{"type":"video","version":"1.0","html":"<iframe src=\"https://www.loom.com/embed/3ab3271cc054487dbfab066de96a603e\" frameborder=\"0\" width=\"1910\" height=\"1432\" webkitallowfullscreen mozallowfullscreen allowfullscreen></iframe>","height":1432,"width":1910,"provider_name":"Loom","provider_url":"https://www.loom.com","thumbnail_height":1432,"thumbnail_width":1910,"thumbnail_url":"https://cdn.loom.com/sessions/thumbnails/3ab3271cc054487dbfab066de96a603e-e19b3c5b4d3dafc4.gif","duration":213.536,"title":"RustSmith ","description":"In this video, I introduce our project RustSmith, which addresses the limitations of small language models in programming, particularly in Rust. We developed a number guessing game that showcases how our system utilizes multiple agents to generate and compile code without manual prompting. I explain the roles of the master agent and the Smith agent in this process, ensuring that any errors are automatically corrected. I encourage you to explore how this project democratizes AI usage and enhances the capabilities of small models."}