Daniela Amodei's Cofounder Test: Go on Vacation Together First

Daniela Amodei has a concrete test for entrepreneurs evaluating a potential cofounder: take a vacation together first. Not a business trip or a working weekend. An actual vacation, sharing a room. "If you feel like, 'Man, all I wanna do is spend more time with you,' that's great," the Anthropic president told students at Stanford Graduate School of Business's View From The Top speaker series on May 12. "But if you're like, 'Really, I'm gonna need a vacation to recover from my vacation,' it might be the wrong choice."

The Logic Behind the Test

The advice is practical rather than sentimental. Amodei is describing a compression test. A vacation surfaces friction that ordinarily takes months of working together to reveal. You are tired, away from your routines, and cannot fall back on professional distance. If the relationship holds under those conditions, it is a reasonable signal that it will hold when a product launch slips or a key hire falls through.

She is drawing directly from experience. Anthropic was founded in 2021 by seven people, all of whom left OpenAI over disagreements about the company's direction. The founding team had unusually long prior relationships. Amodei has known her brother and co-CEO Dario her entire life. Several other cofounders had worked alongside her for fifteen years. When the group decided to start a new company focused on AI safety research, the interpersonal foundation was already tested by years of close collaboration under pressure.

A third factor she raised was consensus on purpose. Interpersonal compatibility alone is not sufficient, Amodei said. The founding team needs genuine alignment on what the company is trying to accomplish, not a shared vague ambition but a worked-through answer to the question of why this particular company needs to exist. Anthropic's cofounders left OpenAI with differing personal motivations but converged on a specific shared conviction about how frontier AI should be built and governed.

What Anthropic Looks for When Hiring

Amodei extended the theme of interpersonal fit beyond the founding team. When hiring at Anthropic, she said, the company looks for "people who are great communicators, who have excellent EQ and people skills, who are kind and compassionate and curious and want to help other people." The technical bar is given, but it is not the primary sorting criterion.

Her own background supports that framing. Amodei studied English Literature and said she would make the same choice again. As AI systems become more capable at the analytical tasks that traditionally defined the value of STEM training, the skills that humanities education sharpens, close reading, arguing from incomplete evidence, writing across audiences, become harder to automate and more differentiating. "Humanities degrees will be more important than ever," she said.

The argument is not abstract. Anthropic's safety work depends on people who can write clearly about model behavior, argue rigorously about values and alignment, and communicate across the gap between technical researchers and policy audiences. Those capacities are not the exclusive province of computer science departments. The earlier Stanford appearance covered related ground, where Amodei argued that safety and commercial success are not in tension, but this talk went further into how that philosophy shapes the people Anthropic actually hires.

"If you're like, 'Really, I'm gonna need a vacation to recover from my vacation,' it might be the wrong choice." Daniela Amodei, Stanford GSB View From The Top, May 2026

Consensus Before You Start

The talk's final theme concerned what cofounders need to agree on before they commit. The founding team needs genuine alignment on the company's purpose, not a shared aspiration but a specific, worked-through answer to the question of why this company should exist and not some adjacent version of it.

Anthropic's founding team had that. The seven cofounders left OpenAI with differing personal motivations but converged on a specific shared conviction: that building frontier AI models responsibly required a company organized around safety research rather than around commercial competition. That shared specificity, Amodei argued, is what allowed the team to navigate disagreements about tactics without fracturing the foundation that held them together.

The lesson she offered to Stanford students was not primarily about technical roadmaps, fundraising strategy, or market timing. It was that the quality of the founding team and the clarity of its shared purpose determine more about a startup's chances than almost anything else that follows. For a company whose central argument is that how you build AI matters as much as what you build, the message is consistent with everything else Anthropic does. Learn more about the company's approach at our Anthropic background page.