<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:dc="http://purl.org/dc/elements/1.1/">
  <channel>
    <title>Forem: Ariel Onoriaga</title>
    <description>The latest articles on Forem by Ariel Onoriaga (@arielonoriaga).</description>
    <link>https://forem.com/arielonoriaga</link>
    <image>
      <url>https://media2.dev.to/dynamic/image/width=90,height=90,fit=cover,gravity=auto,format=auto/https:%2F%2Fdev-to-uploads.s3.amazonaws.com%2Fuploads%2Fuser%2Fprofile_image%2F559471%2F8288486c-c37b-44e6-9a4d-8e46367f14b3.png</url>
      <title>Forem: Ariel Onoriaga</title>
      <link>https://forem.com/arielonoriaga</link>
    </image>
    <atom:link rel="self" type="application/rss+xml" href="https://forem.com/feed/arielonoriaga"/>
    <language>en</language>
    <item>
      <title>Finally, I've fixed my AI workflow</title>
      <dc:creator>Ariel Onoriaga</dc:creator>
      <pubDate>Fri, 08 May 2026 12:36:10 +0000</pubDate>
      <link>https://forem.com/arielonoriaga/finally-ive-fixed-my-ai-workflow-1opo</link>
      <guid>https://forem.com/arielonoriaga/finally-ive-fixed-my-ai-workflow-1opo</guid>
      <description>&lt;h2&gt;
  
  
  I think most developers are approaching AI coding the wrong way
&lt;/h2&gt;

&lt;p&gt;I used to think AI coding quality was mostly about the model.&lt;/p&gt;

&lt;blockquote&gt;
&lt;p&gt;Better model = better code.&lt;/p&gt;
&lt;/blockquote&gt;

&lt;p&gt;After building my repo, I don’t think that anymore.&lt;/p&gt;

&lt;p&gt;I realized AI usually writes bad code for the same reason humans do:&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;Huge files&lt;/li&gt;
&lt;li&gt;Mixed responsibilities&lt;/li&gt;
&lt;li&gt;Hidden business rules&lt;/li&gt;
&lt;li&gt;No architectural boundaries&lt;/li&gt;
&lt;li&gt;Too much implicit context&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;So instead of focusing on prompts, I focused on the repository itself.&lt;/p&gt;

&lt;p&gt;The real question became:&lt;/p&gt;

&lt;h2&gt;
  
  
  “What if the codebase was designed for AI readability and reasoning?”
&lt;/h2&gt;

&lt;p&gt;That completely changed the results.&lt;/p&gt;

&lt;p&gt;The AI stopped generating random boilerplate and started producing code that actually felt:&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;Scalable&lt;/li&gt;
&lt;li&gt;Predictable&lt;/li&gt;
&lt;li&gt;Maintainable&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;I noticed:&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;Shorter prompts&lt;/li&gt;
&lt;li&gt;Better architectural decisions&lt;/li&gt;
&lt;li&gt;Cleaner abstractions&lt;/li&gt;
&lt;li&gt;Less “fighting the AI”&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;And honestly...&lt;/p&gt;

&lt;p&gt;This repository changed how I see software architecture.&lt;/p&gt;

&lt;p&gt;I think one of the most valuable engineering skills in the next few years won’t just be writing code —&lt;/p&gt;

&lt;h2&gt;
  
  
  It’ll be designing systems that humans &lt;em&gt;and AI&lt;/em&gt; can reason about together.
&lt;/h2&gt;

&lt;p&gt;Repo:&lt;/p&gt;

&lt;p&gt;&lt;a href="https://github.com/arielonoriaga/claude-irondev" rel="noopener noreferrer"&gt;Claude IronDev&lt;/a&gt;&lt;/p&gt;

</description>
      <category>ai</category>
      <category>architecture</category>
      <category>productivity</category>
      <category>programming</category>
    </item>
  </channel>
</rss>
