From: aidotengineer
The landscape of building successful companies has fundamentally shifted, moving away from large, bloated teams and endless hiring rounds towards the era of tiny teams [02:54:48], [02:51:00]. This change is largely driven by the advent of AI tooling, enabling smaller groups of highly capable individuals to achieve unprecedented success and even profitability much earlier in their lifetime [02:52:23], [02:55:14]. Startups are generating millions in ARR with teams smaller than many companies’ engineering departments [02:54:42].
The Rise of the Generalist
A core philosophy in this new era is the emphasis on generalists – individuals capable of performing multiple roles and understanding various aspects of the business [02:50:01].
Why Generalists are Key
- More Context Per Head: A small number of people with more context per head means that individuals have more agency and can build things without needing extensive permission or navigating a complex chain of command [00:07:29]. This empowers them to make immediate impact [00:07:50].
- Efficiency and Speed: Smaller, more aligned teams are lean and fast [00:06:39]. They can move faster, avoid constant meetings, and allow engineers deep focus time to build [00:44:52]. This is crucial for startups seeking product-market fit, as it allows for more “shots on goal” by maintaining a lower burn rate [00:07:58], [00:08:36].
- Seamless Context & Fast Feedback Loops: In a traditional large company, context is lost during handoffs between specialized teams, leading to inefficiency and slow feedback loops [02:48:52]. Generalists, however, can work across the stack, ensuring seamless context and very fast feedback cycles [02:49:19].
- Adaptability: Generalists are crucial for adapting to rapidly changing environments, such as the fast-moving AI landscape [02:30:16]. This allows for quick reprioritization and adjustments without top-down mandates [02:30:27].
Characteristics of a Generalist
- Multi-disciplinary Skills: Examples include product engineers who are full-stack developers, great product thinkers, and good at computer network fundamentals [02:58:58]; marketers who can code; and designers who can build [02:59:05].
- Connecting the Dots: A strong generalist, like a head of design who is visual, knows how to code, and can go deep on UX, is empowered to “connect all the dots” and empathize with engineering counterparts to build shippable prototypes [02:27:37].
- Continuous Learners and Teachers: They possess a willingness to learn and adapt, picking up new skills in a fast-paced environment. Crucially, they can also teach others, articulate complex ideas, and convey understanding to persuade others [02:28:46].
- High Agency and Ownership: Generalists are people who can look at a problem and figure out how to solve it, iterate with customers, and prioritize ruthlessly [02:50:03]. They prioritize doing the simplest possible thing over over-complication [02:50:24].
Hiring Generalists
Hiring for generalists requires a specific approach:
- Be Super Picky: Every person on a small team must be absolutely exceptional [00:42:20]. If it’s not a “no-brainer,” do not hire them [00:42:09]. This often means conducting hundreds of interviews and work trials [00:42:15].
- “Product-Led Hiring”: This strategy involves hiring customers who already love the product and have deep insight into its use [00:42:39]. They already understand the product and are inspired to join [00:43:24].
- Work Trials: Bringing candidates into a work trial for several days, paying them as contractors, allows the team to assess if it’s the right fit [00:44:02]. For Data Lab, paid projects (e.g., 10 hours for $1,000) are used to assess fit [02:59:14]. Gamma often defaults to three-month work trials [02:40:08].
- Screening for High Agency: Ask candidates about their most challenging projects, focusing on how they understood and solved the problem, not just the solution. High-agency individuals delve into the underlying layers of the problem and discuss various attempts at solutions [02:38:16].
- In-person Work: For small teams that need to move fast, working in person can be more effective than remote setups, as it minimizes the need for extensive processes and facilitates rapid collaboration and feedback loops [02:50:47].
- Patience: The best hires come when you take the time to find the best person, even if there isn’t an immediate specific role. Rushed hires often result in poor fits [02:54:49].
- Compensation: Pay top-of-market salaries to attract exceptional talent, as hiring fewer, highly compensated individuals can lead to greater productivity [02:54:09].
