<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
  <channel>
    <title>RTSN Studios — Insights</title>
    <link>https://rtsnstudios.com/insights</link>
    <atom:link href="https://rtsnstudios.com/rss.xml" rel="self" type="application/rss+xml"/>
    <description>Long-form essays from a Singapore studio operating brand and operations alongside founder-led teams — on brand drift, methodology transparency, and AI-operated workflows.</description>
    <language>en</language>
    <copyright>RTSN Studios</copyright>
    <managingEditor>kenny@rtsnstudios.com (Kenny)</managingEditor>
    <webMaster>kenny@rtsnstudios.com (Kenny)</webMaster>
    <lastBuildDate>Wed, 03 Jun 2026 09:00:00 +0800</lastBuildDate>
    <generator>Hand-authored static feed (manifesto/website/rss.xml)</generator>

    <item>
      <title>Knowledge Hub: AI Memory</title>
      <link>https://rtsnstudios.com/insights-knowledge-ai-memory</link>
      <guid isPermaLink="true">https://rtsnstudios.com/insights-knowledge-ai-memory</guid>
      <pubDate>Wed, 03 Jun 2026 09:00:00 +0800</pubDate>
      <category>AI That Operates</category>
      <description>How AI agent memory actually works, what RAG can and cannot do, the 100K-token cost inflection point, and what the 2025–2026 vendor landscape means for teams building AI-operated workflows.</description>
    </item>
    <item>
      <title>Knowledge Hub: The Knowledge Flywheel</title>
      <link>https://rtsnstudios.com/insights-knowledge-flywheel</link>
      <guid isPermaLink="true">https://rtsnstudios.com/insights-knowledge-flywheel</guid>
      <pubDate>Wed, 03 Jun 2026 09:00:00 +0800</pubDate>
      <category>AI That Operates</category>
      <description>The four-stage knowledge flywheel for AI-operated systems: execute, capture, distil, deploy. Why most flywheels stall at month three, and the disciplines that keep them compounding.</description>
    </item>
    <item>
      <title>Knowledge Hub: Orchestration</title>
      <link>https://rtsnstudios.com/insights-knowledge-orchestration</link>
      <guid isPermaLink="true">https://rtsnstudios.com/insights-knowledge-orchestration</guid>
      <pubDate>Wed, 03 Jun 2026 09:00:00 +0800</pubDate>
      <category>AI That Operates</category>
      <description>Production evidence on multi-agent orchestration: failure rates, topology tradeoffs, framework choices, and what running a system over months actually requires.</description>
    </item>
    <item>
      <title>Knowledge Hub: Skills + Tools</title>
      <link>https://rtsnstudios.com/insights-knowledge-skills-and-tools</link>
      <guid isPermaLink="true">https://rtsnstudios.com/insights-knowledge-skills-and-tools</guid>
      <pubDate>Wed, 03 Jun 2026 09:00:00 +0800</pubDate>
      <category>AI That Operates</category>
      <description>Function calling, MCP, Agent Skills, and tool governance — what founders and operators need to know about equipping AI agents with reliable capabilities.</description>
    </item>
    <item>
      <title>Knowledge Hub: Governance</title>
      <link>https://rtsnstudios.com/insights-knowledge-governance</link>
      <guid isPermaLink="true">https://rtsnstudios.com/insights-knowledge-governance</guid>
      <pubDate>Wed, 03 Jun 2026 09:00:00 +0800</pubDate>
      <category>AI That Operates</category>
      <description>What AI governance actually requires in 2026: the EU AI Act high-risk rules, Singapore&amp;rsquo;s agentic AI framework, audit trail standards, and the four failure patterns to design around.</description>
    </item>
    <item>
      <title>Knowledge Hub: Costs</title>
      <link>https://rtsnstudios.com/insights-knowledge-costs</link>
      <guid isPermaLink="true">https://rtsnstudios.com/insights-knowledge-costs</guid>
      <pubDate>Wed, 03 Jun 2026 09:00:00 +0800</pubDate>
      <category>AI That Operates</category>
      <description>Token prices fell 1,000x in three years. Agentic workflow costs compound at O(n²). A practical guide to cost discipline for AI-operated systems — from prompt caching to model routing.</description>
    </item>
    <item>
      <title>Knowledge Hub: Open vs Closed Models</title>
      <link>https://rtsnstudios.com/insights-knowledge-open-vs-closed-models</link>
      <guid isPermaLink="true">https://rtsnstudios.com/insights-knowledge-open-vs-closed-models</guid>
      <pubDate>Wed, 03 Jun 2026 09:00:00 +0800</pubDate>
      <category>AI That Operates</category>
      <description>The 2026 state of open-weight vs closed AI models for founders running AI-operated workflows: licensing, cost breakevens, agentic gaps, and the hybrid architecture that holds up.</description>
    </item>

