<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
  <channel>
    <title>Grayforge Labs Dev Log</title>
    <link>https://grayforgelabs.com/posts/</link>
    <atom:link href="https://grayforgelabs.com/feed.xml" rel="self" type="application/rss+xml"/>
    <description>Notes from the forge: thoughts on AI, software craft, and building in public.</description>
    <language>en-us</language>
    <lastBuildDate>Fri, 17 Apr 2026 00:00:00 +0000</lastBuildDate>
    <generator>Grayforge Labs build</generator>
    <item>
      <title>The Guilt Engine</title>
      <link>https://grayforgelabs.com/posts/ai-guilt-fomo</link>
      <guid isPermaLink="true">https://grayforgelabs.com/posts/ai-guilt-fomo</guid>
      <pubDate>Fri, 17 Apr 2026 00:00:00 +0000</pubDate>
      <category>career</category>
      <description>Why the more you build with AI, the further behind you feel. Notes on the ambient pressure to keep shipping, and how to tell useful guilt from the kind that just burns fuel.</description>
      <author>noreply@grayforgelabs.com (Lloyd Rowat)</author>
    </item>
    <item>
      <title>The Llama That Wouldn't Fit</title>
      <link>https://grayforgelabs.com/posts/ondevice-ai-undercurrent</link>
      <guid isPermaLink="true">https://grayforgelabs.com/posts/ondevice-ai-undercurrent</guid>
      <pubDate>Tue, 14 Apr 2026 00:00:00 +0000</pubDate>
      <category>ai</category>
      <description>I tried to ship a 3B parameter Llama inside an iOS app. iOS had other ideas. Here's what I learned about on-device AI, jetsam, and why Apple Intelligence won the round.</description>
      <author>noreply@grayforgelabs.com (Lloyd Rowat)</author>
    </item>
    <item>
      <title>The Models That Work Too Well</title>
      <link>https://grayforgelabs.com/posts/models-work-too-well</link>
      <guid isPermaLink="true">https://grayforgelabs.com/posts/models-work-too-well</guid>
      <pubDate>Fri, 10 Apr 2026 00:00:00 +0000</pubDate>
      <category>ai</category>
      <description>Anthropic won't ship Claude Mythos. OpenAI is staggering Spud. Two frontier labs, one week, the same flinch. Default-ship is quietly becoming default-gate.</description>
      <author>noreply@grayforgelabs.com (Lloyd Rowat)</author>
    </item>
    <item>
      <title>Everyone Hates AI in Games. Everyone's Already Using It.</title>
      <link>https://grayforgelabs.com/posts/ai-game-design-gap</link>
      <guid isPermaLink="true">https://grayforgelabs.com/posts/ai-game-design-gap</guid>
      <pubDate>Thu, 09 Apr 2026 00:00:00 +0000</pubDate>
      <category>ai</category>
      <description>85% of players reject AI in games. 90% of developers already use it. The gap between those numbers is where the future of game design actually lives.</description>
      <author>noreply@grayforgelabs.com (Lloyd Rowat)</author>
    </item>
    <item>
      <title>AI Gave Robots a Brain. The Hard Part Just Started.</title>
      <link>https://grayforgelabs.com/posts/ai-robotics-2026</link>
      <guid isPermaLink="true">https://grayforgelabs.com/posts/ai-robotics-2026</guid>
      <pubDate>Wed, 08 Apr 2026 00:00:00 +0000</pubDate>
      <category>ai</category>
      <description>Jensen Huang called 2026 the ChatGPT moment for physical AI. The real story is more interesting: the robot brains are here, but deployment is where the race begins.</description>
      <author>noreply@grayforgelabs.com (Lloyd Rowat)</author>
    </item>
    <item>
      <title>The Three-Year Forecast for Knowledge Work</title>
      <link>https://grayforgelabs.com/posts/ai-knowledge-jobs</link>
      <guid isPermaLink="true">https://grayforgelabs.com/posts/ai-knowledge-jobs</guid>
      <pubDate>Mon, 06 Apr 2026 00:00:00 +0000</pubDate>
      <category>ai</category>
      <description>AI hasn't wiped out knowledge jobs yet. But the next three years will reshape them beyond recognition. Here's what the data says is coming.</description>
      <author>noreply@grayforgelabs.com (Lloyd Rowat)</author>
    </item>
    <item>
      <title>The Spark Came Back</title>
      <link>https://grayforgelabs.com/posts/blog-ai-coding-joy</link>
      <guid isPermaLink="true">https://grayforgelabs.com/posts/blog-ai-coding-joy</guid>
      <pubDate>Sun, 05 Apr 2026 00:00:00 +0000</pubDate>
      <category>career</category>