Managing Generalist Teams
The Player-Coach Model
For the core leadership team, adopting a “player-coach” model is highly effective [02:30:36]. These individuals have management experience but also actively participate in day-to-day work, remaining close to the product and operations [02:30:46]. This allows them to provide nuanced mentorship and make informed technical tradeoffs and rapid adjustments [02:31:06].
Cultivating an Empowered Culture
- Empowerment and Autonomy: Let great people build [00:45:25]. If you hire exceptional individuals, give them the space to innovate and execute without excessive oversight or meetings [00:45:32].
- Shared Context and Transparency: Regular, transparent company-wide meetings are vital for maintaining shared context in a small team [02:33:16]. This can include discussing metrics, showcasing ongoing work (“wall of work”), and company-wide show-and-tell sessions [02:33:24].
- Strong Shared Values and Low Ego: A strong team culture is built on shared core values, low ego, and high trust [00:09:26]. Everyone should be obsessed with user success and possess grit and resilience to navigate startup challenges [00:09:33].
- Intentional Fun: To counteract the intensity of fast-paced building, incorporate fun activities and team retreats [00:48:23], [00:48:53]. This helps maintain morale and prevents burnout [00:48:21].
- Automation of Internal Processes: Leverage automation tools, potentially even your own product, to handle repetitive internal tasks that would otherwise require dedicated roles or significant time [00:47:18]. This is part of the Innovation in Organizational Design for Startups enabled by new tools [02:24:52].
The Power of Tiny Teams in Crisis
During a period of significant growth for Bolt, Stack Blitz had a team of less than 20 people managing 30-40,000 active customers [00:06:05]. The CEO and Chief of Staff personally handled support tickets [00:06:24]. This was possible due to incredible camaraderie, extreme alignment, and the team’s inherently lean and fast operating style, which they had maintained for seven years [00:06:29].
Practical Examples and Benefits
- Stack Blitz (Bolt.net): Started with less than 20 people and scaled incredibly fast, doubling their ARR, when they were on the brink of shutting down [00:04:27]. They focused on maintaining a lean, fast, and aligned team with high camaraderie [00:06:29].
- Alie: A team of four scaled a portfolio of virally successful products, including Quiz Quizard and Unstuck AI, to $6 million in ARR profitably [02:56:07], [02:56:11]. They leverage “super tools” (e.g., Launch Darkly for traffic balancing, infrastructure changes, and UI experiments) and an organizational structure of “harvesters” (product engineers owning products end-to-end) and “cultivators” (AI software engineers building the company’s agentic operating system) [03:22:55], [03:32:19].
- Gum Loop: Raised a Series A as a team of two and are now nine people, scaling automation across large companies like Instacart, Web Flow, and Shopify [00:37:30]. Their Product-Led Growth (PLG) model eliminates the need for large sales teams [00:40:40]. They automate nearly every part of their business using their own product, freeing up time for building [00:47:24].
- Gamma: Reached over 50 million users with a team of 30, which is roughly one-tenth the size of what a similar company might have been a few years ago [02:25:53]. They prioritize hiring generalists who are continuous learners and teachers, and maintain a “player-coach” model for leadership [02:28:46].
- Data Lab: Achieved 40,000 GitHub stars, seven-figure ARR, and trained state-of-the-art models with a team of three (now four) [02:42:51]. They emphasize that “more people does not equal more productivity” [02:49:50]. Their two-person team handled an entire model training process from customer discussions to production integration, which would typically involve many teams in a larger company [02:48:14]. They advocate for reusing components aggressively, keeping technology simple (e.g., server-rendered HTML instead of complex frameworks), and clean, modular code [02:51:07].
Scaling Productivity, Not Headcount
The goal is to scale productivity, not headcount [02:55:14]. This can be achieved by:
- Raising salary bands to hire more experienced people into the same role [02:55:18].
- Investing in compute resources to multiply researcher/engineer productivity [02:55:25].
- Investing in AI tools that multiply productivity and abstract away complex “edges” of the business [02:55:31].
The ability to maintain a tiny team while achieving significant scale is a deliberate choice [02:53:10]. It requires a commitment to a high cultural bar, prioritizing trust and focusing on meaningful work [02:53:54]. Ultimately, the aim is to preserve the “magic” of building with a small, highly effective group for as long as possible [02:34:07].