The piece explores the current landscape of software investing, particularly in the context of the AI transition. It argues that the market is mistakenly lumping all software companies together. There are two distinct types: those that require human interaction, like Salesforce and HubSpot, and those that operate through APIs and infrastructure, such as MongoDB and Twilio. The first category is under threat as AI can automate tasks previously done by humans, leading to a decline in seat-based SaaS subscriptions. In contrast, the second category stands to benefit from increased demand, as AI agents generate far more API calls than humans ever did, creating a greater need for those underlying services. The article makes a compelling case about the services economy tied to software, particularly IT outsourcing firms like Infosys and Capgemini. These companies rely on labor arbitrage—hiring cheaper labor in countries like India and charging higher rates in the West. AI threatens this model not by making labor cheaper, but by drastically reducing the need for that labor altogether. Following the launch of AI services like Anthropic's Cowork, Indian IT services stocks saw significant drops, indicating market concern over the viability of the outsourcing model in the age of AI. The author asserts that while some firms like Accenture may attempt to pivot towards AI implementation services, traditional Indian outsourcers might struggle due to their business model relying heavily on volume and human resources. The shift in the market means that the winners will be those companies positioned to leverage AI effectively, while those stuck in outdated models will face challenges. This bifurcation in the software market underscores a crucial shift in how technology and labor intersect.