<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Tutorial on Notes from the Rabbit Hole</title><link>https://magnus919.com/tags/tutorial/</link><description>Recent content in Tutorial on Notes from the Rabbit Hole</description><generator>Hugo</generator><language>en</language><copyright>© [Magnus Hedemark](https://github.com/magnus919)</copyright><lastBuildDate>Thu, 14 May 2026 00:00:00 -0400</lastBuildDate><atom:link href="https://magnus919.com/tags/tutorial/index.xml" rel="self" type="application/rss+xml"/><item><title>Get Your Agent a Brain: Getting Started with hermes-cashew</title><link>https://magnus919.com/2026/05/get-your-agent-a-brain-getting-started-with-hermes-cashew/</link><pubDate>Thu, 14 May 2026 00:00:00 -0400</pubDate><guid>https://magnus919.com/2026/05/get-your-agent-a-brain-getting-started-with-hermes-cashew/</guid><description>&lt;p>You know the feeling. You fire up your agent, ask a question about something you worked on together last week, and it stares back at you like you just met. &amp;ldquo;I don&amp;rsquo;t have context from previous sessions.&amp;rdquo; You explain the project again. You restate your preferences. You remind it who you are.&lt;/p>
&lt;p>This isn&amp;rsquo;t an intelligence problem. It&amp;rsquo;s a memory architecture problem.&lt;/p>
&lt;p>Most AI tools treat every conversation as a blank slate. ChatGPT doesn&amp;rsquo;t remember you from yesterday. Claude doesn&amp;rsquo;t carry context from one project to the next. Some bolt on a vector database and call it memory: dump embeddings in, pull the nearest neighbors out, hope the semantic space lines up with what you actually needed. It works about as well as you&amp;rsquo;d expect.&lt;/p></description></item></channel></rss>