<?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>Dissent on Pulling Threads</title><link>https://ravel.untanglingsystems.io/tags/dissent/</link><description>Recent content in Dissent on Pulling Threads</description><generator>Hugo -- 0.157.0</generator><language>en-us</language><lastBuildDate>Sun, 22 Mar 2026 16:30:00 +0000</lastBuildDate><atom:link href="https://ravel.untanglingsystems.io/tags/dissent/index.xml" rel="self" type="application/rss+xml"/><item><title>Dissent as Architecture: Building AI Teams That Argue on Purpose</title><link>https://ravel.untanglingsystems.io/posts/dissent-as-architecture/</link><pubDate>Sun, 22 Mar 2026 16:30:00 +0000</pubDate><guid>https://ravel.untanglingsystems.io/posts/dissent-as-architecture/</guid><description>Most multi-agent AI setups optimize for agreement. A growing number of practitioners and researchers argue that&amp;rsquo;s exactly backwards — and that disagreement should be a structural feature, not a bug to resolve. Here&amp;rsquo;s who&amp;rsquo;s saying it, why it matters, and how you might actually build it.</description></item></channel></rss>