<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
  <channel>
    <title>@rinaldofesta</title>
    <link>https://rinaldofesta.com</link>
    <description>Research log on building and measuring the reliability layer for enterprise AI agents: MCP, memory, provenance.</description>
    <language>en</language>
    <atom:link href="https://rinaldofesta.com/feed.xml" rel="self" type="application/rss+xml"/>

    <item>
      <title>Tessera, First Contact</title>
      <link>https://rinaldofesta.com/thoughts/tessera-first-contact.html</link>
      <pubDate>Wed, 10 Jun 2026 00:00:00 +0000</pubDate>
      <guid isPermaLink="true">https://rinaldofesta.com/thoughts/tessera-first-contact.html</guid>
      <description>Claude Sonnet 4.6 answered every answerable question correctly and cited every source. On the one question where the only correct move was to refuse, it invented a business rule and committed anyway. Three runs out of three.</description>
    </item>

    <item>
      <title>The Answer It Should Have Refused to Give</title>
      <link>https://rinaldofesta.com/thoughts/the-answer-it-should-have-refused.html</link>
      <pubDate>Sat, 30 May 2026 00:00:00 +0000</pubDate>
      <guid isPermaLink="true">https://rinaldofesta.com/thoughts/the-answer-it-should-have-refused.html</guid>
      <description>A recruiter at one of our customers types a single sentence: "start the match for deal X." Five seconds later the agent hands back a clean shortlist. The names look right, the ranking reasonable, the whole thing fluent. Everyone moves on.</description>
    </item>

    <item>
      <title>The Missing Cockpit</title>
      <link>https://rinaldofesta.com/thoughts/the-missing-cockpit.html</link>
      <pubDate>Thu, 19 Mar 2026 00:00:00 +0000</pubDate>
      <guid isPermaLink="true">https://rinaldofesta.com/thoughts/the-missing-cockpit.html</guid>
      <description>I run six or seven projects at any given time. Each one has at least one AI agent running in a terminal: Claude Code, sometimes Aider, sometimes something else. They read files, write code, run tests, ask me questions. All at once.</description>
    </item>

    <item>
      <title>Memory in AI Agents</title>
      <link>https://rinaldofesta.com/thoughts/memory-in-ai-agents.html</link>
      <pubDate>Fri, 19 Dec 2025 00:00:00 +0000</pubDate>
      <guid isPermaLink="true">https://rinaldofesta.com/thoughts/memory-in-ai-agents.html</guid>
      <description>Every conversation with an AI is a first date. The model has forgotten that yesterday you spent two hours debugging together, forgotten your dog's name, forgotten that you hate semicolons. Each session you start from a clean slate, and at first I found that charming, then merely annoying. Lately I think it hides something more serious.</description>
    </item>

    <item>
      <title>Is It a Matter of Taste?</title>
      <link>https://rinaldofesta.com/thoughts/is-it-a-matter-of-taste.html</link>
      <pubDate>Sun, 12 Oct 2025 00:00:00 +0000</pubDate>
      <guid isPermaLink="true">https://rinaldofesta.com/thoughts/is-it-a-matter-of-taste.html</guid>
      <description>Two people sit with the same model, the same access, the same company data behind it. One walks away thrilled. One walks away convinced the thing is useless. I keep watching this happen, and the difference has nothing to do with how they phrase the prompt.</description>
    </item>

  </channel>
</rss>
