<rss version="2.0" xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Hacker News: siris9476</title><link>https://news.ycombinator.com/user?id=siris9476</link><description>Hacker News RSS</description><docs>https://hnrss.org/</docs><generator>hnrss v2.1.1</generator><lastBuildDate>Thu, 03 Sep 2026 09:26:23 +0000</lastBuildDate><atom:link href="https://hnrss.org/user?id=siris9476" rel="self" type="application/rss+xml"></atom:link><item><title><![CDATA[Show HN: PulsarForge – run a 744B MoE model on 32GB RAM with zero GPU (pure C)]]></title><description><![CDATA[
<p>Article URL: <a href="https://github.com/siris9476/pulsarforge">https://github.com/siris9476/pulsarforge</a></p>
<p>Comments URL: <a href="https://news.ycombinator.com/item?id=49535090">https://news.ycombinator.com/item?id=49535090</a></p>
<p>Points: 4</p>
<p># Comments: 0</p>
]]></description><pubDate>Wed, 02 Sep 2026 12:12:07 +0000</pubDate><link>https://github.com/siris9476/pulsarforge</link><dc:creator>siris9476</dc:creator><comments>https://news.ycombinator.com/item?id=49535090</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49535090</guid></item><item><title><![CDATA[New comment by siris9476 in "Show HN: Running 104GB Qwen3.8-Flash-Next on 48GB Mac with at ~12 tok/s"]]></title><description><![CDATA[
<p>32GB dedicated to an N-gram table instead of a draft model is an unusual choice for speculative decoding — what made it win over the more common draft-model approach here?</p>
]]></description><pubDate>Tue, 01 Sep 2026 21:53:30 +0000</pubDate><link>https://news.ycombinator.com/item?id=49528769</link><dc:creator>siris9476</dc:creator><comments>https://news.ycombinator.com/item?id=49528769</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49528769</guid></item></channel></rss>