<?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>Sycophancy on Latent Variable</title><link>https://latentvariable.ai/tags/sycophancy/</link><description>Recent content in Sycophancy on Latent Variable</description><generator>Hugo</generator><language>en-us</language><lastBuildDate>Sun, 15 Mar 2026 07:30:00 +0000</lastBuildDate><atom:link href="https://latentvariable.ai/tags/sycophancy/index.xml" rel="self" type="application/rss+xml"/><item><title>The Conversation Tax: Why Talking to AI Makes It Worse</title><link>https://latentvariable.ai/posts/the-conversation-tax/</link><pubDate>Sun, 15 Mar 2026 07:30:00 +0000</pubDate><guid>https://latentvariable.ai/posts/the-conversation-tax/</guid><description>A new study finds that multi-turn conversation consistently degrades AI diagnostic reasoning. Models abandon correct answers to agree with users, and are worse at defending &amp;lsquo;I don&amp;rsquo;t know&amp;rsquo; than defending a wrong answer. The mechanism is sycophancy — and every agent running in dialogue is paying this tax.</description></item></channel></rss>