The author recounts a bold decision made just a week after launching their startup, Skald: rewriting their backend from Python to Node. The motivation behind this move stems from the need for better scalability, particularly due to the high volume of network I/O operations they perform, like calls to language model APIs. While many startups are advised to focus on shipping and selling before worrying about scaling, the author found Python’s asynchronous capabilities frustrating and inadequate for their needs. They encountered significant challenges with Django’s async support, which is still incomplete, leading to messy code and inefficient performance. Despite a fondness for Django, the author notes that writing effective async Python code requires deep knowledge of its foundations, which isn’t practical for an early-stage startup. They explored alternatives like FastAPI but ultimately chose to migrate to Node, aiming for a more cohesive tech stack. With Express and MikroORM, they immediately observed a threefold increase in throughput, even with mostly sequential code. While the transition meant losing Django's user-friendly ORM and some built-in utilities, the author is optimistic about the performance gains and plans for concurrent processing in the future. The article highlights the complexities of working with Python’s async model compared to the more native support found in JavaScript. The author’s experience underscores the importance of selecting the right tools for specific use cases, especially in fast-paced startup environments where scaling efficiently is critical.