<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
  <channel>
    <title>TheAILAB Blog</title>
    <link>https://theailab.party/blog/</link>
    <description>Articles about automating business work, running AI on your own hardware and automating trading strategies.</description>
    <language>en-au</language>
    <atom:link href="https://theailab.party/blog/feed.xml" rel="self" type="application/rss+xml"/>
    <lastBuildDate>Sat, 10 Oct 2026 18:20:00 GMT</lastBuildDate>
    <item>
      <title>How much VRAM do you need to run an LLM locally?</title>
      <link>https://theailab.party/blog/how-much-vram-to-run-an-llm-locally/</link>
      <guid isPermaLink="true">https://theailab.party/blog/how-much-vram-to-run-an-llm-locally/</guid>
      <pubDate>Sat, 10 Oct 2026 18:20:00 GMT</pubDate>
      <category>Local AI</category>
      <description>Work out the GPU memory a local LLM needs from its parameter count, quantisation and context length, with worked examples from official docs.</description>
    </item>
    <item>
      <title>Why your backtest looks better than your live trading results</title>
      <link>https://theailab.party/blog/why-your-backtest-looks-better-than-live-results/</link>
      <guid isPermaLink="true">https://theailab.party/blog/why-your-backtest-looks-better-than-live-results/</guid>
      <pubDate>Sat, 10 Oct 2026 16:00:00 GMT</pubDate>
      <category>Algorithmic Trading</category>
      <description>Overfitting, look-ahead bias, survivorship bias and missing costs explain why most backtests disappoint in live trading. Here&#39;s how to catch each one.</description>
    </item>
    <item>
      <title>Prompt injection: the security risk every AI automation must plan for</title>
      <link>https://theailab.party/blog/prompt-injection-ai-automation-security/</link>
      <guid isPermaLink="true">https://theailab.party/blog/prompt-injection-ai-automation-security/</guid>
      <pubDate>Sat, 10 Oct 2026 12:00:00 GMT</pubDate>
      <category>Data Privacy &amp; Security</category>
      <description>Prompt injection lets text hidden in emails, documents or web pages hijack an AI assistant. What it is, why there&#39;s no full fix, and how to design around it.</description>
    </item>
  </channel>
</rss>
