Meta’s ambitious A.I. unit is struggling to keep talent despite aggressive hiring. Photo by Matt McClain/The Washington Post via Getty Images
Through a recruiting frenzy that has included $100 million bonus offers, personal meetings and aggressive poaching tactics, Mark Zuckerberg has hired roughly 50 new A.I. researchers at Meta in record time. But keeping that talent at the company is proving to be a different challenge. Several recent hires, along with a few longtime staffers, have already left Meta Superintelligence Labs (MSL) division, the newly formed division tasked with advancing the company’s most ambitious A.I. projects.
MSL is divided into four groups. The largest, known as TBD, is made up of many of the new hires and is focused on researching superintelligence, a form of A.I. with capabilities beyond those of humans. The other three teams are working on A.I. products, infrastructure and Meta’s long-running Fundamental A.I. Research (FAIR) efforts. MSL is led by Alexandr Wang, the founder of Scale AI who joined Meta after the tech giant acquired 49 percent of his startup for $14 billion.
The early days of MSL have been turbulent. Shengjia Zhao, a former OpenAI researcher who joined Meta just a few months ago, reportedly threatened to return to OpenAI within days of arriving, according to the Financial Times. To retain him, Zuckerberg offered him the title of Meta’s chief AI scientist, the FT reported, citing people familiar with the matter.
“Shengjia co-founded MSL and has been our scientific lead since day one,” Meta said in a statement. “We formalized his role once our recruiting had ramped and the team had taken shape.”
While Zhao ultimately decided to stay, others have already walked away—sometimes after only a few weeks on the new superintelligence team. Here’s a look at the shakeups so far inside MSL:
Ethan Knight
The most notable departure from Meta’s superintelligence unit thus far is Ethan Knight, who will reportedly be starting a new role at OpenAI after only a few weeks at Meta. Knight formerly interned for OpenAI’s A.I. safety team between 2018 and 2019.
He arrived at Meta earlier this summer following a stint at Elon Musk’s xAI and about four years at Tesla, where he worked on the company’s computer vision initiative.
Avi Verma
Avi Verma, a prospective member of Meta’s TBD lab, went through onboarding but never showed up for his first day, according to the Financial Times. Instead, he chose to remain at OpenAI, where he’s worked as a researcher since last summer. He also formerly spent more than three years at Tesla as a machine learning engineer and software engineer.
Rishabh agarwal
Rishabh Agarwal, another would-be member of the TBD team, announced his departure in an X post on Aug. 25. “It was a tough decision not to continue with the new Superintelligence TBD lab, especially given the talent and compute density,” said Agarwal, adding that he “felt the pull to take on a different kind of risk” despite an “incredibly compelling” pitch from Zuckerberg and Wang.
Agarwal joined Meta in April and spent five months working on model post-training. A former researcher at Google Brain and DeepMind, he did not disclose his next role.
Chaya Nayak
Not all departures come from recent recruits. Chaya Nayak, who left Meta for OpenAI in August, had been at Meta for about nine years. She joined in 2016 to focus on data projects and, more recently, worked on generative A.I. projects, including Meta’s Llama models and Meta AI assistant, and was part of the company’s broader superintelligence lab.
She will now join OpenAI’s “special initiatives” team led by Irina Kofman. “It feels like the perfect next chapter: to take everything I’ve learned, and pour it into work that will help define what comes next for technology and society,” Nayak wrote in an Aug. 26 LinkedIn post.
Loredana Crisan
Loredana Crisan, too, worked at Meta for nine years before stepping away earlier this month. Having formerly headed Messenger and Instagram DMs, Crisan joined Meta’s A.I. division in February to oversee generative A.I. products.
Crisan is leaving to become chief design officer at Figma. “Now, A.I. carries design into its most empowered chapter yet: where tedium fades while taste, exploration and play take center stage,” she wrote in a post on X.
