<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:content="http://purl.org/rss/1.0/modules/content/"><channel><title>Claude on Ramblings of a cloud engineer</title><link>https://skarlso.github.io/categories/claude/</link><description>Recent content in Claude on Ramblings of a cloud engineer</description><generator>Hugo -- 0.136.0</generator><language>en</language><lastBuildDate>Tue, 17 Feb 2026 01:01:00 +0100</lastBuildDate><atom:link href="https://skarlso.github.io/categories/claude/index.xml" rel="self" type="application/rss+xml"/><item><title>How to use ai efficiently without going brain dead</title><link>https://skarlso.github.io/2026/02/17/how-to-use-ai-efficiently-without-going-braindead/</link><pubDate>Tue, 17 Feb 2026 01:01:00 +0100</pubDate><guid>https://skarlso.github.io/2026/02/17/how-to-use-ai-efficiently-without-going-braindead/</guid><description>&lt;p>I know I &lt;a href="https://skarlso.github.io/2025/06/07/re-my-ai-skeptic-friends-are-all-nuts/">spoke out&lt;/a> against AI in the past, and I still think those thoughts are warranted. That&amp;rsquo;s why I would like to write down how we actually should use AI to do the things we want to do without losing our critical thinking in the process.&lt;/p>
&lt;p>Here&amp;rsquo;s my flow&amp;hellip;&lt;/p>
&lt;h2 id="how-to-use-llms-responsibly">How to use LLMs responsibly&lt;/h2>
&lt;p>There always needs to be feedback. Don&amp;rsquo;t use it to think, use it to ask questions, verify, check, expanding your understanding, your knowledge.&lt;/p></description></item></channel></rss>