<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"><channel><title>Esfis blog</title><link>https://www.esfis.com/blog</link><description>Notes on memory and retrieval.</description><item><title>Introducing Esfis</title><link>https://www.esfis.com/blog/introducing-esfis</link><guid>https://www.esfis.com/blog/introducing-esfis</guid><pubDate>Tue, 08 Sep 2026 00:00:00 GMT</pubDate><description>One local memory for every assistant you use. Why we built it, what it does today, and what we are deliberately not doing.</description></item><item><title>Cascade retrieval, or why we run four searches instead of one</title><link>https://www.esfis.com/blog/cascade-retrieval</link><guid>https://www.esfis.com/blog/cascade-retrieval</guid><pubDate>Wed, 26 Aug 2026 00:00:00 GMT</pubDate><description>Vector search alone loses exact names and dates. Keyword search alone loses paraphrase. We run both, expand through the graph, and rerank. Here is what each stage buys.</description></item><item><title>Memory you can diff</title><link>https://www.esfis.com/blog/memory-you-can-diff</link><guid>https://www.esfis.com/blog/memory-you-can-diff</guid><pubDate>Wed, 12 Aug 2026 00:00:00 GMT</pubDate><description>Every change to your memory is a commit. Here is what that makes possible, from &quot;what did I know in June&quot; to undoing a bad import.</description></item><item><title>One message in, four memories out</title><link>https://www.esfis.com/blog/one-message-four-memories</link><guid>https://www.esfis.com/blog/one-message-four-memories</guid><pubDate>Thu, 30 Jul 2026 00:00:00 GMT</pubDate><description>What extraction actually does to a paragraph, and why we split aggressively instead of storing whole messages.</description></item></channel></rss>