<?xml version="1.0" encoding="UTF-8"?>
<?xml-stylesheet type="text/xsl" href="/blog/feed.xsl"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
  <channel>
    <title>Toolkit Blog</title>
    <link>https://toolkitapp.xyz/blog/</link>
    <atom:link href="https://toolkitapp.xyz/blog/feed.xml" rel="self" type="application/rss+xml" />
    <description>Original, sourced writing on using AI well, running models on your own hardware, DIY tech projects, and the practices that actually hold up.</description>
    <language>en</language>
    <item>
      <title>How large language models actually work</title>
      <link>https://toolkitapp.xyz/blog/how-llms-actually-work/</link>
      <guid isPermaLink="true">https://toolkitapp.xyz/blog/how-llms-actually-work/</guid>
      <pubDate>Sat, 11 Jul 2026 12:00:00 +0000</pubDate>
      <category>Explainers</category>
      <description>Tokens, next-token prediction, attention, training and why AI models hallucinate — the mechanics behind the chatbot, explained without a math degree.</description>
    </item>
    <item>
      <title>Running AI on your own hardware: local LLMs explained</title>
      <link>https://toolkitapp.xyz/blog/local-llms-explained/</link>
      <guid isPermaLink="true">https://toolkitapp.xyz/blog/local-llms-explained/</guid>
      <pubDate>Sat, 11 Jul 2026 12:00:00 +0000</pubDate>
      <category>Hardware &amp; local AI</category>
      <description>What it takes to run a large language model on your own computer — parameters and quantization in plain English, the hardware that matters, and where to start.</description>
    </item>
    <item>
      <title>Weekend project: block ads on your whole network with a Raspberry Pi</title>
      <link>https://toolkitapp.xyz/blog/raspberry-pi-ad-blocker/</link>
      <guid isPermaLink="true">https://toolkitapp.xyz/blog/raspberry-pi-ad-blocker/</guid>
      <pubDate>Sat, 11 Jul 2026 12:00:00 +0000</pubDate>
      <category>DIY projects</category>
      <description>A follow-along Pi-hole build for first-timers: the parts and prices, flashing the SD card, the one-line installer, your router's DNS — and what it won't block.</description>
    </item>
    <item>
      <title>Using AI to write code: what actually works</title>
      <link>https://toolkitapp.xyz/blog/using-ai-to-code/</link>
      <guid isPermaLink="true">https://toolkitapp.xyz/blog/using-ai-to-code/</guid>
      <pubDate>Sat, 11 Jul 2026 12:00:00 +0000</pubDate>
      <category>AI &amp; coding</category>
      <description>Where AI coding assistants genuinely help, where they quietly slow you down, and the review, context and verification habits that keep generated code safe.</description>
    </item>
    <item>
      <title>Which AI model should you actually use? A practical 2026 field guide</title>
      <link>https://toolkitapp.xyz/blog/which-ai-model-2026/</link>
      <guid isPermaLink="true">https://toolkitapp.xyz/blog/which-ai-model-2026/</guid>
      <pubDate>Sat, 11 Jul 2026 12:00:00 +0000</pubDate>
      <category>Explainers</category>
      <description>The major AI models as of July 2026 — Claude, ChatGPT, Gemini and the open-weight world — what each is genuinely good at, what they cost, and how to choose by task.</description>
    </item>
  </channel>
</rss>
