<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0">
<channel>
  <title>AI Learn Home</title>
  <link>https://ailearnhome.blog/</link>
  <description>Practical notes on frontier models, local inference and evaluation.</description>
  <language>en</language>
  <lastBuildDate>Thu, 12 Nov 2026 10:00:00 +0000</lastBuildDate>

  <item>
    <title>Frontier models in 2026: what actually changed</title>
    <link>https://ailearnhome.blog/posts/frontier-models-in-2026.html</link>
    <guid>https://ailearnhome.blog/posts/frontier-models-in-2026.html</guid>
    <pubDate>Thu, 12 Nov 2026 10:00:00 +0000</pubDate>
    <description>Reasoning budgets, long context, native multimodality and the quiet collapse of inference prices.</description>
  </item>
  <item>
    <title>Running a capable model at home</title>
    <link>https://ailearnhome.blog/posts/running-llms-at-home.html</link>
    <guid>https://ailearnhome.blog/posts/running-llms-at-home.html</guid>
    <pubDate>Tue, 03 Nov 2026 10:00:00 +0000</pubDate>
    <description>VRAM arithmetic, quantization, tokens per second, and the point where a local model stops being a toy.</description>
  </item>
  <item>
    <title>Quantization, explained without the math</title>
    <link>https://ailearnhome.blog/posts/quantization-explained.html</link>
    <guid>https://ailearnhome.blog/posts/quantization-explained.html</guid>
    <pubDate>Wed, 21 Oct 2026 10:00:00 +0000</pubDate>
    <description>Why a 4-bit model still works, where quality actually degrades, and what perplexity hides.</description>
  </item>
  <item>
    <title>RAG that works</title>
    <link>https://ailearnhome.blog/posts/rag-that-works.html</link>
    <guid>https://ailearnhome.blog/posts/rag-that-works.html</guid>
    <pubDate>Thu, 08 Oct 2026 10:00:00 +0000</pubDate>
    <description>Chunking, hybrid retrieval, rerankers — and the evaluation step almost everyone skips.</description>
  </item>
  <item>
    <title>Agents and tool use: where they help</title>
    <link>https://ailearnhome.blog/posts/agents-and-tool-use.html</link>
    <guid>https://ailearnhome.blog/posts/agents-and-tool-use.html</guid>
    <pubDate>Thu, 24 Sep 2026 10:00:00 +0000</pubDate>
    <description>Reliable patterns, failure modes, and why sandboxing is not optional.</description>
  </item>
  <item>
    <title>How to evaluate a model for your own task</title>
    <link>https://ailearnhome.blog/posts/evaluating-models.html</link>
    <guid>https://ailearnhome.blog/posts/evaluating-models.html</guid>
    <pubDate>Fri, 11 Sep 2026 10:00:00 +0000</pubDate>
    <description>Golden sets, pairwise judging, and the traps of grading with a language model.</description>
  </item>
</channel>
</rss>
