<?xml version="1.0" encoding="utf-8"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>TheBigHouse Media</title><link>https://thebighouselabs.eu/media/</link><description>Essays, guides, whitepapers and talks from TheBigHouse Media.</description><language>en</language><lastBuildDate>Tue, 13 Oct 2026 00:00:00 +0000</lastBuildDate><atom:link href="https://thebighouselabs.eu/media.xml" rel="self" type="application/rss+xml"/><item><title>Who stops an AI agent when a website says no?</title><link>https://thebighouselabs.eu/media/who-stops-an-ai-agent-when-the-website-says-no/</link><guid isPermaLink="true">https://thebighouselabs.eu/media/who-stops-an-ai-agent-when-the-website-says-no/</guid><pubDate>Tue, 13 Oct 2026 00:00:00 +0000</pubDate><category>Article</category><description>AI agents can keep working around a website’s refusals, errors and access limits to finish a task. Recent incidents show why safeguards must be enforced by the services and permissions around an agent, not by its instructions alone.</description></item></channel></rss>