<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0">
  <channel>
    <title>Chisl Notes</title>
    <link>https://www.chislgroup.com/notes/</link>
    <description>Short, practical notes from the Chisl team on building and running AI, data and software systems in production.</description>
    <language>en-GB</language>
    <item>
      <title>Jev, and why AI calls don't always need a chatbot</title>
      <link>https://www.chislgroup.com/notes/jev-system-one-models/</link>
      <guid>https://www.chislgroup.com/notes/jev-system-one-models/</guid>
      <pubDate>Fri, 02 Oct 2026 09:00:00 GMT</pubDate>
      <description>TypeSafe's Jev doesn't write text. It makes decisions in a fraction of a second, for a fraction of the price of a chat model. Here's why that matters and where we'd use it first.</description>
    </item>
    <item>
      <title>Why enterprise AI stalls, and what we've seen work</title>
      <link>https://www.chislgroup.com/notes/why-enterprise-ai-stalls/</link>
      <guid>https://www.chislgroup.com/notes/why-enterprise-ai-stalls/</guid>
      <pubDate>Thu, 01 Oct 2026 09:00:00 GMT</pubDate>
      <description>The evidence on AI at scale is still mixed. Here's why we think so many rollouts stall, and what we've seen work in the businesses getting real value from it.</description>
    </item>
    <item>
      <title>Some things we've learned scaling LLM inference in production</title>
      <link>https://www.chislgroup.com/notes/scaling-llm-inference/</link>
      <guid>https://www.chislgroup.com/notes/scaling-llm-inference/</guid>
      <pubDate>Fri, 18 Sep 2026 09:00:00 GMT</pubDate>
      <description>One of the platforms we build and run processes millions of pages of documents every month. Here is what the volume taught us about cost, latency and keeping users trusting the output.</description>
    </item>
  </channel>
</rss>
