Why experts writing AI evals is creating the fastest-growing companies in history | Brendan Foody
Mercor CEO Brendan Foody explains how expert-written AI evals became the fastest revenue ramp in history — $1M to $500M in 17 months.
- Mercor grew from $1M to $400M+ ARR in 16 months, called the fastest revenue ascent in history, now tracking toward $500M.
- Mercor works with 6 of the Magnificent 7 and all top 5 AI labs; median expert pay is $95/hour, up to $500/hour for deep specialists.
- The primary bottleneck to model improvement is evals — without a way to measure success, labs cannot run RL hill-climbing at scale.
- Labs need tens of thousands of active expert contributors at any time; top 10% of contributors drive the majority of model improvement.
- Mercor hired the entire Harvard Lampoon comedy writing staff to help make Claude and other models funnier.
- Foody does not believe AGI or superintelligence arrives within 3 years; post-training data quality, not more pre-training compute, is the path forward.
- Jobs with elastic demand — software engineering, product management, consulting — will grow as AI multiplies productivity; accounting-type roles face compression.
- Mercor has never had a customer churn; net revenue retention exceeds 1,600% with zero dedicated sales or marketing staff for its first 18 months.
2025-09-18 · Watch on YouTube