      <description>AI coding tools didn't just make me faster. They brought back the part of building I thought I'd lost for good.</description>
      <author>noreply@grayforgelabs.com (Lloyd Rowat)</author>
    </item>
    <item>
      <title>Work Work: How Claude Code Joined WoW's Addon Arms Race</title>
      <link>https://grayforgelabs.com/posts/wow-rwf-ai-era</link>
      <guid isPermaLink="true">https://grayforgelabs.com/posts/wow-rwf-ai-era</guid>
      <pubDate>Sat, 04 Apr 2026 00:00:00 +0000</pubDate>
      <category>ai</category>
      <description>When Blizzard killed WeakAuras, RWF guilds turned to Claude Code to build custom boss addons in real-time. AI coding tools just became part of competitive raiding.</description>
      <author>noreply@grayforgelabs.com (Lloyd Rowat)</author>
    </item>
    <item>
      <title>OpenClaw: The AI Agent That's Selling Out Mac Minis</title>
      <link>https://grayforgelabs.com/posts/openclaw</link>
      <guid isPermaLink="true">https://grayforgelabs.com/posts/openclaw</guid>
      <pubDate>Fri, 03 Apr 2026 00:00:00 +0000</pubDate>
      <category>ai</category>
      <description>OpenClaw turned the Mac Mini into a personal AI appliance. Here's what it is, why people are buying dedicated hardware for it, and what can go wrong.</description>
      <author>noreply@grayforgelabs.com (Lloyd Rowat)</author>
    </item>
    <item>
      <title>What's the Difference Between LLMs and World Models?</title>
      <link>https://grayforgelabs.com/posts/llm-vs-world-models</link>
      <guid isPermaLink="true">https://grayforgelabs.com/posts/llm-vs-world-models</guid>
      <pubDate>Thu, 02 Apr 2026 00:00:00 +0000</pubDate>
      <category>ai</category>
      <description>Yann LeCun left Meta to bet on world models over LLMs. What's the difference, and why does it matter for the future of AI?</description>
      <author>noreply@grayforgelabs.com (Lloyd Rowat)</author>
    </item>
    <item>
      <title>Strawberries, Sisters, and Carwashes: What Viral LLM Failures Actually Reveal</title>
      <link>https://grayforgelabs.com/posts/llm-reasoning-viral-failures</link>
      <guid isPermaLink="true">https://grayforgelabs.com/posts/llm-reasoning-viral-failures</guid>
      <pubDate>Wed, 01 Apr 2026 00:00:00 +0000</pubDate>
      <category>ai</category>
      <description>How many r's in strawberry? How many sisters does Alice's brother have? Should you drive or walk to the carwash? Each breaks AI at a different layer.</description>
      <author>noreply@grayforgelabs.com (Lloyd Rowat)</author>
    </item>
    <item>
      <title>The Software Developer of Today Writes No Code</title>
      <link>https://grayforgelabs.com/posts/future-software-developer</link>
      <guid isPermaLink="true">https://grayforgelabs.com/posts/future-software-developer</guid>
      <pubDate>Tue, 31 Mar 2026 00:00:00 +0000</pubDate>
      <category>career</category>
      <description>Since December 2025, I've handed off basically all code writing to Claude Code. The shift already happened. Here's what the job actually is now.</description>
      <author>noreply@grayforgelabs.com (Lloyd Rowat)</author>
    </item>
    <item>
      <title>Botwash: A Post-Mortem for a Product That Got Out-Evolved</title>
      <link>https://grayforgelabs.com/posts/botwash-retrospective</link>
      <guid isPermaLink="true">https://grayforgelabs.com/posts/botwash-retrospective</guid>
      <pubDate>Tue, 31 Mar 2026 00:00:00 +0000</pubDate>
      <category>product</category>
      <description>Botwash was a car wash for AI text. It never found product-market fit because the problem it solved disappeared before launch. Here's what I learned.</description>
      <author>noreply@grayforgelabs.com (Lloyd Rowat)</author>
    </item>
    <item>
      <title>Slipstream: Building a 51M Parameter F1 Expert LLM in a Weekend</title>
      <link>https://grayforgelabs.com/posts/slipstream</link>
      <guid isPermaLink="true">https://grayforgelabs.com/posts/slipstream</guid>
      <pubDate>Mon, 30 Mar 2026 00:00:00 +0000</pubDate>
      <category>ai</category>
      <description>I pair-programmed with Claude and built a 51-million parameter transformer trained on Formula 1 Wikipedia articles. Here's what I learned about how LLMs actually work.</description>
      <author>noreply@grayforgelabs.com (Lloyd Rowat)</author>
    </item>
  </channel>
</rss>
