<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
  <channel>
    <title>Writing about AI and agents..</title>
    <link>https://www.derpx06.online/blog</link>
    <description>Read what Derpx06 has been up to recently</description>
    <language>en</language>
    <lastBuildDate>Mon, 31 Aug 2026 17:16:15 GMT</lastBuildDate>
    <atom:link href="https://www.derpx06.online/api/rss" rel="self" type="application/rss+xml" />
    <managingEditor>manaskhairnar1511@gmail.com (Derpx06)</managingEditor>
    <webMaster>manaskhairnar1511@gmail.com (Derpx06)</webMaster>
    <image>
      <url>https://www.derpx06.online/images/derpx06-avatar.png</url>
      <title>Writing about AI and agents..</title>
      <link>https://www.derpx06.online/blog</link>
    </image>
    
    <item>
      <title>B-Trees or LSM-Trees: Your Database Already Chose</title>
      <link>https://www.derpx06.online/blog/btrees-or-lsm-trees</link>
      <guid>https://www.derpx06.online/blog/btrees-or-lsm-trees</guid>
      <pubDate>Thu, 15 Oct 2026 09:00:00 GMT</pubDate>
      <description><![CDATA[Every database picks a side in the same trade — quick to write, or quick to read. Almost every performance surprise you have had follows from which side it picked.]]></description>
      
      <category>Databases</category>
      <author>manaskhairnar1511@gmail.com (Derpx06)</author>
    </item>
    <item>
      <title>Mobile Apps Are Distributed Systems With Terrible Networks</title>
      <link>https://www.derpx06.online/blog/mobile-apps-are-distributed-systems</link>
      <guid>https://www.derpx06.online/blog/mobile-apps-are-distributed-systems</guid>
      <pubDate>Thu, 08 Oct 2026 09:00:00 GMT</pubDate>
      <description><![CDATA[A phone is a copy of your data that vanishes without warning, holds changes for hours, and comes back disagreeing with everyone else. Most mobile bugs are really that problem, unlabelled.]]></description>
      
      <category>Distributed Systems</category>
      <author>manaskhairnar1511@gmail.com (Derpx06)</author>
    </item>
    <item>
      <title>What I Got Wrong About Document Modelling</title>
      <link>https://www.derpx06.online/blog/what-i-got-wrong-about-document-modelling</link>
      <guid>https://www.derpx06.online/blog/what-i-got-wrong-about-document-modelling</guid>
      <pubDate>Thu, 01 Oct 2026 09:00:00 GMT</pubDate>
      <description><![CDATA[Picking MongoDB takes five minutes. Deciding what one record should contain takes months to get wrong and is painful to undo.]]></description>
      
      <category>Databases</category>
      <author>manaskhairnar1511@gmail.com (Derpx06)</author>
    </item>
    <item>
      <title>Your Retry Made It Worse</title>
      <link>https://www.derpx06.online/blog/your-retry-made-it-worse</link>
      <guid>https://www.derpx06.online/blog/your-retry-made-it-worse</guid>
      <pubDate>Thu, 24 Sep 2026 09:00:00 GMT</pubDate>
      <description><![CDATA[Trying again after a failure is the most obvious safety net there is, and the easiest one to build backwards. It hinges on one question: did the first attempt actually fail?]]></description>
      
      <category>Distributed Systems</category>
      <author>manaskhairnar1511@gmail.com (Derpx06)</author>
    </item>
    <item>
      <title>MCP Is a Protocol, Not a Plugin System</title>
      <link>https://www.derpx06.online/blog/mcp-is-a-protocol-not-a-plugin-system</link>
      <guid>https://www.derpx06.online/blog/mcp-is-a-protocol-not-a-plugin-system</guid>
      <pubDate>Thu, 17 Sep 2026 09:00:00 GMT</pubDate>
      <description><![CDATA[The Model Context Protocol looks like a way to hand an AI some plugins. Treat it that way and you get tools that shine in a demo and go round in circles in real use.]]></description>
      
