<?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 performance</title>
    <link>https://minutesof.com/dylan-patel/on/performance/</link>
    <description>Everything Dylan Patel has said on performance: 3 verbatim quotes between February 2023 and July 2026, each with a timestamp and a link to the recording…</description>
    <language>en</language>
    <lastBuildDate>Sun, 30 Aug 2026 16:22:18 +0000</lastBuildDate>
    <atom:link href="https://minutesof.com/dylan-patel/on/performance/feed.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>Patel predicts another 50-100% Blackwell performance improvement from software optimization of standard librar</title>
      <link>https://minutesof.com/q/1220feb8-33a2-4114-958d-7e559ad3806a/</link>
      <guid isPermaLink="true">https://minutesof.com/q/1220feb8-33a2-4114-958d-7e559ad3806a/</guid>
      <description>“I see another 50% to doubling in the cards, just from the fact that there is so much low hanging fruit or performance left on the table if you are using the general open libraries, right? Whether it&#x27;s cuDNN, VLM, SGLANG, all these other things,” — Together AI</description>
      <pubDate>Fri, 03 Oct 2025 20:23:30 +0000</pubDate>
      <category>performance</category>
    </item>
    <item>
      <title>Patel predicts general-purpose Blackwell users could see another performance doubling over the next year from</title>
      <link>https://minutesof.com/q/e17fa153-5a23-488d-99aa-a4fe4fc4a705/</link>
      <guid isPermaLink="true">https://minutesof.com/q/e17fa153-5a23-488d-99aa-a4fe4fc4a705/</guid>
      <description>“for the general purpose user, yeah, you&#x27;re going to have another doubling of performance potentially on the tables over the next year.” — Together AI</description>
      <pubDate>Fri, 03 Oct 2025 20:23:30 +0000</pubDate>
      <category>performance</category>
    </item>
    <item>
      <title>Patel reports PyTorch 2.0 delivered 30% performance improvement on 7,000 popular GitHub models.</title>
      <link>https://minutesof.com/q/5a9a7e9d-e509-4a4c-a7ad-30fdd24c4722/</link>
      <guid isPermaLink="true">https://minutesof.com/q/5a9a7e9d-e509-4a4c-a7ad-30fdd24c4722/</guid>
      <description>“So I think I think I think, the PyTorch Foundation tested the 7,000 most popular most starred models on GitHub, and they got, like, a 30% increase.” — Gradient Flow</description>
      <pubDate>Thu, 02 Feb 2023 12:00:39 +0000</pubDate>
      <category>performance</category>
    </item>
  </channel>
</rss>
