<?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>Rant on blog.yfzhou</title><link>https://blog.yfzhou.fyi/tags/rant/</link><description>Recent content in Rant on blog.yfzhou</description><generator>Hugo</generator><language>en-us</language><managingEditor>mail@yfzhou.fyi (yfzhou)</managingEditor><webMaster>mail@yfzhou.fyi (yfzhou)</webMaster><lastBuildDate>Thu, 27 Aug 2026 08:00:00 +0800</lastBuildDate><atom:link href="https://blog.yfzhou.fyi/tags/rant/index.xml" rel="self" type="application/rss+xml"/><item><title>Don't Let LLMs Play Telephone With Your Ideas</title><link>https://blog.yfzhou.fyi/posts/prompt/</link><pubDate>Thu, 27 Aug 2026 08:00:00 +0800</pubDate><author>mail@yfzhou.fyi (yfzhou)</author><guid>https://blog.yfzhou.fyi/posts/prompt/</guid><description>&lt;p&gt;A prompt is for communicating an idea to an LLM.&lt;/p&gt;
&lt;p&gt;If you tell your idea to an AI agent and ask it to write down in its own words, say, a &amp;ldquo;skill&amp;rdquo; document which is meant to be a reusable piece of prompt to attach to another LLM session, it becomes a classic game of telephone. The model inevitably misinterprets your unspoken context or invent its own rules that you never intended. So the process is guaranteed to be lossy, therefore worse than your original words.&lt;/p&gt;</description></item></channel></rss>