<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Prompting on Philipp Metzler</title><link>https://chili-it.de/tags/prompting/</link><description>Recent content in Prompting on Philipp Metzler</description><generator>Hugo</generator><language>en-us</language><copyright>This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.</copyright><lastBuildDate>Mon, 21 Sep 2026 00:00:00 +0000</lastBuildDate><atom:link href="https://chili-it.de/tags/prompting/index.xml" rel="self" type="application/rss+xml"/><item><title>Prompt in Your Own Language: Clarity Beats the English Tax</title><link>https://chili-it.de/posts/prompt-in-your-own-language/</link><pubDate>Mon, 21 Sep 2026 00:00:00 +0000</pubDate><guid>https://chili-it.de/posts/prompt-in-your-own-language/</guid><description>&lt;p&gt;Here&amp;rsquo;s a question I keep coming back to: when I talk to an LLM, should I write in my native language, or switch to English?&lt;/p&gt;
&lt;p&gt;The case for English seems obvious. Most models are trained on more of it. The docs, the benchmarks, the forum threads the model learned from, all overwhelmingly English. So English output should be a little sharper.&lt;/p&gt;
&lt;p&gt;And it is. But less than you&amp;rsquo;d think.&lt;/p&gt;
&lt;p&gt;The gap has narrowed. On the current frontier models, German, my native language, is one of the best-covered languages after English. The quality difference is low single digits, not night and day. That advantage is real, but it&amp;rsquo;s small, and it&amp;rsquo;s shrinking with every model generation.&lt;/p&gt;</description></item></channel></rss>