The 100-person AI lab that became Anthropic and Google's secret weapon | Edwin Chen (Surge AI)

What It Covers

Edwin Chen built Surge AI into the fastest company to hit $1 billion revenue in four years with under 100 people, completely bootstrapped, by providing high-quality AI training data.

Key Questions Answered

• •Elite team scaling: Fire 90% of people to move faster - small elite teams of 60-70 people can generate $1 billion revenue by eliminating distractions and focusing top performers on core work.
• •Data quality framework: Quality requires thousands of signals tracking worker expertise, keyboard strokes, review scores, and model performance improvements - not just checking boxes or throwing bodies at problems.
• •Model differentiation strategy: AI models will become increasingly differentiated by company values and objective functions rather than commoditized - choose models optimizing for productivity over endless engagement and iteration cycles.
• •Benchmark gaming problem: Academic benchmarks contain wrong answers and encourage hill-climbing on objective metrics rather than real-world performance - use human expert evaluations across diverse conversational topics instead.
• •Reinforcement learning environments: Build simulation worlds with Gmail, Slack, code bases where models learn through trial and reward over long time horizons - mimicking how humans learn through practice.