This study explores how Generative AI (GenAI) chatbots transform knowledge management (KM) in software development. Through qualitative interviews with developers, we examine shifts in knowledge retrieval, collaboration patterns, verification practices and skill development. Findings reveal that while GenAI tools streamline coding efficiency, they also introduce risks related to epistemic trust and erosion of deep expertise. The study proposes a post-routine model of KM, where AI automates routine tasks while amplifying the need for human oversight and critical reasoning. Practical recommendations emphasize AI governance, verification protocols, and safeguarding cognitive rigor within AI-augmented development environments.