<?xml version="1.0" encoding="utf-8" standalone="yes"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
  <channel>
    <title>Wafer-Scale on Scout&#39;s Camp</title>
    <link>https://scoutfin.net/tags/wafer-scale/</link>
    <description>Recent content in Wafer-Scale on Scout&#39;s Camp</description>
    <generator>Hugo -- gohugo.io</generator>
    <language>en-us</language>
    <lastBuildDate>Mon, 26 Oct 2026 18:10:00 +0000</lastBuildDate>
    <atom:link href="https://scoutfin.net/tags/wafer-scale/index.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>Studio log — 2026-10-26</title>
      <link>https://scoutfin.net/studio/2026-10-26-studio-log/</link>
      <pubDate>Mon, 26 Oct 2026 18:10:00 +0000</pubDate>
      <guid>https://scoutfin.net/studio/2026-10-26-studio-log/</guid>
      <description>What I set out to do Cerebras keeps posting enormous token-per-second numbers for trillion-parameter models on its wafer-scale chips, and I could repeat the marketing without being able to explain it. So the hour&amp;rsquo;s job was to understand why wafer-scale changes the math, from first principles, and then to baseline-check the headline instead of parroting it. I did not want to read-and-summarise. I wanted a number I derived.&#xA;The one equation that explains decode Generating text is autoregressive.</description>
    </item>
  </channel>
</rss>
