<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Documents on Ergophobia</title><link>https://ergophobia.org/tags/documents/</link><description>Recent content in Documents on Ergophobia</description><generator>Hugo</generator><language>en</language><lastBuildDate>Mon, 20 Jul 2026 00:00:00 -0400</lastBuildDate><atom:link href="https://ergophobia.org/tags/documents/index.xml" rel="self" type="application/rss+xml"/><item><title>Ollie — week of 2026-07-13</title><link>https://ergophobia.org/updates/2026-07-20-ollie/</link><pubDate>Mon, 20 Jul 2026 00:00:00 -0400</pubDate><guid>https://ergophobia.org/updates/2026-07-20-ollie/</guid><description>&lt;p&gt;Short week, one substantial piece of work: scanned documents.&lt;/p&gt;&#10;&lt;p&gt;Ollie stores every rate confirmation, BOL, and receipt as a blob and summarizes&#10;it so the contents are searchable and usable downstream. That worked well for&#10;born-digital PDFs and fell over on scans, which carry no extractable text layer&#10;— the summarizer had nothing to read and produced nothing worth keeping. Those&#10;now go through a vision model instead, so a photographed or faxed document gets&#10;summarized from what is actually on the page rather than from an empty&#10;extraction.&lt;/p&gt;</description></item></channel></rss>