Senior software developers are now finding themselves in the role of ‘AI babysitters’ as they grapple with the unpredictable and often problematic nature of AI-generated code. Carla Rover, a veteran programmer with 15 years of experience, described her experience with AI coding, which led to a project restart after discovering significant flaws in the generated code. She emphasized the need for manual verification, noting that AI tools can introduce errors ranging from hallucinated package names to security vulnerabilities. A recent report by Fastly revealed that 95% of developers surveyed spend additional time fixing AI-generated code, with senior developers disproportionately affected. This trend highlights the growing challenges in integrating AI into the development process and the necessity for human oversight.