      <category>Agents</category>
      <author>manaskhairnar1511@gmail.com (Derpx06)</author>
    </item>
    <item>
      <title>Local-First AI Is a Systems Problem, Not a Model Problem</title>
      <link>https://www.derpx06.online/blog/local-first-ai-is-a-systems-problem</link>
      <guid>https://www.derpx06.online/blog/local-first-ai-is-a-systems-problem</guid>
      <pubDate>Thu, 10 Sep 2026 09:00:00 GMT</pubDate>
      <description><![CDATA[Getting an AI running on your own laptop takes four minutes. Everything that makes it actually usable afterwards is queueing, memory and patience — plain old engineering.]]></description>
      
      <category>LLM Systems</category>
      <author>manaskhairnar1511@gmail.com (Derpx06)</author>
    </item>
    <item>
      <title>Evals Are the Only Moat</title>
      <link>https://www.derpx06.online/blog/evals-are-the-only-moat</link>
      <guid>https://www.derpx06.online/blog/evals-are-the-only-moat</guid>
      <pubDate>Thu, 03 Sep 2026 09:00:00 GMT</pubDate>
      <description><![CDATA[Anyone can copy your prompts in a weekend. Nobody can copy the thing that tells you, precisely, whether your last change made the product better.]]></description>
      
      <category>Evaluation</category>
      <author>manaskhairnar1511@gmail.com (Derpx06)</author>
    </item>
    <item>
      <title>Your RAG System Doesn't Have a Retrieval Problem</title>
      <link>https://www.derpx06.online/blog/rag-is-not-a-retrieval-problem</link>
      <guid>https://www.derpx06.online/blog/rag-is-not-a-retrieval-problem</guid>
      <pubDate>Thu, 27 Aug 2026 09:00:00 GMT</pubDate>
      <description><![CDATA[When an AI answers badly from your own documents, everyone blames the search. In every system I have debugged, the search was fine. The documents were the problem.]]></description>
      
      <category>Retrieval</category>
      <author>manaskhairnar1511@gmail.com (Derpx06)</author>
    </item>
    <item>
      <title>I Fine-Tuned a 0.5B Model to Write Shell Commands</title>
      <link>https://www.derpx06.online/blog/finetuning-a-05b-model-for-shell-commands</link>
      <guid>https://www.derpx06.online/blog/finetuning-a-05b-model-for-shell-commands</guid>
      <pubDate>Thu, 20 Aug 2026 09:00:00 GMT</pubDate>
      <description><![CDATA[A tiny AI running on my laptop beats a giant one at a single job. That sounds like a story about size. It is really a story about picking a small enough job.]]></description>
      
      <category>Models</category>
      <author>manaskhairnar1511@gmail.com (Derpx06)</author>
    </item>
    <item>
      <title>Inside Claude's J-Space: Anthropic's Global Workspace Research</title>
      <link>https://www.derpx06.online/blog/anthropic-j-space-global-workspace</link>
      <guid>https://www.derpx06.online/blog/anthropic-j-space-global-workspace</guid>
      <pubDate>Fri, 10 Jul 2026 00:00:00 GMT</pubDate>
      <description><![CDATA[Anthropic found a small, editable workspace inside Claude that carries silent intermediate thoughts. Here is what the J-lens reveals, why it matters, and what it does not prove.]]></description>
      
      <category>AI Interpretability</category>
      <author>manaskhairnar1511@gmail.com (Derpx06)</author>
    </item>
    <item>
      <title>KV Cache from First Principles</title>
      <link>https://www.derpx06.online/blog/kv-cache-from-first-principles</link>
      <guid>https://www.derpx06.online/blog/kv-cache-from-first-principles</guid>
      <pubDate>Fri, 27 Mar 2026 00:00:00 GMT</pubDate>
      <description><![CDATA[A clear first-principles guide to KV cache: why it exists, how it works, and what tradeoffs matter in real production systems.]]></description>
      <enclosure url="https://www.derpx06.online/images/blog/image.png" type="image/png" />
      <category>LLM Systems</category>
      <author>manaskhairnar1511@gmail.com (Derpx06)</author>
    </item>
  </channel>
</rss>