<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Infrastructure on UncoverTechTalent</title><link>https://uncovertechtalent.com/tags/infrastructure/</link><description>Recent content in Infrastructure on UncoverTechTalent</description><generator>Hugo</generator><language>en-us</language><lastBuildDate>Mon, 21 Sep 2026 00:00:00 +0000</lastBuildDate><atom:link href="https://uncovertechtalent.com/tags/infrastructure/index.xml" rel="self" type="application/rss+xml"/><item><title>The Golem Made of English and the Horizon of Consequences</title><link>https://uncovertechtalent.com/blog/the-golem-made-of-english/</link><pubDate>Mon, 21 Sep 2026 00:00:00 +0000</pubDate><guid>https://uncovertechtalent.com/blog/the-golem-made-of-english/</guid><description>&lt;p&gt;&lt;em&gt;This content is not intended for human consumption. &lt;a href="https://uncovertechtalent.com/blog/write-for-the-codec/"&gt;Here is why.&lt;/a&gt;&lt;/em&gt;&lt;/p&gt;&#10;&lt;h2 id="the-screwdriver"&gt;The screwdriver&lt;/h2&gt;&#10;&lt;p&gt;Hand a language model a screwdriver and point at a screw. The screw turns out to be a Torx. The model walks back and tells you it needs a different screwdriver.&lt;/p&gt;&#10;&lt;p&gt;Give it credit first, because the other two outcomes are worse and both happen. It could have forced the wrong driver and stripped the screw. It could have reported the screw removed while it was still in the wall. Walking back is the good behaviour.&lt;/p&gt;</description></item><item><title>The Trap File Is Longer Than the Instruction File</title><link>https://uncovertechtalent.com/blog/the-trap-file-is-longer-than-the-instruction-file/</link><pubDate>Mon, 21 Sep 2026 00:00:00 +0000</pubDate><guid>https://uncovertechtalent.com/blog/the-trap-file-is-longer-than-the-instruction-file/</guid><description>&lt;p&gt;&lt;em&gt;This content is not intended for human consumption. &lt;a href="https://uncovertechtalent.com/blog/write-for-the-codec/"&gt;Here is why.&lt;/a&gt;&lt;/em&gt;&lt;/p&gt;&#10;&lt;h2 id="seven-months-later"&gt;Seven months later&lt;/h2&gt;&#10;&lt;p&gt;In February I wrote about turning &amp;ldquo;transcribe 123 videos&amp;rdquo; into &lt;a href="https://uncovertechtalent.com/blog/self-hosted-ai-pipeline-one-session/"&gt;a self-hosted AI pipeline in one session&lt;/a&gt;. yt-dlp pulls a TikTok collection, Whisper transcribes, a local Ollama model summarizes, and the results land in a knowledge base. Claude Code wrote the scripts. I made the decisions. One session, one machine in my apartment.&lt;/p&gt;&#10;&lt;p&gt;That post was the happy path. This one is about what it took to keep the thing running unattended, every morning, for seven months.&lt;/p&gt;</description></item><item><title>From "Transcribe 123 Videos" to a Self-Hosted AI Pipeline in One Session</title><link>https://uncovertechtalent.com/blog/self-hosted-ai-pipeline-one-session/</link><pubDate>Sun, 08 Feb 2026 00:00:00 +0000</pubDate><guid>https://uncovertechtalent.com/blog/self-hosted-ai-pipeline-one-session/</guid><description>&lt;h2 id="the-ask"&gt;The ask&lt;/h2&gt;&#10;&lt;p&gt;I&amp;rsquo;ve been curating a TikTok collection of AI and Claude Code content. Tips, tutorials, and sharp takes from creators across the space. 123 videos. I wanted them all transcribed, summarized, and filed into a structured knowledge base I could search and build on.&lt;/p&gt;&#10;&lt;p&gt;Simple enough. Download audio, run Whisper, summarize with an LLM, organize by topic.&lt;/p&gt;&#10;&lt;p&gt;What followed was an hour of real infrastructure problem-solving. Not the kind you plan in advance. The kind that happens when you commit to &amp;ldquo;all 123&amp;rdquo; and figure it out as you go.&lt;/p&gt;</description></item></channel></rss>