{"type":"video","version":"1.0","html":"<iframe src=\"https://www.loom.com/embed/b28dd8b897a749b3a96dda8cf29cc33f\" frameborder=\"0\" width=\"1152\" height=\"864\" webkitallowfullscreen mozallowfullscreen allowfullscreen></iframe>","height":864,"width":1152,"provider_name":"Loom","provider_url":"https://www.loom.com","thumbnail_height":864,"thumbnail_width":1152,"thumbnail_url":"https://cdn.loom.com/sessions/thumbnails/b28dd8b897a749b3a96dda8cf29cc33f-6da23f56582526ea.gif","duration":306.993,"title":"Essential Challenges in Software Engineering 🤖 - Part II (Spatial Reasoning Spec)","description":"In this video, I explore the capabilities of large language models (LLMs) in handling arithmetic versus spatial reasoning tasks. I found that while LLMs excelled at arithmetic through visual activities and prompting, they struggled significantly with spatial tasks, even when provided with detailed examples and external libraries. My attempts to guide the model in creating geometric patterns revealed its limitations in understanding spatial relationships.\n\nPart I - https://www.loom.com/share/0c9639885c244e7889bb11843aa948d6 \nPart II - https://www.loom.com/share/b28dd8b897a749b3a96dda8cf29cc33f\nPart III - https://www.loom.com/share/3abd569af3b04335b8ad144b772e73e0"}