Evan Armstrong argues that the recent downturn in the software market, which saw $300 billion in market cap erased from various companies, stems from fears that AI-generated code could diminish the value of traditional software. While companies like Salesforce and newer startups like Figma were hit hard, Armstrong believes the narrative oversimplifies a complex situation. Rather than making software companies less valuable across the board, AI is reshaping the landscape by redistributing value within the software stack. He identifies a shift toward three distinct layers of software: commoditized layers that handle basic functions and a new "context layer" that holds significant potential. The context layer consists of the institutional knowledge that guides AI agents in executing tasks effectively. It encompasses the nuances of how organizations operate, which can’t be captured by traditional documentation like markdown files. This layer is crucial because it transforms raw data and workflows into actionable intelligence, allowing AI agents to make informed decisions. Armstrong emphasizes that while software companies face a valuation problem, particularly those relying on traditional SaaS models, the rise of AI creates opportunities for those that can harness this context layer. Companies that succeed in integrating and managing this knowledge will likely emerge as leaders in the evolving software ecosystem. The context layer not only holds the key to maximizing the efficiency of AI agents but also represents a new frontier for business value in software.