<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Zhipu on Danilo Falcão da Silva</title><link>https://falcao.org/tags/zhipu/</link><description>Recent content in Zhipu on Danilo Falcão da Silva</description><generator>Hugo</generator><language>en-us</language><lastBuildDate>Sat, 20 Jun 2026 01:30:00 -0300</lastBuildDate><atom:link href="https://falcao.org/tags/zhipu/index.xml" rel="self" type="application/rss+xml"/><item><title>How the Internet Lost Its Mind Over GLM-5.2</title><link>https://falcao.org/posts/glm-5-2-internet-lost-its-mind/</link><pubDate>Sat, 20 Jun 2026 01:30:00 -0300</pubDate><guid>https://falcao.org/posts/glm-5-2-internet-lost-its-mind/</guid><description>&lt;p>On June 17, 2026, Zhipu AI — the lab that ships under the Z.ai brand —
released GLM-5.2 under an MIT license. Open weights. Free to download,
free to fine-tune, free to run wherever you want. Within forty-eight
hours, my feeds had decided this was either the death of Anthropic, the
triumph of Chinese AI, or the most overhyped model of the year.&lt;/p>
&lt;p>It is none of those things. It is something more interesting and more
boring at the same time: the first open-weights model that I can&amp;rsquo;t
dismiss as &amp;ldquo;good for an open model.&amp;rdquo; Let me walk through what&amp;rsquo;s actually
in the box, what the internet got right, and where it falls apart the
moment you put real work through it.&lt;/p></description></item></channel></rss>