<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"><channel><title>Adrian Dussan</title><description>Writing on software, infrastructure and the craft of building systems that can be rebuilt from nothing.</description><link>https://adriandussan.com/</link><language>en</language><item><title>How to Write Prompts That Don&apos;t Hallucinate: Proven Techniques with Examples</title><link>https://adriandussan.com/articles/how-to-write-prompts-that-dont-hallucinate/</link><guid isPermaLink="true">https://adriandussan.com/articles/how-to-write-prompts-that-dont-hallucinate/</guid><description>Language models are fluent even when they are wrong. This guide explains why hallucinations happen and walks through ten practical prompting techniques that reduce them, from grounding and explicit exits to citations, output constraints, and self-verification, each with clear before-and-after examples.</description><pubDate>Sun, 06 Sep 2026 01:22:23 GMT</pubDate></item><item><title>The first thing ever written here</title><link>https://adriandussan.com/articles/first-proof-1788544276163/</link><guid isPermaLink="true">https://adriandussan.com/articles/first-proof-1788544276163/</guid><description>Written through the editing interface over the encrypted tunnel, to prove the system stores and serves a word.</description><pubDate>Fri, 04 Sep 2026 17:51:16 GMT</pubDate></item></channel></rss>