<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
  <channel>
    <title>The Minutes of Dylan Patel on opus</title>
    <link>https://minutesof.com/dylan-patel/on/opus/</link>
    <description>Everything Dylan Patel has said on opus: 2 verbatim quotes from May 2026, each with a timestamp and a link to the recording it came from.</description>
    <language>en</language>
    <lastBuildDate>Mon, 31 Aug 2026 13:10:15 +0000</lastBuildDate>
    <atom:link href="https://minutesof.com/dylan-patel/on/opus/feed.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>Patel reports kernel programmers use Codex to generate code then have Opus rewrite it, not reverse.</title>
      <link>https://minutesof.com/q/c776bb8a-664b-48d6-9d85-372aa9b92246/</link>
      <guid isPermaLink="true">https://minutesof.com/q/c776bb8a-664b-48d6-9d85-372aa9b92246/</guid>
      <description>“I think it was tree down. Tree down is like, yeah, dude, Codex is like so dumb, but I always just have it created and then this and it works and it&#x27;s smarter, but the code is sloppy, I have Opus rewrite it. But you can&#x27;t go the other way around.” — SemiAnalysis Weekly</description>
      <pubDate>Wed, 06 May 2026 00:55:06 +0000</pubDate>
      <category>opus</category>
    </item>
    <item>
      <title>Patel reports kernel programmers use Codex to generate code then Opus to clean it, not the reverse workflow.</title>
      <link>https://minutesof.com/q/bc52f0cc-dd82-4db4-89fa-99f109589046/</link>
      <guid isPermaLink="true">https://minutesof.com/q/bc52f0cc-dd82-4db4-89fa-99f109589046/</guid>
      <description>“But you can&#x27;t go the other way around. You can&#x27;t have Opus write the thing and then have Codex fix this.” — SemiAnalysis Weekly</description>
      <pubDate>Wed, 06 May 2026 00:55:06 +0000</pubDate>
      <category>opus</category>
    </item>
  </channel>
</rss>
