Grant Lee serves 50 million users with a team of 30

Traditional scaling models rely on narrow specialization. Once a company hits a certain growth trajectory, the standard playbook dictates hiring deep experts for every micro-niche. Grant Lee, co-founder of Gamma, argues that the AI era demands the opposite: the rise of the high-agency generalist. These individuals do not just inhabit a role; they connect dots across disciplines.

Lee points to his head of design as the archetype. This individual handles visual aesthetics, conducts deep UX research, and writes production-ready code. By understanding the underlying engineering constraints, a generalist avoids the friction of handing off impossible designs. They possess the empathy required to understand what can actually be built, allowing a tiny team to outpace massive organizations bogged down by cross-departmental coordination.

Leading from the trenches with player coaches

Management in the old world often becomes a cascading hierarchy of directors and VPs who primarily move information. In a lean AI startup, that overhead is a liability. Lee implements a "player coach" model across Gamma, requiring every leader to remain active in the day-to-day work.

This approach draws a direct parallel to the Quarterback in American football. Rather than waiting for a top-down mandate from a sideline coach, the player coach reads the field in real-time and adjusts the play. When the technology moves as fast as large language models do, leadership must have the technical context to make immediate trade-offs. You cannot mentor an engineer on a complex refactor if you haven't touched the codebase in six months.

Tribal knowledge over corporate hierarchy

Scaling doesn't always mean adding headcount. Instead, Gamma scales through brand and culture, which Lee views as two sides of the same coin. A small, high-density team functions more like a tribe than a corporation. This tribal structure relies on continuity; when every member shares the same core values and operational context, the need for formal processes evaporates.

To maintain this, Lee uses an evergreen "living culture deck" and holds frequent company-wide "show and tell" sessions. These rituals are not just about updates; they are about dog-fooding the product and reinforcing shared context. This transparency ensures that even as the user base grows to 50 million, the team maintains the high-speed communication of a single room.

Proving agency through work trials

Resumes are notoriously poor indicators of high agency. To find people who can navigate the "idea maze," Gamma prioritizes work trials—sometimes lasting up to three months—over traditional interviewing. Lee has found that failures in hiring almost always stem from roles where no work trial was performed. By asking candidates to drill into the fifth or sixth layer of a problem, Lee sifts for those who truly understand the "why" behind their technical decisions rather than those who simply follow orders.