<?xml version="1.0" encoding="utf-8" standalone="yes"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
  <channel>
    <title>Coding Agents on Igor Kulman</title>
    <link>https://blog.kulman.sk/tags/coding-agents/</link>
    <description>Recent content in Coding Agents on Igor Kulman</description>
    <generator>Hugo -- gohugo.io</generator>
    <language>en-us</language>
    <lastBuildDate>Wed, 02 Sep 2026 05:29:12 +0100</lastBuildDate><atom:link href="https://blog.kulman.sk/tags/coding-agents/index.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>Why Pi is my AI coding harness</title>
      <link>https://blog.kulman.sk/pi-is-now-my-ai-harness/</link>
      <pubDate>Wed, 02 Sep 2026 05:29:12 +0100</pubDate>
      
      <guid>https://blog.kulman.sk/pi-is-now-my-ai-harness/</guid>
      <description>I want to experiment with AI models without rebuilding my coding workflow every time I switch. Pi has become the stable layer around those experiments.
By “harness”, I mean the command-line environment, tools, extensions, and instructions around the model.
I have been experimenting with AI models and coding tools quite a lot lately. At the moment, I pay $20/month for ChatGPT Plus and $20/month for Claude Pro. I get GitHub Copilot Pro for free as an open-source maintainer, and I sometimes use OpenCode Zen to try the models it offers for free.</description>
    </item>
    
  </channel>
</rss>
