“I would much rather people just try and make a decision, and then it’s wrong, and a week later they adjust and change, than they spend like three months not making a decision.” This radical bias for action, articulated by Harvey Co-Founder and CEO Winston Weinberg, encapsulates the necessary operating philosophy for startups navigating the chaotic, high-stakes environment of foundational AI. In a recent interview on the Long Strange Trip podcast with HubSpot Co-Founder and Sequoia Partner Brian Halligan, Weinberg pulled back the curtain on what it truly takes to scale an application layer company from zero to a $190 million run rate in under three years—a journey defined less by stable growth and more by perpetual organizational crisis management.
Weinberg spoke with Halligan about the intense internal pressure that accompanies exponential growth. While Harvey, which provides generative AI tools for elite legal firms, appears externally to be a runaway success, the internal reality is one of constant breakage. Halligan noted the company is "absolutely ripping," having quadrupled its revenue run rate and doubled its headcount to 500 employees over the last year. Yet, Weinberg candidly admitted that this success is deceptive: the foundational "machinery is still not there." The core challenge facing leaders in this epoch is that the market and the technology are shifting so rapidly that any organizational structure or process built today will inevitably fail within months. This leads to Weinberg's central, almost exhausting mandate: leaders must "reinvent yourself as a founder every like four months," otherwise "you just break."
This sentiment speaks directly to the core insight of decision velocity. In the current market climate, the speed of iteration dramatically outweighs the pursuit of perfection. Weinberg emphasizes that the biggest threat to growth is not making the wrong choice, but failing to choose at all. He explicitly states that his single most important hiring criterion is a "bias for action." In a complex, fast-moving environment, spending months debating a critical structural or product decision ensures that by the time consensus is reached, the market has already moved, rendering the decision irrelevant. The only way to survive is to institutionalize rapid decision-making, even if it carries a high risk of immediate error, allowing for quick adjustments rather than prolonged paralysis.
The chaos of hypergrowth manifests particularly acutely in organizational structure and delegation. As a founder, Weinberg confessed that he fell into the trap of retaining tasks that were beneath his pay grade and leverage potential—at one point, spending six hours trying to book a complex flight himself. This failure to delegate effectively illustrates a critical choke point for scaling founders: the unwillingness or inability to release control over functions that are performing well, even if they could perform exponentially better with specialized leadership. For Weinberg, the solution involves intentionally hiring leaders who can build leverage quickly, rather than simply expanding headcount. His primary interview test for executive candidates is demanding they draw out their future organization charts for the next three, six, and twelve months. The inability to articulate this future structure, and the steps required to fill those roles, indicates a lack of strategic foresight necessary for managing extreme scale.
Weinberg highlighted the crucial difference between merely hiring leaders and hiring leaders who can effectively manage cross-functional initiatives. His promotion of Katie Burke to COO formalized a role that manages the "glue" of the organization, freeing the CEO to focus on individual high-leverage tasks. He explains that if a leader is not actively creating leverage for themselves, especially in roles that require coordinating across engineering, sales, and product, they will fail to deliver the core competence required of their function. This focus on clear ownership is paramount, especially in product development. Weinberg insists that every critical initiative must have a single directly responsible individual (DRI), stating that almost every problem he encounters has "like six DRIs on it," leading to diffused accountability. By ruthlessly limiting key priorities (P0s) and assigning singular ownership, the organization avoids the trap of hedging bets and achieves clarity.
Harvey’s go-to-market strategy further reflects this philosophy of tackling the hardest challenges first. While conventional wisdom suggests starting with mid-tier firms that have shorter sales cycles and lower compliance hurdles, Harvey deliberately targeted the most complex, global law firms and large corporate enterprises. Weinberg noted that selling to a major bank and getting through its brutal security and implementation reviews grants instant credibility. If you can satisfy the stringent demands of a Fortune 100 corporate legal department, "you can sell everybody else." This counter-intuitive approach ensures the core product is hardened and validated against the most rigorous use cases, providing a robust platform for broader market expansion. The success of this strategy is evident in the fact that 42% of Harvey's revenue now comes from Fortune 100 companies, a powerful testament to the value of prioritizing difficult, high-leverage customers. Ultimately, Weinberg’s experience confirms that surviving the AI hypergrowth cycle demands leaders who are not only comfortable with chaos but who actively dismantle and rebuild their organizations every few months, prioritizing rapid, decisive action over static perfection.
