TL;DR

Mercor, a Silicon Valley startup valued at $10bn, pays approximately $2m daily to 30,000 experts who train AI to perform knowledge work in their own fields. Roles span consulting to radiography, with average hourly wages exceeding $95—but workers face uncomfortable questions about training their potential replacements.

A New Category of Work

Founded by three school friends, Mercor recruits teams of professionals to guide and assess AI output, teaching models how to approach problems in fields from journalism to real estate. CEO Brendan Foody—who at 22 became one of San Francisco’s youngest billionaires—describes it as building a new “category of work.”

The work differs markedly from earlier AI training, which relied heavily on low-paid workers in the global south labelling basic data. Mercor’s contractors are highly qualified: an Oxford Economics study found 41% hold master’s or doctoral degrees.

“It’s training the LLM to do the job,” says one business graduate working on consulting projects. In a tough labour market, she finds the work interesting but acknowledges displacement concerns: “Working with this kind of model we now have the sense that it could be scary.”

The Uncomfortable Bargain

Labour economists have flagged the precarious nature of this work. Zoe Cullen at Harvard Business School suggests AI trainers could retain revenue stakes in models that use their expertise: “If what you’re teaching the model to do is your core expertise, by definition you’re reducing your labour power.”

Foody acknowledges “there is definitely going to be some job displacement” but envisions a future where “everyone is managing dozens or hundreds of agents” and humans “do things once, creating a rubric for AI to do the same task many times over.”

Looking Forward

The model raises fundamental questions about knowledge work. Anton Korinek at the University of Virginia notes training still follows a “human master, AI apprentice model”—but warns: “the teacher always gets replaced by her student.” Whether that replacement is imminent or distant remains uncertain, but Mercor’s $10bn valuation suggests investors are betting on the former.