<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Ai on C Shells</title><link>https://cshellsbytheseashore.dev/tags/ai/</link><description>Recent content in Ai on C Shells</description><generator>Hugo</generator><language>en-GB</language><lastBuildDate>Sat, 26 Sep 2026 20:00:00 +0100</lastBuildDate><atom:link href="https://cshellsbytheseashore.dev/tags/ai/index.xml" rel="self" type="application/rss+xml"/><item><title>Questions we couldn't afford to answer</title><link>https://cshellsbytheseashore.dev/2026/questions-we-couldnt-afford/</link><pubDate>Sat, 26 Sep 2026 20:00:00 +0100</pubDate><guid>https://cshellsbytheseashore.dev/2026/questions-we-couldnt-afford/</guid><description>Most talk about models is about writing more code, sooner. The bigger change is that answers which used to cost days now cost an hour, and in an agency that changes what good work looks like.</description></item><item><title>We don't know our most important dependency</title><link>https://cshellsbytheseashore.dev/2026/models-are-dependencies/</link><pubDate>Sat, 26 Sep 2026 00:00:00 +0000</pubDate><guid>https://cshellsbytheseashore.dev/2026/models-are-dependencies/</guid><description>We pin every package in the repo, then hand a large share of our work to a model we know only by reputation. What it would take to know it properly.</description></item></channel></rss>