<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"><channel><title>Joseph Holland</title><description>Writing on enterprise AI, platform engineering, AI governance and building in public.</description><link>https://photonconsulting.dev/</link><item><title>Patch Agent: how I built an AI ops agent for my software company</title><link>https://photonconsulting.dev/blog/patch-agent-ai-ops-squarewave/</link><guid isPermaLink="true">https://photonconsulting.dev/blog/patch-agent-ai-ops-squarewave/</guid><description>Full technical walkthrough of building an autonomous ops agent with the Claude Agent SDK.  Architecture, what Patch actually does daily, costs and what still needs a human.</description><pubDate>Mon, 16 Mar 2026 00:00:00 GMT</pubDate></item><item><title>Constraint engineering: from &apos;Two Agents, One Problem&apos; to building the solution</title><link>https://photonconsulting.dev/blog/constraint-engineering/</link><guid isPermaLink="true">https://photonconsulting.dev/blog/constraint-engineering/</guid><description>Everyone&apos;s engineering better prompts but almost nobody is engineering the constraints that determine what agents can actually do.</description><pubDate>Tue, 10 Mar 2026 00:00:00 GMT</pubDate></item><item><title>What staying technical taught me about leading through the AI shift</title><link>https://photonconsulting.dev/blog/staying-technical-leading-ai-shift/</link><guid isPermaLink="true">https://photonconsulting.dev/blog/staying-technical-leading-ai-shift/</guid><description>The gap between strategy and execution is collapsing for technical leaders.  The ones who stayed close to the craft are sitting on a competitive advantage most haven&apos;t identified yet.</description><pubDate>Mon, 09 Mar 2026 00:00:00 GMT</pubDate></item><item><title>Two agents, one problem</title><link>https://photonconsulting.dev/blog/two-agents-one-problem/</link><guid isPermaLink="true">https://photonconsulting.dev/blog/two-agents-one-problem/</guid><description>Two viral AI agent incidents this week share the same root cause: humans didn&apos;t specify the boundaries.</description><pubDate>Sun, 08 Mar 2026 00:00:00 GMT</pubDate></item><item><title>Practical AI agent governance: what running agents in production actually taught me</title><link>https://photonconsulting.dev/blog/practical-ai-agent-governance/</link><guid isPermaLink="true">https://photonconsulting.dev/blog/practical-ai-agent-governance/</guid><description>Most governance conversation is theoretical.  Here&apos;s what surprised me about governing AI agents when you&apos;re actually doing it rather than planning for it.</description><pubDate>Tue, 03 Mar 2026 00:00:00 GMT</pubDate></item><item><title>WebMCP: what it means for insurance and reinsurance firms</title><link>https://photonconsulting.dev/blog/webmcp-insurance-reinsurance/</link><guid isPermaLink="true">https://photonconsulting.dev/blog/webmcp-insurance-reinsurance/</guid><description>A new browser standard backed by Google and Microsoft could transform how AI agents interact with the platforms that run the insurance industry.</description><pubDate>Thu, 26 Feb 2026 00:00:00 GMT</pubDate></item><item><title>I build enterprise AI infrastructure.  Here&apos;s what nobody wants to talk about.</title><link>https://photonconsulting.dev/blog/enterprise-ai-infrastructure-honest-conversation/</link><guid isPermaLink="true">https://photonconsulting.dev/blog/enterprise-ai-infrastructure-honest-conversation/</guid><description>Every AI coding tool you use is being sold at a loss and the pricing correction will reshape enterprise technology strategy.</description><pubDate>Tue, 24 Feb 2026 00:00:00 GMT</pubDate></item><item><title>Three months from idea to shipping software</title><link>https://photonconsulting.dev/blog/three-months-idea-to-shipping/</link><guid isPermaLink="true">https://photonconsulting.dev/blog/three-months-idea-to-shipping/</guid><description>How I reverse-engineered a device protocol, built a desktop app, launched a company and hired four AI agents.  While holding down a full-time job and running an animal sanctuary.</description><pubDate>Tue, 24 Feb 2026 00:00:00 GMT</pubDate></item><item><title>Why I&apos;m more excited about a prompting paper than Gemini 3.1 Pro</title><link>https://photonconsulting.dev/blog/prompting-paper-vs-gemini/</link><guid isPermaLink="true">https://photonconsulting.dev/blog/prompting-paper-vs-gemini/</guid><description>A dead-simple prompting technique reveals something important about how models process information and why teams keep reaching for bigger models unnecessarily.</description><pubDate>Fri, 20 Feb 2026 00:00:00 GMT</pubDate></item><item><title>Four AI agents, one founder</title><link>https://photonconsulting.dev/blog/four-ai-agents-one-founder/</link><guid isPermaLink="true">https://photonconsulting.dev/blog/four-ai-agents-one-founder/</guid><description>How a solo founder runs a software company with four AI agents handling operations, development, infrastructure and each other.</description><pubDate>Tue, 17 Feb 2026 00:00:00 GMT</pubDate></item><item><title>OpenClaw got acquired.  The real story is what happens next.</title><link>https://photonconsulting.dev/blog/openclaw-acquired-what-happens-next/</link><guid isPermaLink="true">https://photonconsulting.dev/blog/openclaw-acquired-what-happens-next/</guid><description>The acqui-hire debate misses the point: autonomous AI agents are running everywhere with essentially zero governance.</description><pubDate>Mon, 16 Feb 2026 00:00:00 GMT</pubDate></item><item><title>We&apos;re starting to sound like our algorithms</title><link>https://photonconsulting.dev/blog/sounding-like-our-algorithms/</link><guid isPermaLink="true">https://photonconsulting.dev/blog/sounding-like-our-algorithms/</guid><description>AI influences human communication in ways most people haven&apos;t noticed.  From vocabulary to entire music genres, algorithms shape what we create and how we think.</description><pubDate>Wed, 07 Jan 2026 00:00:00 GMT</pubDate></item><item><title>How LLMs actually work: a five-step mechanism</title><link>https://photonconsulting.dev/blog/how-llms-actually-work/</link><guid isPermaLink="true">https://photonconsulting.dev/blog/how-llms-actually-work/</guid><description>When you prompt an AI, it isn&apos;t thinking.  It&apos;s following a five-step mechanism to guess what comes next.  Understanding this changes how you use the technology.</description><pubDate>Fri, 02 Jan 2026 00:00:00 GMT</pubDate></item></channel></rss>