Why top AI talent is leaving Meta: Researchers head for the door, using Mark Zuckerberg’s own words to justify exit

deutsche mark Zuckerberg’s ride to impel Meta to the head of unreal intelligence (AI) is reportedly encountering early difficulties. Several leading researchers have left the company’s newly formed Meta Superintelligence Labs (MSL) only months after its announcement.

Lavish recruitment drive

According to Futurism, MSL intended to signal Meta’s entry into the highest tier of AI research. Zuckerberg himself spearheaded the hiring effort, reportedly offering unprecedented salaries and contracts worth hundreds of millions of dollars to lure experts away from rivals such as OpenAI, DeepMind and Apple.

Despite these lavish offers, talent has begun to slip away. According to Wired, at least three researchers have resigned in recent weeks. 

Avi Verma and Ethan Knight, both formerly at OpenAI, have chosen to go back after short stints at Meta. Meanwhile, Rishabh Agarwal, an Indian researcher recruited from Google DeepMind on a reported $1 million package, confirmed that August would mark his final month at the lab.

Zuckerberg’s own words used against him

In a statement on X, Rishabh Agarwal acknowledged the “talent and compute density” within MSL but said he was ready for “a different kind of risk”. He signed off with a direct quote from Zuckerberg: “In a world that’s changing so fast, the biggest risk you can take is not taking any risk.” The remark, first made by the Meta founder in 2011, has since been widely interpreted as researchers turning his own words against him.

But why are researchers leaving Meta AI Lab?

The departures come amid reports of turbulence inside Meta’s AI operations. Frequent reorganisations, shifting strategies and close oversight from senior leadership have contributed to unease. The recent decision to split the AI workforce into four separate teams has further fuelled uncertainty.

The situation highlights a broader challenge facing companies attempting to recruit frontier AI talent. As DeepMind co-founder Demis Hassabis has observed, many top researchers are motivated less by salaries than by the opportunity to shape the safe development of artificial general intelligence, reported Economic Times. 

Benjamin Mann, co-founder of Anthropic, put it more starkly: “My best case at Anthropic is we affect the future of humanity. My best case at Meta is we make money.”

OpenAI gains from Meta’s losses

For OpenAI, the timing is advantageous. Alongside the return of Verma and Knight, the company has welcomed Chaya Nayak, a veteran of nearly a decade at Meta, who will lead special initiatives. Her exit represents another significant blow to Zuckerberg’s plans.

Despite appointing prominent figures such as former Scale AI chief Alexandr Wang and ex-GitHub CEO Nat Friedman to senior roles, Meta’s superintelligence effort risks being undermined by instability and cultural frictions.

Zuckerberg has long styled himself as a risk-taker willing to bet big on the future. For now, however, it is his recruits who are invoking that philosophy as they head for the door. Whether Meta can stem further departures will be critical to its standing in the global race to build advanced AI.

---

---