<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"><channel><title>The shad0w blog</title><link>https://2796gaurav.github.io/shad0w/blog/</link>
<description>System Zero, certified LLM shortcuts, and shad0w in production.</description><language>en</language><item><title>Zero-code System Zero for the OpenAI Decisions API</title><link>https://2796gaurav.github.io/shad0w/blog/decisions-api-proxy.html</link><guid>https://2796gaurav.github.io/shad0w/blog/decisions-api-proxy.html</guid>
<pubDate>Fri, 09 Oct 2026 09:00:00 +0000</pubDate><description>Put a certified reflex in front of client.decisions.create with three commands and no code change. Shadow mode first, the dashboard, the agreement check, the cost maths and the limits.</description></item><item><title>Snake at 120 moves a second: System 0 vs System One models, at game speed</title><link>https://2796gaurav.github.io/shad0w/blog/game-speed.html</link><guid>https://2796gaurav.github.io/shad0w/blog/game-speed.html</guid>
<pubDate>Fri, 09 Oct 2026 09:00:00 +0000</pubDate><description>shad0w learned snake and Pong from an LLM&#x27;s answers. Then it raced three System One decision models on an Apple M5 Pro&#x27;s GPU and a hosted LLM through the same games, with every decision timed. Here is what happened, and the Pong you can play.</description></item><item><title>94.3%, 51.3%, 0%: honest numbers from shadowing a hosted LLM</title><link>https://2796gaurav.github.io/shad0w/blog/honest-numbers.html</link><guid>https://2796gaurav.github.io/shad0w/blog/honest-numbers.html</guid>
<pubDate>Fri, 09 Oct 2026 09:00:00 +0000</pubDate><description>We put shad0w in front of a hosted LLM through the official OpenAI client and measured how much shad0w answered, how often it agreed with the LLM, how fast, and where it answered nothing at all.</description></item><item><title>Adding a category without breaking production</title><link>https://2796gaurav.github.io/shad0w/blog/adding-a-category.html</link><guid>https://2796gaurav.github.io/shad0w/blog/adding-a-category.html</guid>
<pubDate>Thu, 08 Oct 2026 09:00:00 +0000</pubDate><description>A trained shortcut and a changing list of options are a dangerous mix. The rule shad0w follows, and what happens when you add, remove or rename an option.</description></item><item><title>System Zero: the reflex layer below your LLM</title><link>https://2796gaurav.github.io/shad0w/blog/system-zero.html</link><guid>https://2796gaurav.github.io/shad0w/blog/system-zero.html</guid>
<pubDate>Thu, 08 Oct 2026 09:00:00 +0000</pubDate><description>shad0w is the shadow of your LLM. It watches your LLM, learns to answer like it, and takes the repeat traffic, so the LLM only gets what shad0w is not sure about. What System Zero is, why a cache is not enough, and where shad0w stays silent.</description></item><item><title>How shad0w proves it agrees with your LLM</title><link>https://2796gaurav.github.io/shad0w/blog/how-the-certificate-works.html</link><guid>https://2796gaurav.github.io/shad0w/blog/how-the-certificate-works.html</guid>
<pubDate>Wed, 07 Oct 2026 09:00:00 +0000</pubDate><description>The certificate behind shad0w, in plain words: a held-out test, a confidence line, an allowance for bad luck, and what the promise does and does not cover.</description></item></channel></rss>