{"type":"rich","version":"1.0","provider_name":"Transistor","provider_url":"https://transistor.fm","author_name":"Ortec","title":"How far can you trust an agent? - Cynthia","html":"<iframe width=\"100%\" height=\"180\" frameborder=\"no\" scrolling=\"no\" seamless src=\"https://share.transistor.fm/e/a08e0caa\"></iframe>","width":"100%","height":180,"duration":461,"description":"\"Please fix the review comments.\" When Cynthia received this pull request for the third time, she realized something was off, not with the code, but with the way she was working. \n \nCynthia is Senior AI Engineer at ORTEC and leads the team behind Art, ORTEC's AI assistant. A recent move to GitHub pushed her team's use of coding agents further than they expected, and that one frustrating pull request turned into a much bigger question about how far you can actually trust an agent to work on its own. \n \nIn this episode, Cynthia shares what her team built to answer that question, and how it changed the way she thinks about her own job. ","thumbnail_url":null,"thumbnail_width":300,"thumbnail_height":300}