Meta's newly launched Superintelligence Lab is facing significant challenges as at least eight top AI researchers have left the company, with some returning to OpenAI after less than a month. The lab, announced by CEO Mark Zuckerberg in June, aims to develop advanced AI systems rivaling OpenAI's capabilities. Despite offering nine-figure compensation packages to lure top talent, Meta struggles with talent retention, internal reorganizations, and cultural friction.
The departed researchers include Avi Verma and Ethan Knight, who rejoined OpenAI after brief stints at Meta, and Rishabh Agarwal, who cited a desire to take on different risks. Chaya Nayak, a nearly nine-year Meta veteran who directed product for generative AI, also joined OpenAI for special initiatives. Other notable departures include Bert Maher, a 12-year Meta veteran who built PyTorch and joined Anthropic, and Tony Liu, who led PyTorch GPU systems and stepped away after eight years.
Meta's approach to building its superintelligence lab has been questioned, with some speculating that the company's bureaucratic and recruitment issues may be hindering its progress. The lab's structure and expectations may not have aligned with the recruiting pitch, leading to these departures. Despite this, Meta's spokesperson Dave Arnold said that some attrition is normal, especially during intense recruiting processes.
The departures raise questions about Meta's ability to retain talent and compete in the AI race. With OpenAI and other competitors attracting and retaining top researchers, Meta faces significant challenges in achieving its AI ambitions. The company's repeated reorganizations of its AI teams and struggles with internal friction may be contributing to the turbulence.
