Apple's Key Multimodal AI Researcher to Join Meta's Superintelligence Lab
Apple Struggles Amid Wave of Talent Departures
Zhang was part of the Apple Foundation Model (AFM) team, which built Apple's AI platform technologies. He is the fourth AI researcher to move from Apple to Meta in the past month. Previously, Luo Ming Fang, who led the AFM team and spearheaded AI model development, also moved to Meta with a compensation package reportedly exceeding $200 million.
The AFM team, consisting of dozens of engineers and researchers based in both Cupertino, California and New York, developed the core technologies behind the Apple Intelligence platform launched last year. Bloomberg reported that the recent series of key personnel departures has thrown the AFM team into turmoil. Internal concerns over Apple's unclear AI strategy and the limitations of its proprietary models are fueling these departures. There is also a growing sense of crisis within Apple that relying solely on in-house developed models makes it difficult to keep pace with competitors such as OpenAI and Meta.
In contrast, Meta is aggressively recruiting top AI talent from rival companies. Meta recently established a research lab with the goal of developing superintelligent AI that surpasses human capabilities, and has recruited a large number of researchers from competitors including OpenAI, Google, Apple, and Anthropic.
Meanwhile, according to the IT media outlet Wired, Meta has also approached more than 10 researchers from Thinking Machine Lab, an AI startup founded by former OpenAI CTO Mira Murati, offering them positions. One of these researchers reportedly received an unprecedented compensation package exceeding $1 billion.
After leaving OpenAI last September, Murati founded her own startup, which has since attracted several key former OpenAI employees. Industry observers believe Meta's aggressive talent acquisition strategy will further accelerate the race for dominance in next-generation AI.
© The Asia Business Daily(www.asiae.co.kr). All rights reserved.


