<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
  <channel>
    <title>The Minutes of Dylan Patel on interconnect</title>
    <link>https://minutesof.com/dylan-patel/on/interconnect/</link>
    <description>Everything Dylan Patel has said on interconnect: 2 verbatim quotes between February 2023 and July 2026, each with a timestamp and a link to the recording…</description>
    <language>en</language>
    <lastBuildDate>Mon, 31 Aug 2026 13:10:14 +0000</lastBuildDate>
    <atom:link href="https://minutesof.com/dylan-patel/on/interconnect/feed.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>Patel explains NVLink connects only 72 GPUs while Google ICI connects 8,000 chips without switches, creating a</title>
      <link>https://minutesof.com/q/5ca387af-4b9c-4a90-9c49-5374505ffeb1/</link>
      <guid isPermaLink="true">https://minutesof.com/q/5ca387af-4b9c-4a90-9c49-5374505ffeb1/</guid>
      <description>“NVIDIA, the NVLink can only connect 72 GPUs. For Google, their ICI can connect 8,000 chips at super high bandwidth, but you have to pass through other chips to get there because there&#x27;s no switch.” — The AGI Post</description>
      <pubDate>Mon, 06 Jul 2026 22:07:54 +0000</pubDate>
      <category>interconnect</category>
    </item>
    <item>
      <title>Patel argues interconnects between servers are now the key bottleneck, not individual chip speed.</title>
      <link>https://minutesof.com/q/9e80e3ac-3937-4220-880d-5bc04e1351f9/</link>
      <guid isPermaLink="true">https://minutesof.com/q/9e80e3ac-3937-4220-880d-5bc04e1351f9/</guid>
      <description>“The interconnects is the biggest bottleneck right? And so the problem set is it moved from you know just one, you know, hey, my chip is faster.” — Gradient Flow</description>
      <pubDate>Thu, 02 Feb 2023 12:00:39 +0000</pubDate>
      <category>interconnect</category>
    </item>
  </channel>
</rss>
