Why experts writing AI evals is creating the fastest-growing companies in history | Brendan Foody

The era of AI evaluations has become the primary bottleneck and business opportunity for foundational model labs. As models become the product, evaluations function as the essential product requirement document, measuring success and guiding reinforcement learning. Mercor, a high-growth labor marketplace, facilitates this by connecting AI labs with top-tier professionals—such as software engineers, lawyers, and doctors—who provide the high-caliber feedback necessary to refine model capabilities. This shift moves beyond traditional crowdsourced data toward expert-driven reinforcement learning from AI feedback. Rather than signaling widespread job displacement, this evolution creates a new category of work centered on building, testing, and training AI systems. Success in this landscape requires leveraging AI to increase productivity, focusing on elastic industries where demand scales with efficiency, and maintaining high standards for expert-driven model alignment.