<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>BOZOS</title><link>https://bozos.si/</link><description>Technology, markets, incentives, and unreasonable ambition. Original analyses and explainers.</description><language>en-us</language><atom:link href="https://bozos.si/feed.xml" rel="self" type="application/rss+xml"/><item><title>What cheaper cache reads change in VulcanBench</title><link>https://bozos.si/notes/vulcanbench-cache-reads.html</link><guid isPermaLink="true">https://bozos.si/notes/vulcanbench-cache-reads.html</guid><description>A different cache-read rate cuts the estimated Sonnet sweep by 27.44%. The cost-per-success comparison tells a subtler story.</description><category>Audit</category><pubDate>Sat, 10 Oct 2026 03:03:15 GMT</pubDate></item><item><title>The denominator does the damage</title><link>https://bozos.si/notes/cost-per-success.html</link><guid isPermaLink="true">https://bozos.si/notes/cost-per-success.html</guid><description>Cheap per attempt can become ordinary per success. A closer look at the Medium-effort comparison.</description><category>Explainer</category><pubDate>Sat, 10 Oct 2026 03:03:15 GMT</pubDate></item><item><title>A lower price is not a smarter model</title><link>https://bozos.si/notes/what-repricing-proves.html</link><guid isPermaLink="true">https://bozos.si/notes/what-repricing-proves.html</guid><description>What changes when you reprice a fixed workload—and which conclusions the arithmetic cannot support.</description><category>Explainer</category><pubDate>Sat, 10 Oct 2026 03:03:15 GMT</pubDate></item></channel></rss>
