<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
  <channel>
    <title>hardware: everyone on the record — The Minutes</title>
    <link>https://minutesof.com/on/hardware/</link>
    <description>Everyone on the record about hardware: 8 verbatim quotes from 4 people, April 2023 to August 2026, in one chronology, each with a timestamp and a link to…</description>
    <language>en</language>
    <lastBuildDate>Sun, 30 Aug 2026 22:33:00 +0000</lastBuildDate>
    <atom:link href="https://minutesof.com/on/hardware/feed.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>Dylan Patel: Patel argues silicon supply can be optimized for either high throughput or high interactivity use</title>
      <link>https://minutesof.com/q/6a3bd7b5-c74b-46a6-9dd5-8db72aa9576f/</link>
      <guid isPermaLink="true">https://minutesof.com/q/6a3bd7b5-c74b-46a6-9dd5-8db72aa9576f/</guid>
      <description>“Supply of silicon can go many ways. You can either leverage it to high throughput things or high interactivity things.” — Dylan Patel, SemiAnalysis</description>
      <pubDate>Mon, 17 Aug 2026 15:00:06 +0000</pubDate>
      <category>hardware</category>
    </item>
    <item>
      <title>Dylan Patel: Patel says InferenceX now has over $80 million of compute across multiple vendors.</title>
      <link>https://minutesof.com/q/eb8b0b35-dc54-4b31-af11-cfea01d719e6/</link>
      <guid isPermaLink="true">https://minutesof.com/q/eb8b0b35-dc54-4b31-af11-cfea01d719e6/</guid>
      <description>“we have over $80,000,000 of compute GPUs from NVIDIA AMD, TPUs from, Google, Tranium from Amazon, and we run this benchmark constantly on the newest inference engine, newest drivers, newest, PyTorch version,” — Dylan Patel, RAISE Summit</description>
      <pubDate>Thu, 16 Jul 2026 16:44:57 +0000</pubDate>
      <category>hardware</category>
    </item>
    <item>
      <title>Dylan Patel: Patel claims hardware improved 30x from Hopper to Blackwell for DeepSeek on optimized deployments</title>
      <link>https://minutesof.com/q/1aeaaf0c-1080-4d21-a79c-2bc3d127baf5/</link>
      <guid isPermaLink="true">https://minutesof.com/q/1aeaaf0c-1080-4d21-a79c-2bc3d127baf5/</guid>
      <description>“from Hopper to Blackwell, which is all we&#x27;ve had over the last three years, roughly 30x improvement on DeepSeek, on the most optimized deployment, which you can see on InferenceX there&#x27;s about a 30x improvement.” — Dylan Patel, Sequoia Capital</description>
      <pubDate>Tue, 30 Jun 2026 12:00:25 +0000</pubDate>
      <category>hardware</category>
    </item>
    <item>
      <title>Dylan Patel: Patel explains Bumpgate issue where thermal expansion differences caused GPU solder ball failures</title>
      <link>https://minutesof.com/q/0a5f3bdf-1c4b-45e3-ae44-c701f4cf2fee/</link>
      <guid isPermaLink="true">https://minutesof.com/q/0a5f3bdf-1c4b-45e3-ae44-c701f4cf2fee/</guid>
      <description>“And what ended up happening is because of that different rate of expansion, the solder balls connecting the chip and the board would crack.” — Dylan Patel, Matthew Berman</description>
      <pubDate>Mon, 30 Jun 2025 17:27:19 +0000</pubDate>
      <category>hardware</category>
    </item>
    <item>
      <title>David Sacks: Sacks cites Dylan Patel&#x27;s estimate that DeepSeek has 50,000 Hopper chips across different models.</title>
      <link>https://minutesof.com/q/a4ede3f9-51af-4927-b48a-d925bdf7fed7/</link>
      <guid isPermaLink="true">https://minutesof.com/q/a4ede3f9-51af-4927-b48a-d925bdf7fed7/</guid>
      <description>“Dylan Patel, who&#x27;s leading semiconductor analyst, has estimated that DeepSeac has about 50,000 hoppers. And specifically, he said they have about 10,000 h one hundreds, They have 10,000 h eight hundreds and 30,000 h twenties.” — David Sacks, All-In Podcast</description>
      <pubDate>Fri, 31 Jan 2025 21:52:00 +0000</pubDate>
      <category>hardware</category>
    </item>
    <item>
      <title>Dylan Patel: Patel claims four to five percent of NVIDIA GPUs fail during initial burn-in testing within two w</title>
      <link>https://minutesof.com/q/36c00658-21e4-44c9-9919-c879522edbb7/</link>
      <guid isPermaLink="true">https://minutesof.com/q/36c00658-21e4-44c9-9919-c879522edbb7/</guid>
      <description>“Buy a good chip from NVIDIA, and then it ends up failing like constantly, right?” — Dylan Patel, Scaling Intelligence</description>
      <pubDate>Tue, 12 Nov 2024 04:17:00 +0000</pubDate>
      <category>hardware</category>
    </item>
    <item>
      <title>Brad Gerstner: Gerstner estimates Apple needs 75 gigs of DRAM on phones to run 7B on-device models.</title>
      <link>https://minutesof.com/q/4c94cf62-8cc0-4ef5-9679-00a68b01f29c/</link>
      <guid isPermaLink="true">https://minutesof.com/q/4c94cf62-8cc0-4ef5-9679-00a68b01f29c/</guid>
      <description>“I think the question that Apple&#x27;s probably struggling with, Bill, is in order to do that, even a 7B model, you probably have to have, you know, 75 gigs of DRAM on your phone, right?” — Brad Gerstner, BG2 Pod</description>
      <pubDate>Thu, 16 May 2024 19:31:21 +0000</pubDate>
      <category>hardware</category>
    </item>
    <item>
      <title>David Friedberg: Friedberg argues Silicon Valley media coverage misses substantial hardware and biotech invest</title>
      <link>https://minutesof.com/q/b8f7ebed-51a0-4b43-a621-d7bd5cb2b1ae/</link>
      <guid isPermaLink="true">https://minutesof.com/q/b8f7ebed-51a0-4b43-a621-d7bd5cb2b1ae/</guid>
      <description>“There are many biotech companies. There are a number of kind of consumer hardware, there&#x27;s also a number of kind of enterprise hardware, lab automation.” — David Friedberg, Consumer VC with Mike Gelb</description>
      <pubDate>Sun, 30 Apr 2023 12:51:57 +0000</pubDate>
      <category>hardware</category>
    </item>
  </channel>
</rss>
