AI researchers lured with high salaries are leaving Meta, quoting Mark Zuckerberg’s own advice on their way out

When Mark Zuckerberg announced Meta’s Superintelligence Labs (MSL) earlier this year, it was framed as a moonshot project meant to catapult the company to the forefront of the artificial intelligence race. The Facebook founder opened his checkbook wide, reportedly dangling offers in the hundreds of millions of dollars—sometimes even close to a billion, according to Futurism—to lure top AI minds from rivals like OpenAI, DeepMind, and Apple.

But just months after the big unveil, the lab is already facing turbulence. According to Wired, at least three researchers have quit, two of whom are returning to OpenAI after only brief stints at Meta.

The exits begin

Avi Verma and Ethan Knight, both previously associated with OpenAI, resigned from MSL and are now headed back to their former employer. Another high-profile exit is Rishabh Agarwal, an Indian researcher poached from Google DeepMind at a reported $1 million salary. Agarwal had joined Meta in April but announced on August 25 that this would be his last week at the lab.



In a post on X, he admitted the decision was difficult given the “talent and compute density” at MSL but added that he felt drawn to “a different kind of risk” after more than seven years at Google Brain, DeepMind, and Meta.

Notably, Agarwal quoted Zuckerberg himself in his farewell: “In a world that’s changing so fast, the biggest risk you can take is not taking any risk.” The remark has since been widely viewed as researchers turning Zuckerberg’s own mantra against him.

Why researchers are walking away

While neither Verma nor Knight commented on their departures, Wired notes that Meta’s AI operations have been plagued by reorganizations, shifting priorities, and reports of micromanagement at the very top. The company recently split its AI workforce into four groups, creating further uncertainty inside MSL.

Meanwhile, Agarwal’s exit highlights a tension many in the field acknowledge: the most talented researchers are not solely motivated by eye-popping salaries. DeepMind cofounder Demis Hassabis told Lex Fridman that true frontier scientists want to “help influence how AGI plays out and steward the technology safely into the world”—not simply chase paychecks.

This sentiment is echoed across the industry. Anthropic’s cofounder Benjamin Mann recently remarked, “My best case at Anthropic is we affect the future of humanity. My best case at Meta is we make money.”

OpenAI welcomes Meta defectors

Meta’s losses appear to be OpenAI’s gain. Alongside Verma and Knight, longtime Meta executive Chaya Nayak has also joined OpenAI, where she will work on special initiatives. Her exit after nearly a decade at Meta is seen as another blow to Zuckerberg’s plans.

The timing is notable. OpenAI, which had criticized Meta’s recruitment style as “distasteful,” now finds its bench strengthened by former Meta hires who chose mission alignment over compensation.

What the departures mean

Meta has invested heavily to close the gap with rivals in artificial general intelligence research, appointing former Scale AI chief Alexandr Wang and GitHub’s former CEO Nat Friedman to leadership positions. Yet the high-profile resignations suggest that a generous paycheck may not be enough to build and sustain a cohesive frontier AI team.

For Zuckerberg, the irony is sharp: after positioning himself as a visionary risk-taker, it is his recruits who are citing his own words as they walk out the door.

Whether Meta can stabilize its superintelligence project or continues to bleed talent will determine how it fares in the next phase of the AI race.

The broader AI talent war

Meta’s troubles also come as competition for elite AI researchers intensifies across Silicon Valley. Elon Musk’s xAI has reportedly lured away at least 14 engineers from Meta this year, despite Zuckerberg’s massive retention bonuses. Rather than outbidding rivals, Musk is pushing a “performance-first” culture built on equity and relentless pace, a contrast to Meta’s cash-heavy offers. Industry leaders warn that the escalating pay arms race risks creating toxic dynamics, exposing a deeper cultural divide between money-driven recruitment and mission-driven work.

AI’s growing impact on workers

Beyond the lab wars, artificial intelligence is already reshaping the job market. A new Stanford study warns that entry-level positions are eroding fastest in AI-exposed fields such as software engineering, customer service, and marketing, with employment for 22 to 25-year-olds dropping sharply. Older workers remain stable, but researchers caution that the loss of these “first rungs” could destabilize entire career ladders, making young professionals the earliest casualties of AI disruption.