<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Interactive-Fiction on Guy Freeman</title><link>https://gfrm.in/categories/interactive-fiction/</link><description>Recent content in Interactive-Fiction on Guy Freeman</description><generator>Hugo</generator><language>en</language><lastBuildDate>Mon, 23 Mar 2026 00:00:00 +0000</lastBuildDate><atom:link href="https://gfrm.in/categories/interactive-fiction/index.xml" rel="self" type="application/rss+xml"/><item><title>Teaching Zork to a Bayesian</title><link>https://gfrm.in/posts/teaching-zork/</link><pubDate>Tue, 28 Apr 2026 00:00:00 +0000</pubDate><guid>https://gfrm.in/posts/teaching-zork/</guid><description>&lt;div class="callout callout-note"&gt;
 This is Part 2 of a series. For the axioms and types underneath, see &lt;a href="https://gfrm.in/posts/three-types/"&gt;Part 1: Three Types and a Funeral&lt;/a&gt;. For the state-representation consequences, see &lt;a href="https://gfrm.in/posts/loop-problem/"&gt;Part 3: The Loop Problem&lt;/a&gt;.
&lt;/div&gt;

&lt;p&gt;Every AI agent demo involves web search, retrieval, or API calls &amp;mdash; tasks where querying everything is merely expensive. A LangChain ReAct agent that hammers all four tools on every question wastes money but still gets answers. The penalty is economic, not existential.&lt;/p&gt;</description></item></channel></rss>