Meta’s Already Bleeding AI Talent Two Months Into Hiring Spree

Meta CEO Mark Zuckerberg wants to be the first to achieve AI superintelligence, a type of artificial intelligence that is smarter than the average human, and a concept that some experts are still highly skeptical.

To do so, Zuckerberg announced the formation of Meta Superintelligence Labs in late June, a team dedicated to helping AI catch up to and beat its peers in the AI race. 

With that, he sparked a multi-million dollar hiring frenzy. Meta began poaching talent from competitors, the main target seemingly being OpenAI, promising them multi-year deals worth millions of dollars. OpenAI executives at the time described the feeling as though “someone has broken into our home and stolen something.”

But OpenAI may have gotten the final laugh, at least for now. Just two months in, the superintelligence team is already bleeding talent, WIRED and Business Insider report. 

At least three AI researchers have resigned from Meta’s superintelligence lab, and two of these researchers returned to work for their previous employer, OpenAI, after less than a month of employment at Meta, the report claims. 

“During an intense recruiting process, some people will decide to stay in their current job rather than starting a new one. That’s normal,” a Meta spokesperson told Gizmodo.

A separate Business Insider article found that at least eight employees have left the superintelligence team, including not only researchers but also engineers and a senior product leader. Some of the employees, BI found, switched over to OpenAI after having been at Meta for years.

Sign of the times?

Late last month, Meta executives said they expect to shell out between $66 billion and $72 billion this year, and it will be even more next year, driven by data centers and AI hiring. The announcement did not frighten investors at the time, it seems, as the company’s stock went up in response to the news. In an investor call that week, Zuckerberg assured investors that the company was reaping the benefits of its AI push and was likely to see more thanks to the superintelligence initiative.

But then, just a few weeks later, Meta made a surprise decision to split up its AI division, the Meta Superintelligence Labs, into four smaller groups less than two months after its formation. Along with the restructuring, the New York Times reported that the company was also looking at downsizing its AI division.

Meta’s AI catch-up efforts so far the past few months have been ill-received by users and riddled with controversy, specifically regarding its chatbots. 

The company came under fire and found itself in the midst of a Senate probe after a Reuters report found that Meta allowed its generative AI assistants and chatbots to engage in “sensual” conversations with minors, reaffirm racist beliefs, and give false medical information.

Texas attorney-general Ken Paxton’s office also opened an investigation into Meta’s chatbots, which it claims have impersonated licensed mental health professionals.

OpenAI vs Meta

Meta’s AI corporate overhaul over the summer was marked by its talent raid of OpenAI researchers. Zuckerberg even allegedly tried and failed to recruit OpenAI co-founder Ilya Sutskever and John Schulman.

OpenAI CEO Sam Altman called Meta’s approach at the time mercenary and said the company lacked culture.

Now, according to reports, some of those hires are returning to OpenAI after less than a month at Meta, along with some Meta veterans.

Chaya Nayak, Meta’s director of generative AI product management, who has been with the company for almost nine years, shared in a LinkedIn post yesterday that she will be joining the special initiatives team at OpenAI.

The WIRED report also claims that former OpenAI researcher Shengjia Zhao, who made headlines late last month when he joined Meta as chief scientist of the superintelligence lab, actually tried to return to OpenAI before the announcement was made. Zhao even allegedly went so far as to sign employment paperwork with his former employer.

If the reports are true, it looks like Meta’s quest for artificial superintelligence and AI market domination is off to a rocky start.