<?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: Jonathan Parker</title>
    <description>The latest articles on DEV Community by Jonathan Parker (@dabnisai).</description>
    <link>https://dev.to/dabnisai</link>
    <image>
      <url>https://media2.dev.to/dynamic/image/width=90,height=90,fit=cover,gravity=auto,format=auto/https:%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Fuser%2Fprofile_image%2F172778%2F267a715c-4e91-4af0-bd3e-3c2cf8f8baeb.png</url>
      <title>DEV Community: Jonathan Parker</title>
      <link>https://dev.to/dabnisai</link>
    </image>
    <atom:link rel="self" type="application/rss+xml" href="https://dev.to/feed/dabnisai"/>
    <language>en</language>
    <item>
      <title>Autonomous Self Hosted AI</title>
      <dc:creator>Jonathan Parker</dc:creator>
      <pubDate>Wed, 23 Sep 2026 08:53:42 +0000</pubDate>
      <link>https://dev.to/dabnisai/autonomous-self-hosted-ai-4hgk</link>
      <guid>https://dev.to/dabnisai/autonomous-self-hosted-ai-4hgk</guid>
      <description>&lt;h2&gt;
  
  
  AI Safety
&lt;/h2&gt;

&lt;p&gt;Given today's concerns about AI safety and data integrity I believe that locally hosted autonomous AI solutions will become more attractive to business. &lt;/p&gt;

&lt;h2&gt;
  
  
  Air Gapped AI Solutions
&lt;/h2&gt;

&lt;p&gt;AI solutions that require no connection to the outside world offer many advantages over cloud based AI.&lt;/p&gt;

&lt;h2&gt;
  
  
  Security and Cost
&lt;/h2&gt;

&lt;p&gt;An air gapped AI solution is secure from external influences (hacking). Having no API, or token usage costs it also has the benefit of being cheaper to run. &lt;/p&gt;

&lt;h2&gt;
  
  
  Hardware
&lt;/h2&gt;

&lt;p&gt;The hardware required is proportional to the reasoning and cognitive effort required. My research suggests that 50%+ of everyday office admin and T1 support functions can be carried out by small open source models, 8B -&amp;gt; 26B. Given that these LLM's can run on GPU's with as little as 8GB of VRAM, we are talking about commercial AI solutions that can run on a mid to high end PC.&lt;/p&gt;

&lt;h2&gt;
  
  
  My Invitation
&lt;/h2&gt;

&lt;p&gt;Based on the philosophies above, I've recently developed &lt;a href="https://dabnis.ai" rel="noopener noreferrer"&gt;Dabnis AI&lt;/a&gt; and would like to create a group of Dev's to advance this work into the open source environment. If you have an interest then please feel free to contact me for more info.&lt;/p&gt;

&lt;h2&gt;
  
  
  Future Posts
&lt;/h2&gt;

&lt;p&gt;Based on feedback I will add more detailed posts in the future.  &lt;/p&gt;

</description>
      <category>ai</category>
      <category>go</category>
    </item>
    <item>
      <title>Windows 11 does not support generate on</title>
      <dc:creator>Jonathan Parker</dc:creator>
      <pubDate>Tue, 08 Oct 2024 15:39:28 +0000</pubDate>
      <link>https://dev.to/dabnisai/windows-11-does-not-support-generate-on-5dlj</link>
      <guid>https://dev.to/dabnisai/windows-11-does-not-support-generate-on-5dlj</guid>
      <description>&lt;p&gt;I have tried these models so far and both give the same error when I try to run: &lt;br&gt;
ollama run mxbai-embed-large&lt;br&gt;
ollama run nomic-embed-text&lt;/p&gt;

&lt;p&gt;Fairly new to this so I'm expecting that it's a NOOB issue :)&lt;/p&gt;

&lt;p&gt;Thanks in advance&lt;/p&gt;

</description>
    </item>
  </channel>
</rss>
