<?xml version="1.0" encoding="UTF-8"?><oembed><type>video</type><version>1.0</version><html>&lt;iframe src=&quot;https://www.loom.com/embed/fb6b6732166f410aa6872ba9c32ca274&quot; frameborder=&quot;0&quot; width=&quot;1280&quot; height=&quot;960&quot; webkitallowfullscreen mozallowfullscreen allowfullscreen&gt;&lt;/iframe&gt;</html><height>960</height><width>1280</width><provider_name>Loom</provider_name><provider_url>https://www.loom.com</provider_url><thumbnail_height>960</thumbnail_height><thumbnail_width>1280</thumbnail_width><thumbnail_url>https://cdn.loom.com/sessions/thumbnails/fb6b6732166f410aa6872ba9c32ca274-04e0054e8dcc20c7.gif</thumbnail_url><duration>246.471</duration><title>Autonomous Vibe Finder Architecture with Safety Loops</title><description>Hi everyone, I am Raymond, and today I am excited to show you a gentec vibe finder I built as an autonomous creator. It uses a plan, act, reflect, refine loop where it first plans from fuzzy natural language into a structured profile, selects a scoring strategy, then reflects and retries until the energy matches, with a reliability layer that blocks contradictory songs and a RAG knowledge base for expert insights. In my live demo, it handled high energy pump, relaxed but intense rock with safety conflicts, and tired but focusing. The system currently has a 100% rate on its safety and reliability criteria. I did not ask you to do anything during the demo.</description></oembed>