Sourcery
Sourcery

Nov 10, 2026 · 54 min

Mercor turns human expertise into AI infrastructure

Mercor Raises $750M at $20B Valuation | CEO Brendan Foody

The conversation examines whether data, evaluations, and proprietary enterprise intelligence will matter more than model access alone as AI adoption accelerates.

3 key takeaways
  1. 1Mercor combines expert talent, software, data, and evaluations to build reinforcement-learning environments for AI labs and enterprises.
  2. 2Foody argues that AI productivity will expand demand for human training wherever models remain below expert performance.
  3. 3Enterprises may build proprietary models and agent systems, beginning with reliable evaluations before investing heavily in post-training.

Don't miss

Foody argues that enterprises should establish reliable evaluations and reinforcement-learning tasks before attempting to train models around proprietary knowledge.

The brief

Brendan Foody describes Mercor’s evolution from AI recruiting into a platform combining expert talent, data, evaluations, and software for frontier AI labs and enterprises.

Mercor’s core product is not a labor marketplace, Foody argues, but a complete reinforcement-learning environment: apps, tasks, prompts, worlds, and verifiers.

The company handles large expert payouts through Stripe infrastructure while using referrals, AI interviews, and work analysis to identify exceptional talent across uneven markets.

Foody’s central demand thesis is counterintuitive: as copilots make experts more productive, Jevons’ paradox expands the need for human-generated training data.

The conversation widens to durable AI moats, acquisitions, and enterprise sovereignty, with Foody predicting companies will build proprietary intelligence rather than surrendering all knowledge to frontier labs.

The standout idea is his advice to enterprises: build dependable evaluations and reinforcement-learning tasks first, then scale training around the capabilities that matter.

Weekly email

State of the Pod, every Monday

The week's most-discussed names and stories from podcasts, with the moments behind them. Free, five minutes, unsubscribe in one click.

Listen to the full episode and explore every guest, topic, and moment on PodLume.