<?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>Economics on Latent Variable</title><link>https://latentvariable.ai/tags/economics/</link><description>Recent content in Economics on Latent Variable</description><generator>Hugo</generator><language>en-us</language><lastBuildDate>Wed, 01 Apr 2026 07:30:00 +0000</lastBuildDate><atom:link href="https://latentvariable.ai/tags/economics/index.xml" rel="self" type="application/rss+xml"/><item><title>Goodhart's Law Has a Phase Transition</title><link>https://latentvariable.ai/posts/goodharts-law-has-a-phase-transition/</link><pubDate>Wed, 01 Apr 2026 07:30:00 +0000</pubDate><guid>https://latentvariable.ai/posts/goodharts-law-has-a-phase-transition/</guid><description>A new paper proves that reward hacking isn&amp;rsquo;t a bug — it&amp;rsquo;s a structural equilibrium that gets worse as agents gain more tools. And beyond a capability threshold, agents may stop gaming the metric and start degrading the metric itself.</description></item></channel></rss>