Meta’s Chief AI Scientist to Depart for New Startup
Meta’s leading AI scientist is reportedly set to leave the company, which marks a significant departure as CEO Mark Zuckerberg aims to develop a “superintelligence.”
Yann LeCun, 65, a Turing Award laureate and an influential figure in AI, has informed colleagues of his plan to exit Meta in the coming months, according to sources familiar with the situation.
The French-American scientist is currently in initial discussions to secure funding for his new venture, officials indicated.
Shares of Meta Inc. dipped by 1.2% on Tuesday morning as investors expressed concerns about whether Zuckerberg could validate the substantial investments in new technologies.
Zuckerberg is striving to revitalize the AI division as Meta faces stiff competition from industry leaders like OpenAI and Google.
He has urged employees to accelerate product rollouts, marking a shift from the long-term research focus of Meta’s Fundamental AI Research Lab (FAIR) that LeCun has led since 2013.
Meta, the parent company of Facebook, Instagram, and WhatsApp, invested over $14 billion this summer to acquire a 49% share in Scale AI and appointed its founder, Alexandr Wang, to head the new “superintelligence” division.
LeCun, who used to report to chief product officer Chris Cox, will now be under Wang’s leadership, as reported.
Zuckerberg has also attracted talent from rival firms with compensation packages exceeding $100 million, leading to dissatisfaction among long-time Meta employees.
In July, Meta brought on Shengjia Zhao, a co-creator of OpenAI’s ChatGPT, as the chief scientist for its “superintelligence” division.
These new teams are focused on scaling language models to compete with big players like ChatGPT, especially after setbacks with Meta’s Llama 4 model and its AI chatbot struggling to gain traction.
LeCun and Zuckerberg seem to have differing perspectives on AI’s future.
The award-winning scientist views the company’s large-scale language models as valuable but contends they can’t fully mirror human reasoning. In contrast, Zuckerberg is optimistic about AI’s capabilities, suggesting that much of Meta’s coding work could be accomplished within a year.
At FAIR, LeCun has been developing a new generation of AI systems that can understand the physical world through language, video, and spatial data. However, he has warned that these models might take a decade for full development.
His upcoming project will further this global model development, according to insiders.
Representatives for LeCun did not immediately respond to requests for comments.
Zuckerberg has indicated that the costs associated with Mehta’s “superintelligence” lab could reach hundreds of billions. However, Meta’s leaders, like other tech company heads, are under increasing pressure to demonstrate that their hefty spending will yield returns. Recently, fears surrounding AI’s overvaluation triggered a significant sell-off in tech stocks.
In late October, Meta’s stock crashed over 12%, erasing nearly $240 billion from its market value after Zuckerberg mentioned that AI expenditures might surpass $100 billion next year.
The company has also seen a loss of key personnel, such as Joelle Pineau, the vice president of research, who left to join Canadian AI startup Cohere in May. Last month, Meta cut 600 positions in its AI research team to reduce costs and streamline operations.
