<?xml version="1.0" encoding="utf-8" standalone="yes"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
	<channel>
		<title>SemIf on transtorm.ai</title>
		<link>https://transtorm.ai/tags/semif/</link>
		<description>Recent content in SemIf on transtorm.ai</description>
		<generator>Hugo</generator>
		<language>en</language>
		
		
		
		
			<lastBuildDate>Sat, 26 Sep 2026 00:00:00 +0000</lastBuildDate>
		
			<atom:link href="https://transtorm.ai/tags/semif/index.xml" rel="self" type="application/rss+xml" />
			<item>
				<title>Qwen3.8-Flash-Next applies the fundus laterality rule backwards</title>
				<link>https://transtorm.ai/blog/qwen-fundus-laterality-reversed-rule/</link>
				<pubDate>Sat, 26 Sep 2026 00:00:00 +0000</pubDate>
				<guid>https://transtorm.ai/blog/qwen-fundus-laterality-reversed-rule/</guid>
				<description>Qwen3.8-Flash-Next locates the optic disc correctly in 120 decisions out of 120, then gets the eye wrong on 51 images out of 60. The protocol, the numbers and the fix we adopted.</description>
			</item>
			<item>
				<title>Decision models (1): you don&#39;t need a thousand words to decide</title>
				<link>https://transtorm.ai/blog/decision-models-1-you-dont-need-a-thousand-words-to-decide/</link>
				<pubDate>Fri, 18 Sep 2026 00:00:00 +0000</pubDate>
				<guid>https://transtorm.ai/blog/decision-models-1-you-dont-need-a-thousand-words-to-decide/</guid>
				<description>For three years, we have been asking models built to write to make decisions. A new family of models does nothing but decide. First part: what a decision model is, with Jev and SemIf.</description>
			</item>
			<item>
				<title>Decision models (2): what speed changes</title>
				<link>https://transtorm.ai/blog/decision-models-2-what-speed-changes/</link>
				<pubDate>Fri, 18 Sep 2026 00:00:00 +0000</pubDate>
				<guid>https://transtorm.ai/blog/decision-models-2-what-speed-changes/</guid>
				<description>A decision model that answers in a hundred milliseconds is not used the way a language model that answers in ten seconds is. Decision loops, comparing hypotheses, the generator-and-judge pairing, and the return of the small specialized model: the second part of our article on Jev and SemIf.</description>
			</item>
	</channel>
</rss>
