How AI labs use Mercor to get the data companies won’t share

At a recent TechCrunch Disrupt event, Mercor's CEO Brendan Foody unveiled a groundbreaking strategy that AI labs are adopting to access industry knowledge without the burdensome costs of formal data contracts. By recruiting former senior employees from investment banks, consulting firms, and legal practices, these labs are gaining invaluable insights to enhance automation in their respective fields. Mercor positions itself as a critical link between these skilled professionals and AI developers. As Foody highlighted, major players like OpenAI, Anthropic, and Meta rely on Mercor to tap into the expertise of individuals who are familiar with specific workflows. He noted that companies such as Goldman Sachs may be hesitant to share data that could lead to the automation of their operations, prompting the need for contractors who can help train AI models without compromising proprietary information. Foody, a 22-year-old entrepreneur, revealed that Mercor compensates these industry experts up to $200 per hour to contribute to AI training by filling out forms and drafting reports. The startup has rapidly expanded, boasting tens of thousands of contractors and a daily payout exceeding $1.5 million. Despite these costs, Mercor remains profitable, thanks to the willingness of AI labs to invest in the critical knowledge that drives their innovations. Since its launch less than three years ago, Mercor has skyrocketed its annual recurring revenue to about $500 million and recently achieved a valuation of $10 billion. This meteoric rise has raised concerns among traditional firms that may fear losing their competitive edge as industry knowledge becomes accessible through the marketplace. While Foody admits his business could highlight market inefficiencies, he refrains from labeling it as exploiting a loophole. He believes that some companies are beginning to recognize the inevitable shift in the workforce landscape, akin to the transformation ushered in by ride-sharing platforms like Uber. However, the ethical implications of this new gig economy are complex. Foody expressed a commitment to preventing corporate espionage, noting that many of Mercor's contractors are still employed in their previous roles and are instructed to avoid sharing confidential documents. Yet, he acknowledges the challenges of fully enforcing these guidelines. Mercor has distinguished itself as a pioneer in recruiting high-skilled workers for AI training in the U.S., contrasting with earlier data vendors that primarily relied on low-cost labor overseas for basic tasks. As the demand for expert insights grows, competitors like Surge and Scale AI are quickly adapting their strategies to include this approach. Despite recent challenges faced by rivals, Mercor's valuation has soared fivefold in the past year, although it still lags behind larger competitors. Currently, a significant portion of Mercor's revenue is generated from a handful of AI labs, but Foody envisions partnerships expanding into sectors such as law, finance, and healthcare. He confidently predicts that advancements in AI will eventually surpass the capabilities of the top consulting and legal firms, heralding a transformative era for the economy that promises to benefit all.