<?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>prompt-engineering on Sparse Notes</title>
    <link>https://sparsenotes.com/tags/prompt-engineering/</link>
    <description>Recent content in prompt-engineering on Sparse Notes</description>
    <image>
      <url>https://sparsenotes.com/images/og-default.png</url>
      <link>https://sparsenotes.com/images/og-default.png</link>
    </image>
    <generator>Hugo -- gohugo.io</generator>
    <lastBuildDate>Sat, 07 Nov 2026 00:00:00 +0000</lastBuildDate><atom:link href="https://sparsenotes.com/tags/prompt-engineering/index.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>Your Old Prompts Are Making Claude Worse (Lance Martin, Anthropic)</title>
      <link>https://sparsenotes.com/posts/2026/11/unhobbling-the-frontier-lance-martin/</link>
      <pubDate>Sat, 07 Nov 2026 00:00:00 +0000</pubDate>
      
      <guid>https://sparsenotes.com/posts/2026/11/unhobbling-the-frontier-lance-martin/</guid>
      <description>Lance Martin argues most LLM applications sit inside the cost/quality frontier because of fixable misconfigurations: stale prompt habits, broken prompt caching, unclimbable evals, overfit hill-climbing and harnesses not built for multi-day agents.</description>
    </item>
    
  </channel>
</rss>
