Why don't more developers "use the platform"?

The tools you use to build your applications will increasingly favor platform-native features over custom code.

For years, web developers have been urged to "use the platform," meaning to leverage built-in browser APIs and features rather than building custom solutions. Historically, browsers lagged behind the ecosystem, forcing developers to create their own libraries, like jQuery, to fill gaps. This led to ingrained habits and a preference for familiar npm packages, even when platform equivalents existed. Documentation was also a barrier, with scattered resources for platform APIs compared to the rich documentation of third-party libraries. Furthermore, some developers find building custom solutions more engaging and educational, leading to a sense of ownership and deeper learning, akin to the "IKEA effect." This often results in reinventing the wheel, even when native, more efficient platform features are available. The article suggests this tendency to build custom solutions over using platform features is not unique to the web and can be observed in other tech stacks.

The historical lag in browser development and fragmented documentation fostered a generation of developers accustomed to building custom solutions, a habit that persists due to familiarity with ecosystems like npm. This "platform-skeptic" mindset, where developers opt for custom code despite existing platform APIs, is driven by a combination of inertia, the allure of pre-packaged libraries, and the enjoyment derived from the creation process itself. The article highlights that this phenomenon is not limited to web development but extends to any domain where developers work with platforms they may not fully understand, suggesting a deep-seated tendency to build rather than integrate.

The rise of AI coding assistants presents a dual-edged sword for this trend. Optimistically, AI could leverage its encyclopedic knowledge of platform APIs to generate more efficient and correct code, reducing the need for custom solutions. However, a pessimistic view suggests AI might exacerbate the problem by prioritizing novel code generation over existing, optimized platform features, potentially leading to an explosion of custom, less maintainable code if not guided by rigorous testing and a deep understanding of platform capabilities.