Gamma has reached more than 50 million users with a team of 30, using a smaller-team model than the usual startup scaling playbook.
2
Generalists who can learn, teach, and work across design, code, and user research can connect product decisions more effectively than narrowly specialized roles.
3
Player coaches stay close to the work while mentoring and prioritizing for the team, which helps leaders adapt quickly as AI changes the work.
Summary
Grant Lee argues that founders should apply first-principles thinking to organization design as well as product development. Gamma has passed 50 million users with a team of 30, around one-tenth the size Lee says a similar company might have used a few years earlier. He describes hiring generalists who can move between disciplines, such as Gamma's head of design, who codes prototypes, researches users, and mentors the team. Leaders at Gamma are player coaches. They manage people while continuing to do the work, which keeps them close to technical tradeoffs and changing priorities. Lee also says small companies need to invest in culture and brand early. Gamma maintains a living culture deck, runs regular company-wide meetings, and uses show-and-tell to preserve shared context. In the Q&A, he recommends patience during early AI exploration, earlier investment in experimentation infrastructure, founders doing unfamiliar jobs before hiring for them, and work trials for ambiguous roles.
Founders can apply first principles to team design
Lee says founders often use first-principles thinking to decide how to build products, but they can apply the same approach to building teams and designing organizations. The traditional model adds layers as a startup grows: a VP hires directors, directors hire their reports, and the pattern spreads across functions. Lee says Gamma has passed 50 million users with a team of 30, and that the company is roughly one-tenth the size it might have been under a few-years-ago startup scaling model. He does not claim Gamma has finished learning this approach. The team is still adapting and sharing what it sees.
Generalists connect product decisions across disciplines
Gamma's head of design was its first hire and combines visual design, coding, and deep user-experience work. That range lets him understand what engineering can actually ship, build prototypes, put them in front of users, and bring feedback back into development. As the product became more complex, his responsibilities changed rather than narrowing. Lee says strong generalists like learning and teaching. Gamma looks for people who can explain a domain clearly, persuade others, and teach a new skill during the interview process. The point is practical connection across design, engineering, and user research.
Player coaches make decisions close to fast-moving work
Lee uses an American football analogy. A player coach is on the field and can adjust to what is happening without waiting for the head coach to call every play. He applies this to AI, which is changing quickly and requires frequent reprioritization. Gamma's leadership team members are player coaches. Engineering leaders have management experience but continue to code and stay involved in daily work. That gives them context when people need mentoring, coaching, prioritization, or technical tradeoffs. Lee says the model is working for Gamma today, although the company does not yet know how it will scale.
Small teams need culture systems from the beginning
Lee says brand and culture should develop together. A company's brand reflects its culture, while culture describes its operating values. A small team has less room for a poor fit because one person's behavior can affect the whole group. Gamma has kept a living culture deck since the beginning and rewrites it often. The team uses it to describe its values, onboard employees, and compare those values with how people actually behave. Lee says this creates continuity and shared tribal knowledge, so people do not need to be repeatedly retrained or re-onboarded.
Regular shared work preserves context as Gamma grows
Gamma holds three standing company-wide meetings. At the start of the week, the team goes deeply into metrics and reviews a 'wall of work' where everyone shows what they are doing. On Wednesdays and Fridays, the company holds show-and-tell sessions. Employees use Gamma to present a small project or a shipped feature, which also gives the team a chance to use its own product. Lee says these routines help a small team retain the feeling of working together in one room around a long-term vision. He prefers preserving that experience over building a company alone.
AI speed does not remove the need for early thinking
Asked what he would change about building the team during the recent AI wave, Lee says Gamma started four years earlier and had to spend time in the idea maze, finding a real user need and deciding which problem to solve. He warns that AI can make it tempting to start building immediately. In hindsight, he wishes Gamma had spent more time considering how quickly the technology was changing before its first AI launch two years earlier. He would have thought earlier about infrastructure decisions that become hard to undo at scale. He also would have put more weight behind experimentation infrastructure, especially with a large user base.
Founders should learn unfamiliar functions before hiring for them
When asked about adding communication, legal, or other specialists without damaging the company's culture, Lee says founders and leaders should try doing the job first. He personally handled much of Gamma's marketing, sales, and customer experience work, even though he did not always do it well. That gave him a baseline understanding of the work, its nuances, and what good performance might look like. He then looks for a player coach in the role rather than creating a large reporting structure. The approach applies the same hands-on standard outside engineering.
Lee says Gamma's biggest hiring failures happened when the role was unclear and the company skipped a work trial. In a trial, a candidate does the actual job for a period ranging from two days to three months. Gamma usually defaults to three months when the person is between jobs or doing fractional work. Lee says every work trial has worked out for Gamma so far, across more than five trials. The hires that skipped trials had a high failure rate, especially when Gamma had not yet defined what it was hiring for. A trial lets both sides discover whether the role and the person fit.
"What if you had player coaches on the field that are able to actually understand how can we adapt?"07:27
Who should watch
Founders deciding whether to add managers and specialists as their startup grows will get a concrete alternative built around generalists and player coaches.
Engineering and product leaders working in fast-changing AI markets can use Lee's hiring and decision-making examples to keep leaders close to the work.
Teams that are adding their first non-technical functions will find advice on learning the job before hiring and using work trials for unclear roles.
