Mercor CEO Highlights "Human Paradox" as AI's Future Relies on Human-Built Environments

Brendan Foody, CEO of Mercor, recently published an essay titled "The Economy will Become an RL Environment Machine," asserting that "the future of AI is human," a statement that has garnered significant attention, including from venture capitalist Spencer Peterson. The article, published on the Mercor Blog on September 15, 2025, posits a "great paradox" where the advancement of artificial intelligence, despite fears of job displacement, will intrinsically rely on human involvement in its development and evaluation.

Foody's essay argues that while technological revolutions historically spark concerns over job loss, they ultimately create new categories of work. He contends that the AI revolution is no different, fostering a new class of workers dedicated to training and guiding AI agents. "This is the great paradox: the future of AI is human," Peterson quoted from the essay in a recent social media post, emphasizing the critical role humans will play.

The core of Foody's argument centers on the concept of Reinforcement Learning (RL) environments, which he believes will become the foundation of the future economy. He explains that as RL models become increasingly effective, the challenge shifts to building robust, real-world evaluation environments. These environments, essential for assessing AI performance in complex tasks, will necessitate human labor to construct, verify, and refine them.

Mercor, a company focused on connecting graduate-level experts with AI labs, is positioned to address this emerging need for specialized human talent. Foody highlights that the market for humans teaching models will persist as long as there are tasks agents cannot perform, requiring human expertise to create evaluations and train agents for longer-horizon and collaborative tasks. The essay concludes that the AI revolution will create an industry driven by people who will shape AI’s judgment and ensure its outputs meet human standards.