<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Posts on gegare.com</title><link>https://gegare.com/posts/</link><description>Recent content in Posts on gegare.com</description><generator>Hugo</generator><language>en-us</language><lastBuildDate>Fri, 29 May 2026 00:00:00 +0000</lastBuildDate><atom:link href="https://gegare.com/posts/index.xml" rel="self" type="application/rss+xml"/><item><title>If The Mac Mini is the Brain, OpenClaw is the Nervous System - OpenClaw Installation</title><link>https://gegare.com/posts/if-the-mac-mini-is-the-brain-openclaw-is-the-nervous-system---openclaw-installation/</link><pubDate>Fri, 29 May 2026 00:00:00 +0000</pubDate><guid>https://gegare.com/posts/if-the-mac-mini-is-the-brain-openclaw-is-the-nervous-system---openclaw-installation/</guid><description>&lt;p&gt;I never really used AI all that much prior to this. I&amp;rsquo;ve had GPT or Claude write me up some scripts or helped me do some research in the past, but nothing huge like some of these developers out here. But when I heard about OpenClaw and its ability to ease of access ratio, I have to say it was one of those &amp;ldquo;oh wow&amp;rdquo; moments. OpenClaw is the piece that turns a local language model into an actual AI assistant. Without it, the Mac Mini is just a &amp;ldquo;dumb&amp;rdquo; model with no real applications other than chat, kind of like a brain without a body. With it, you have an agent with a persistent identity, vault access, tool calling, Telegram integration, and the ability to spawn specialist sub-agents. This post covers how I deployed it, the problems that I came across, and the solutions implemented to get around them.&lt;/p&gt;</description></item><item><title>Building the Librarian's Brain - Mac Mini M4 as a Headless AI Inference Server</title><link>https://gegare.com/posts/building-the-ai-agents-brain---mac-mini-m4-as-a-headless-ai-inference-server/</link><pubDate>Mon, 04 May 2026 00:00:00 +0000</pubDate><guid>https://gegare.com/posts/building-the-ai-agents-brain---mac-mini-m4-as-a-headless-ai-inference-server/</guid><description>&lt;h2 id="overview"&gt;
 Overview
 &lt;a class="heading-link" href="#overview"&gt;
 &lt;i class="fa-solid fa-link" aria-hidden="true" title="Link to heading"&gt;&lt;/i&gt;
 &lt;span class="sr-only"&gt;Link to heading&lt;/span&gt;
 &lt;/a&gt;
&lt;/h2&gt;
&lt;p&gt;For those of you just joining, I am working on a personal AI assistant that is intended to manage my digital life, synthesizing all the information I come across on a daily basis. I am calling it the Librarian, after Neal Stephenson&amp;rsquo;s AI agent from his novel &amp;ldquo;Snow Crash&amp;rdquo;. But where to actually begin?&lt;/p&gt;
&lt;p&gt;Well, first off, the Librarian needs a brain. Every query I send it, every research task, every intelligence synthesis — all of that reasoning has to happen somewhere. That somewhere is a Mac Mini M4 sitting on a shelf in my server rack, running completely headless, accessible only over SSH, serving local AI inference to the rest of my homelab stack.&lt;/p&gt;</description></item><item><title>Hello World</title><link>https://gegare.com/posts/hello-world/</link><pubDate>Mon, 27 Apr 2026 16:33:39 -0500</pubDate><guid>https://gegare.com/posts/hello-world/</guid><description>&lt;p&gt;Hello, my name is Dylan, and I&amp;rsquo;m looking to create something that I think is pretty cool.&lt;/p&gt;
&lt;p&gt;Have you ever read the Neal Stephenson book &amp;ldquo;Snow Crash&amp;rdquo;? It&amp;rsquo;s a satirical cyberpunk novel with an over the top world of mafia pizza delivery services and a virtual &amp;ldquo;Metaverse&amp;rdquo; (not to be confused with the company Meta&amp;rsquo;s failed VR platform) where most people spend their time living out their virtual fantasies. Very entertaining, I recommend it. Well, within this book there is an AI agent named The Librarian. This AI is the interface in which people access the nation&amp;rsquo;s privatized intelligence agencies. Essentially, its a talking encyclopedia that has full situational awareness of its user, knows the answers to any questions the user may have, and is capable of filling the gaps in knowledge the user seeks answers to. This got me thinking, our current LLM infrastructure can realistically do this same thing. Of course cloud models are getting more and more advanced at an especially fast pace, but I&amp;rsquo;m more interested in what someone could accomplish with local models, where the data the model interacts with is solely isolated on the users&amp;rsquo; machines. I want to create my own, self-hosted, self-contained version of Stephenson&amp;rsquo;s Snow Crash Librarian.&lt;/p&gt;</description></item></channel></rss>