The day I started believing (davidpoblador.com)

A software engineer shares a transformative experience with AI coding assistants, specifically Claude Opus 4.5, which reshaped their understanding of AI's capabilities. Initially viewing these tools as limited and primarily useful for generating boilerplate code, the engineer's perspective shifted dramatically when they realized that the latest models could manage complex, multi-hour tasks coherently, thus expanding their practical utility beyond simple code generation. This shift was particularly evident in a challenging home networking project where the AI not only provided guidance but executed intricate tasks autonomously, significantly reducing the time and effort required. This development is significant for the AI/ML community as it challenges previously held assumptions about the limitations of AI in coding tasks and suggests that the role of experienced engineers will evolve rather than diminish. The author notes a doubling of the task complexity that AI can handle every seven months, indicating rapid advancements in the technology. Rather than relegating senior engineers to cleanup roles, AI tools are enhancing their reach, empowering them to tackle projects previously deemed too daunting. This evolution triggers a necessary reevaluation of team structures, workflows, and the nature of expertise within software development.