20VC: Mercor CEO on Why Application Layer Companies Have No Defensibility, The Model is the Product | Token Spend Will Exceed Headcount Spend in 5 Years | The True Cost of Hiring AI Researchers in the Valley Today with Brendan Foody

AI-driven automation is fundamentally reshaping the global economy, shifting the focus from traditional software layers to the underlying frontier models. As reasoning capabilities improve, enterprises are increasingly adopting agentic workflows that manage complex, end-to-end tasks, rendering traditional software moats less defensible. This transition is driving a massive reallocation of capital, where companies now spend more on AI tokens than on human headcount. While frontier models like those from OpenAI and Anthropic are poised for multi-trillion-dollar valuations, the API layer faces rapid commoditization due to low switching costs. Success in this new paradigm requires a "forward-deployed" approach, where AI is deeply integrated into specific enterprise workflows to capture tacit knowledge. Ultimately, the rapid scaling of these agentic systems suggests that AI will not only displace existing roles but also create unprecedented demand for new job categories focused on training and managing autonomous agents.