<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:dc="http://purl.org/dc/elements/1.1/">
  <channel>
    <title>DEV Community: Vic Luna</title>
    <description>The latest articles on DEV Community by Vic Luna (@serial-theorist).</description>
    <link>https://dev.to/serial-theorist</link>
    <image>
      <url>https://media2.dev.to/dynamic/image/width=90,height=90,fit=cover,gravity=auto/https:%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Fuser%2Fprofile_image%2F3973474%2Fd728ae71-046e-4e16-9fba-1470f8b942b5.webp</url>
      <title>DEV Community: Vic Luna</title>
      <link>https://dev.to/serial-theorist</link>
    </image>
    <atom:link rel="self" type="application/rss+xml" href="https://dev.to/feed/serial-theorist"/>
    <language>en</language>
    <item>
      <title>Mati: personal AI</title>
      <dc:creator>Vic Luna</dc:creator>
      <pubDate>Tue, 17 Nov 2026 16:59:02 +0000</pubDate>
      <link>https://dev.to/serial-theorist/mati-personal-ai-4a1</link>
      <guid>https://dev.to/serial-theorist/mati-personal-ai-4a1</guid>
      <description>&lt;p&gt;I’ve been working on Mati, a personal AI that runs entirely locally with encrypted memory. My goal was to model memory as a continuous process (a phase field) rather than a static database. I wanted to share the architecture and the honest limitations. For example, in my tests, field-based retrieval scored 0.036, while lexical retrieval scored 0.838. I document all of this, including the negative results, in the ARCHITECTURE.md. It’s built for CPUs, respects privacy by design, and is Apache-2.0. I’d love feedback from this community.(&lt;a href="https://github.com/vml520/Mati" rel="noopener noreferrer"&gt;https://github.com/vml520/Mati&lt;/a&gt;)&lt;/p&gt;

</description>
      <category>ai</category>
      <category>programming</category>
      <category>python</category>
      <category>architecture</category>
    </item>
  </channel>
</rss>
