<?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/38b24ae89f514650be4223a9dcb0de1d&quot; frameborder=&quot;0&quot; width=&quot;1996&quot; height=&quot;1497&quot; webkitallowfullscreen mozallowfullscreen allowfullscreen&gt;&lt;/iframe&gt;</html><height>1497</height><width>1996</width><provider_name>Loom</provider_name><provider_url>https://www.loom.com</provider_url><thumbnail_height>1497</thumbnail_height><thumbnail_width>1996</thumbnail_width><thumbnail_url>https://cdn.loom.com/sessions/thumbnails/38b24ae89f514650be4223a9dcb0de1d-fc7c03cb06c90403.gif</thumbnail_url><duration>209.31</duration><title>4o vs o1 in Highest Order Multi-Step Reasoning Problems</title><description>In this video, I dive into the complexities of creating challenging questions for ChatGPT and how it responds to them. I highlight the importance of language, presentation, and actionable insights while also showcasing a mermaid diagram that illustrates my points. I encourage you to review the embedded task and provide your thoughts on the reasoning process I discuss. Your feedback will be invaluable as we refine our approach.</description></oembed>