The New Hiring Filter

One of the most important hiring questions for founders is not simply whether a candidate is exceptional. It is whether their expertise addresses the company’s greatest uncertainty.

Early-stage companies are typically surrounded by a different kind of fog. For some, the uncertainty is technical. For others, it is product-market fit, distribution, trust, regulation, research, or user behavior. Some of the most effective hiring plans begin by identifying where that uncertainty sits and then concentrating talent against it.

ElevenLabs provides a useful example. In many startups, the first hires are overwhelmingly engineering-focused. ElevenLabs hired researchers early because the company’s central uncertainty was whether it could develop the proprietary models its product would depend on. Research was not a later-stage capability. It was foundational to the company’s existence.

Nevis made a different choice. Its early leverage point was product judgment, leading the company to hire a founding designer before hiring engineers. In a new category, the scarce resource was not simply the ability to build. It was the ability to determine what deserved to be built in the first place.

At first glance, these examples point in different directions: one company invested early in research, while another prioritized design. The underlying logic, however, is the same. Both organizations hired first against their most important unresolved question.

This may be one of the most important shifts occurring in company building today. The default sequence of engineering first, sales second, and support functions later is no longer universally correct. The first ten hires should not necessarily be inherited from an old startup playbook. In our view they should reflect the company’s most important unresolved question.

Victoria Weller, VP Operations at ElevenLabs, points to another version of this lesson when discussing rapid scaling. Recruiting and people infrastructure often need to arrive earlier than founders anticipate. A hiring surge can feel temporary until it becomes the company’s new operating rhythm. By the time a team needs thirty, sixty, or one hundred additional employees, the systems for finding, evaluating, onboarding, and integrating them can no longer be improvised.

AI is placing increasingly powerful tools in the hands of almost every knowledge worker. Tasks that once required specialized expertise including — writing code, creating mockups, generating content, building dashboards, analyzing data, or orchestrating workflows — can now be accelerated or partially automated. As the cost of execution falls, however, the source of value does not disappear. It simply moves.

The differentiator becomes less about producing output and more about determining which output matters. What should be built? Which customer problem deserves attention? Which workflow should be redesigned? Which product experience will create delight rather than merely functionality?

Alex Murashko, Founding Product Designer at Nevis, sees this shift clearly within design. If generating interfaces and prototypes becomes increasingly accessible, then the value of design can no longer reside solely in production. It must move toward direction, prioritization, and judgment.

Nev Flynn, an early design leader at ElevenLabs, frames it through the lens of craft. As technical capability becomes more widely distributed, the difference between good and great products often comes down to taste, clarity, and judgment. The question is not simply whether a team can ship. It is whether they can consistently ship experiences that feel obvious and inevitable to the user.

Phil Fernandez, former CEO of Marketo and longtime board advisor, offers an important caution. AI tools do not magically turn engineers into designers, nor do they eliminate the need for deep expertise. Design is not merely the production of screens. It requires empathy, sequencing, context, and an understanding of human behavior. A company that mistakes cheaper output for better decision-making may move faster while simultaneously moving further from customer needs.

The same pattern extends well beyond design. Across product, revenue operations, data, finance, customer success, and people functions, some of the highest-leverage employees combine deep expertise with the ability to operate across boundaries. They still need a functional major, but they also need a builder’s mindset that allows them to identify problems, run experiments, create leverage through AI, and translate learning across the organization.

This is not the end of specialization. Deep expertise remains essential. But expertise alone is becoming less sufficient than it once was. Increasingly, the highest-performing employees resemble scouts as much as operators — capable of navigating uncertainty, creating clarity, and helping the organization explore terrain that has not yet been mapped.

Hiring decisions rarely remain confined to recruiting. Over time, they become embedded in culture, shaping what behaviors are rewarded, what forms of contribution are respected, and ultimately how the organization learns and operates. Companies that consistently hire for pedigree create one set of norms. Companies that hire for experimentation, ownership, and learning create another.

This becomes particularly important in fast-growing organizations. Weller is clear that at ElevenLabs, new hires cannot simply absorb culture after they arrive. When headcount is growing rapidly, each individual contributes to defining and reinforcing culture from their first day. The combination she looks for is telling: ambitious, low-ego, and hands-on. Ambition creates momentum. Humility allows teams to learn. A hands-on orientation keeps leaders connected to the reality of the work itself.

In lean AI-native organizations, individual leverage is so high that these qualities matter even more. Fernandez highlights the growing risk of key-person dependency. When a small number of highly capable employees build critical agents, workflows, and operating systems, organizations can become fragile if that knowledge remains trapped inside individual minds. In a traditional company, losing one employee may create a temporary gap. In a lean AI-native company, the exposure is generally far greater.

The answer is not bureaucracy. It is intentional knowledge sharing. Documentation, shared practices, and repeatable learning loops are not administrative overhead. They are the scaffolding that allows lean organizations to remain resilient as they scale.

People compound their impact inside systems. Organizations can hire curious people and punish experimentation. They can hire builders and bury them in the process. They can hire adaptable leaders and then evaluate them using metrics designed for a different era. Eventually, the system wins.

Prabhakar is particularly clear on the risks of measuring the wrong thing. AI adoption is not the same as AI impact. If companies turn tool usage into a performance metric, employees will use the tools because they are being measured on it,  — but that does not necessarily mean the work improves. The more important question is whether AI is changing how people spend their time and where they create value. Are teams doing more meaningful work? Are they learning faster? Are they making better decisions? Are leaders modeling the behaviors they hope others will adopt?

The hiring filter matters enormously. But the environment those people enter matters just as much. All of the work described in this piece — sequencing hires against uncertainty, interviewing for adaptability, building culture deliberately — may only compound if the organization is also rowing in the same direction. The companies that get this right can do more than hire impressive people. They can build the conditions for those people to learn, exercise judgment, and create meaningful leverage, even as the work continues to change.