<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0"><channel>
<title>rods.sh</title><link>https://rods.sh/en/</link><description>Mainframe veteran, product owner and builder. Daily notes on AI.</description><language>en-US</language>
<item><title>Treat AI like a sharp junior teammate, not an oracle</title><link>https://rods.sh/en/posts/2026-11-05-ai-as-a-junior-teammate/</link><guid>https://rods.sh/en/posts/2026-11-05-ai-as-a-junior-teammate/</guid><pubDate>Thu, 05 Nov 2026 00:00:00 GMT</pubDate><description>The most useful mental model I&#39;ve found for working with AI every day.</description></item>
<item><title>Context is the new working storage</title><link>https://rods.sh/en/posts/2026-11-04-context-is-working-storage/</link><guid>https://rods.sh/en/posts/2026-11-04-context-is-working-storage/</guid><pubDate>Wed, 04 Nov 2026 00:00:00 GMT</pubDate><description>Old programmers obsessed over every byte of memory. Prompting well is the same habit with a new name.</description></item>
<item><title>What the mainframe already taught us about AI agents</title><link>https://rods.sh/en/posts/2026-11-03-mainframe-lessons-for-ai-agents/</link><guid>https://rods.sh/en/posts/2026-11-03-mainframe-lessons-for-ai-agents/</guid><pubDate>Tue, 03 Nov 2026 00:00:00 GMT</pubDate><description>Checkpoints, restartability and audit trails aren&#39;t legacy ideas. They&#39;re exactly what agents are missing.</description></item>
</channel></rss>
