<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"><channel><title>jcqb.dev — Jacques Bronkhorst</title><description>Field notes from the enterprise trenches and the home lab — AI, .NET, system design, and teaching a house to behave. By Jacques Bronkhorst.</description><link>https://jcqb.dev/</link><language>en-us</language><item><title>Consistent AI-generated covers for a static blog</title><link>https://jcqb.dev/writing/ai-generated-covers/</link><guid isPermaLink="true">https://jcqb.dev/writing/ai-generated-covers/</guid><description>Brand consistency with AI images isn&apos;t a prompting problem — it&apos;s a pipeline problem. Here&apos;s the one I built in a weekend.</description><pubDate>Sat, 20 Jun 2026 00:00:00 GMT</pubDate><category>ai</category></item><item><title>Skills vs MCP: when 50 tool calls is the smell</title><link>https://jcqb.dev/writing/skills-vs-mcp/</link><guid isPermaLink="true">https://jcqb.dev/writing/skills-vs-mcp/</guid><description>45 MCP tools and the parameter fatigue to match — forever tweaking arguments to do something simple. The fix was a skill that read my grudging docs.</description><pubDate>Wed, 01 Apr 2026 00:00:00 GMT</pubDate><category>ai</category></item><item><title>Ship the agent&apos;s code review, then ship an agent to doubt it</title><link>https://jcqb.dev/writing/ship-doubtful-agents/</link><guid isPermaLink="true">https://jcqb.dev/writing/ship-doubtful-agents/</guid><description>A single agent pass is plausible-but-unverified. The fix isn&apos;t more prompting — it&apos;s a separate agent whose only job is to refute the first one.</description><pubDate>Sun, 01 Mar 2026 00:00:00 GMT</pubDate><category>ai</category></item><item><title>Trust in AI is at an all-time low and your tests are why</title><link>https://jcqb.dev/writing/your-tests-are-why/</link><guid isPermaLink="true">https://jcqb.dev/writing/your-tests-are-why/</guid><description>84% of developers use AI tools. 46% don&apos;t trust the output. The gap isn&apos;t a PR problem — it&apos;s a testing problem, and I watched it happen.</description><pubDate>Wed, 01 Oct 2025 00:00:00 GMT</pubDate><category>ai</category></item><item><title>I gave an LLM root and lived to write about it</title><link>https://jcqb.dev/writing/i-gave-an-llm-root/</link><guid isPermaLink="true">https://jcqb.dev/writing/i-gave-an-llm-root/</guid><description>I handed a language model root on my gaming PC. It restarted my network to investigate the packet loss it had just caused. Hard lessons followed.</description><pubDate>Tue, 01 Jul 2025 00:00:00 GMT</pubDate><category>ai</category><category>agents</category><category>linux</category></item><item><title>Your conversation doesn&apos;t survive the terminal</title><link>https://jcqb.dev/writing/conversation-survive-terminal/</link><guid isPermaLink="true">https://jcqb.dev/writing/conversation-survive-terminal/</guid><description>Inspiration strikes in your browser; your agent lives in the terminal — here&apos;s the handoff schema that bridges the gap.</description><pubDate>Sun, 01 Jun 2025 00:00:00 GMT</pubDate><category>ai</category></item></channel></rss>