Meta Superintelligence Labs is already losing key staff less than two months after launch

Meta's ambitious endeavor to spearhead 'superintelligence' appears to be faltering as it experiences significant employee turnover just weeks after its launch. In a surprising twist, at least eight key personnel, including seasoned researchers and engineers, have opted to leave the newly formed Meta Superintelligence Labs (MSL). This division was championed by CEO Mark Zuckerberg, who aimed to democratize access to advanced AI capabilities. Among those departing are long-time Meta employees who played crucial roles in developing the company’s foundational AI systems. The departures include veterans and newer recruits alike, all lured by Zuckerberg's vision. A Meta spokesperson acknowledged that some level of attrition is expected in such a large organization, wishing the departing employees well. The exits come on the heels of Zuckerberg's substantial financial investments aimed at narrowing the competition with AI leaders like OpenAI and Google. To attract talent to the new division, Meta has reportedly offered lucrative compensation packages worth hundreds of millions, creating friction between newer and veteran employees. The atmosphere has become tense, particularly among those who have been with the company prior to this new focus on superintelligence. Notable departures this month include Bert Maher, who had dedicated over a decade to Meta and was instrumental in developing tools like PyTorch, an essential open-source library for AI training. Maher recently joined Anthropic, expressing enthusiasm about enhancing their AI chatbot, Claude. Another significant loss is Tony Liu, who managed teams focused on PyTorch GPU systems, key for Meta’s AI model training, and now plans to share insights through a newsletter. Chi-Hao Wu, a machine learning expert, also announced his exit to take on the role of chief AI officer at Memories.ai. Wu pointed out that the frequent organizational changes at Meta may have contributed to feelings of instability among AI team members. The company's AI division has undergone multiple restructurings this year, including the dissolution of teams. In light of this, Meta's communications chief has criticized media coverage of these changes as overly focused on internal dynamics. Other notable exits include Aram Markosyan, who worked on AI safety and features for smart glasses, and several former Meta employees transitioning to OpenAI, Meta's primary competitor. Chaya Nayak, who contributed to generative AI initiatives, is now moving to OpenAI to focus on specialized projects. Despite the turmoil, Meta's spokesperson downplayed the situation, suggesting that the decision-making process during recruitment can lead some individuals to opt for their existing roles instead of new opportunities. As Meta continues to grapple with these challenges, the path forward for its superintelligence ambitions remains uncertain.