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

Surge AI's early story begins with a deliberate rejection of the conventional Silicon Valley playbook. Edwin Chen describes building a small, elite, bootstrapped company rather than a large organization dependent on venture funding and publicity. The result, as Lenny frames it, was more than a billion dollars in revenue in under four years with fewer than one hundred people. Chen presents this as the combination of a lean organization and AI-enabled efficiency, while also warning that the ratio should not be treated as a universal benchmark.

Chen says avoiding the Silicon Valley publicity and fundraising machine made growth harder, because Surge had to win through product quality and researcher word of mouth rather than investor announcements. That constraint also selected for customers who understood data and cared about quality. In his account, close mission alignment with early customers produced feedback that helped the company improve, while the absence of a large organization allowed a small group to focus on the work it considered important.

The company defines its work as teaching AI models what is good and bad through human data, training methods, rubrics, verifiers, and reinforcement-learning environments. Chen rejects the idea that quality can be purchased simply by adding more workers. His example is an eight-line poem about the moon: checking line count and the presence of the word moon is easy, but judging originality, imagery, emotional force, and insight requires a richer and more subjective standard. Surge therefore gathers signals about expertise and performance across workers, tasks, and projects. The account makes clear why this work is difficult to reduce to a single score, and why the company treats measurement technology as part of the product itself.