<?xml version="1.0" encoding="utf-8"?>
<feed xmlns="http://www.w3.org/2005/Atom">
  <title>PHNTM One — Build log</title>
  <subtitle>Engineering notes from building a private, local AI appliance by hand.</subtitle>
  <link href="https://www.phntmcore.com/feed.xml" rel="self"/>
  <link href="https://www.phntmcore.com/blog"/>
  <id>https://www.phntmcore.com/blog</id>
  <updated>2026-10-08T04:00:00Z</updated>
  <author><name>Jacob DeCamp</name></author>
<entry><title>The model was the easy part: turning a Pi into a desk assistant</title><link href="https://www.phntmcore.com/blog/from-pi-project-to-ai-appliance"/><id>https://www.phntmcore.com/blog/from-pi-project-to-ai-appliance</id><published>2026-10-08T04:00:00Z</published><updated>2026-10-08T04:00:00Z</updated><summary>Jacob DeCamp on building PHNTM One: shared RAM, local documents, original notes, honest failure reporting and the limits of offline AI.</summary></entry>
  
  <entry>
    <title>Software 0.8: autonomy, on the record</title>
    <link href="https://www.phntmcore.com/changelog"/>
    <id>https://www.phntmcore.com/changelog</id>
    <published>2026-09-26T00:00:00Z</published>
    <updated>2026-09-26T00:00:00Z</updated>
    <summary>Autonomous mode, repairs with a verified undo, a running score of every fix, and the September security work: every software release from the first builds to 0.8.</summary>
  </entry>
  <entry>
    <title>Running a private AI assistant on a Raspberry Pi 5</title>
    <link href="https://www.phntmcore.com/blog/private-ai-on-raspberry-pi-5"/>
    <id>https://www.phntmcore.com/blog/private-ai-on-raspberry-pi-5</id>
    <published>2026-08-20T00:00:00Z</published>
    <updated>2026-09-30T00:00:00Z</updated>
    <summary>The Raspberry Pi 5 hardware and local voice stack, the 8 GB memory budget, dated performance evidence, a real hardware clip and repeatable privacy checks.</summary>
  </entry>
  <entry>
    <title>The mini-PC I returned: choosing hardware for a local AI appliance</title>
    <link href="https://www.phntmcore.com/blog/choosing-hardware-for-a-local-ai-appliance"/>
    <id>https://www.phntmcore.com/blog/choosing-hardware-for-a-local-ai-appliance</id>
    <published>2026-08-20T00:00:00Z</published>
    <updated>2026-08-20T00:00:00Z</updated>
    <summary>Why an N5095 mini-PC got returned (no AVX2, an ACPI interrupt storm), what the prototype taught us, and why production PHNTM One units are built on the Raspberry Pi 5.</summary>
  </entry>
</feed>