    <item>
      <title>Why your brand book sits in a folder no one opens</title>
      <link>https://rtsnstudios.com/insights-brand-book-unopened</link>
      <guid isPermaLink="true">https://rtsnstudios.com/insights-brand-book-unopened</guid>
      <pubDate>Fri, 22 May 2026 09:00:00 +0800</pubDate>
      <category>The Drift Problem</category>
      <description>A brand book is a document, not a system. Four failure modes that keep a brand book dormant — the PDF problem, the searchability problem, the team-not-onboarded problem, and the structural no-API gap.</description>
    </item>

    <item>
      <title>Anatomy of S2 Verbal: walking through Ann Siang Chambers</title>
      <link>https://rtsnstudios.com/insights-anatomy-s2-verbal</link>
      <guid isPermaLink="true">https://rtsnstudios.com/insights-anatomy-s2-verbal</guid>
      <pubDate>Fri, 22 May 2026 09:00:00 +0800</pubDate>
      <category>Show the Work</category>
      <description>The methodology made visible — Section 2 Verbal walked end-to-end through six steps, using Ann Siang Chambers as the worked example. Not described, built.</description>
    </item>

    <item>
      <title>AI-operated vs AI-generated: the distinction that matters</title>
      <link>https://rtsnstudios.com/insights-ai-operated-vs-generated</link>
      <guid isPermaLink="true">https://rtsnstudios.com/insights-ai-operated-vs-generated</guid>
      <pubDate>Fri, 22 May 2026 09:00:00 +0800</pubDate>
      <category>AI That Operates</category>
      <description>AI-generated and AI-operated are different practices. They produce different artefacts, require different governance, and only one is sustainable as a service-tier discipline.</description>
    </item>

    <item>
      <title>Why the Workshop is two days, not five</title>
      <link>https://rtsnstudios.com/insights-workshop-two-days</link>
      <guid isPermaLink="true">https://rtsnstudios.com/insights-workshop-two-days</guid>
      <pubDate>Fri, 22 May 2026 09:00:00 +0800</pubDate>
      <category>The Founder's Brand</category>
      <description>The Discovery Workshop is two days by design. The reasoning is operational and cognitive — discovery quality is constraint-shaped, and founder attention degrades reliably after the second day.</description>
    </item>

    <item>
      <title>The Knowledge Hub: AI-operating knowledge</title>
      <link>https://rtsnstudios.com/insights-knowledge-hub</link>
      <guid isPermaLink="true">https://rtsnstudios.com/insights-knowledge-hub</guid>
      <pubDate>Fri, 22 May 2026 09:00:00 +0800</pubDate>
      <category>AI That Operates</category>
      <description>A modular, self-updating research surface. Seven topics on running AI inside a real business — memory, knowledge flywheels, orchestration, skills, governance, costs, and the open-vs-closed-model question. Updated monthly.</description>
    </item>

    <item>
      <title>What brand drift actually costs</title>
      <link>https://rtsnstudios.com/insights-brand-drift-cost</link>
      <guid isPermaLink="true">https://rtsnstudios.com/insights-brand-drift-cost</guid>
      <pubDate>Tue, 19 May 2026 09:00:00 +0800</pubDate>
      <category>The Drift Problem</category>
      <description>81% of companies produce off-brand content regularly; marketing leaders spend a fifth of their week correcting it. Brand drift is a documented operational tax — here is what the research shows.</description>
    </item>

  </channel>
</rss>
