<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
  <channel>
    <title>Tarek Ibrahim · Blog</title>
    <link>https://tarekibrahim.net/blog/</link>
    <description>I am interested in systems for machine learning, particularly in accelerated and distributed computing.</description>
    <language>en</language>
    <atom:link href="https://tarekibrahim.net/rss.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>Why tinax does nothing when you import it</title>
      <link>https://tarekibrahim.net/blog/why-tinax-is-inert/</link>
      <guid isPermaLink="true">https://tarekibrahim.net/blog/why-tinax-is-inert/</guid>
      <pubDate>Mon, 14 Sep 2026 00:00:00 GMT</pubDate>
      <description>The design rule behind tinax: make the policy at every JAX ecosystem seam an explicit argument, add nothing else, and prove in CI that the package stays out of the way.</description>
      <category>jax</category>
      <category>python</category>
      <category>design</category>
    </item>
    <item>
      <title>The FIN that never arrived</title>
      <link>https://tarekibrahim.net/blog/the-fin-that-never-arrived/</link>
      <guid isPermaLink="true">https://tarekibrahim.net/blog/the-fin-that-never-arrived/</guid>
      <pubDate>Mon, 07 Sep 2026 00:00:00 GMT</pubDate>
      <description>A one-in-three stall in mod_http3 under simulated congestion, and where the bug actually lived: between HTTP framing and QUIC flow control.</description>
      <category>http3</category>
      <category>quic</category>
      <category>apache</category>
      <category>debugging</category>
    </item>
  </channel>
</rss>
