDate>
    <atom:link href="https://cuigh.com/tags/developer-culture/index.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>Why you can&#39;t unit test for taste in the age of AI</title>
      <link>https://cuigh.com/posts/you-cant-unit-test-for-taste-2026/</link>
      <pubDate>Fri, 26 Jun 2026 11:12:19 +0800</pubDate>
      <guid>https://cuigh.com/posts/you-cant-unit-test-for-taste-2026/</guid>
      <description>Karl Tryggvason posted &amp;#34;You can&amp;#39;t unit test for taste&amp;#34; on June 24. It hit 254 points and 117 comments on HN within a day. He was building a running app and wanted to enrich virtual routes with points of interest. AI gave him a list of museums. Every entry was correct. Some were just boring. He couldn&amp;#39;t write a test for that. The same week, GitHub&amp;#39;s monthly leaderboard pushed two &amp;#39;taste&amp;#39; skills past 60k stars combined. This post is about the thing those tools can&amp;#39;t reach.</description>
    </item>
  </channel>
</rss>
