<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"><channel><title>Savanty — Skelf-Research</title><description>Notes from Savanty: natural-language to Answer Set Programming, the typed self-repair loop, and what mathematical guarantees actually mean in practice.</description><link>https://savanty.skelfresearch.com/</link><language>en-us</language><item><title>The promise of &apos;mathematically guaranteed&apos; — what it actually means here</title><link>https://savanty.skelfresearch.com/blog/mathematically-guaranteed-what-it-means/</link><guid isPermaLink="true">https://savanty.skelfresearch.com/blog/mathematically-guaranteed-what-it-means/</guid><description>Marketing copy makes the phrase do a lot of work. Inside the pipeline, the guarantee has a precise scope, a precise boundary, and an honest gap. This is where each one lives.</description><pubDate>Fri, 29 May 2026 00:00:00 GMT</pubDate></item><item><title>From English to model: where translation breaks</title><link>https://savanty.skelfresearch.com/blog/from-english-to-model-where-translation-breaks/</link><guid isPermaLink="true">https://savanty.skelfresearch.com/blog/from-english-to-model-where-translation-breaks/</guid><description>The LLM is a translator. Translators get things wrong. The interesting question is which kinds of wrong are recoverable from inside the loop — and how.</description><pubDate>Fri, 08 May 2026 00:00:00 GMT</pubDate></item><item><title>Why LLMs alone don&apos;t solve optimization problems</title><link>https://savanty.skelfresearch.com/blog/llms-alone-dont-solve-optimization/</link><guid isPermaLink="true">https://savanty.skelfresearch.com/blog/llms-alone-dont-solve-optimization/</guid><description>A large language model can describe a shift schedule. Asking it to produce one is a different request entirely — and the failure mode is invisible until you look.</description><pubDate>Wed, 22 Apr 2026 00:00:00 GMT</pubDate></item></channel></rss>