In this reflective essay, the author explores the complexities of recent advancements in AI, particularly in programming. They ponder the influence of both large language model (LLM) developments and traditional, pre-built systems on AI decision-making processes, questioning how much of AI behavior is generative versus guided by manual instructions from software engineers. The author illustrates their evolving view on AI, inspired by a video from Hank Green, while also acknowledging the obscured details that AI companies maintain about their models.