Yann LeCun, Meta's AI Chief, Departing to Establish Independent Artificial Intelligence Research Firm
In a significant move within the artificial intelligence landscape, Yann LeCun, Meta's Chief AI Scientist, has announced his departure from the tech giant to launch an independent AI research company. LeCun, a Turing Award winner and one of the foremost figures in the field, is set to leave Meta at the end of the year, marking the end of a twelve-year tenure.
LeCun's new venture will focus on developing what he calls "Advanced Machine Intelligence" (AMI), a research agenda centered on creating AI systems that possess a deeper understanding of the physical world, persistent memory, reasoning capabilities, and the ability to plan complex action sequences. This approach contrasts with the current industry trend of focusing on large language models (LLMs), which LeCun has publicly expressed skepticism about, viewing them as a potentially limited path toward achieving human-level AI. He believes LLMs are useful but not the ultimate solution for better-than-human AI.
LeCun joined Facebook, later Meta, in 2013 and played a pivotal role in establishing the company's AI research division, initially known as Facebook AI Research (FAIR). While he stepped down as the director of FAIR in 2018, he remained Meta's Chief AI Scientist, contributing to various AI advancements within the company. His work at Meta included developing applications such as targeted marketing based on AI chatbot interactions. LeCun's departure follows Meta's recent shift towards more commercially driven AI endeavors. In June, Meta invested $14.3 billion in the AI data company Scale and recruited its CEO, Alexandr Wang, to spearhead a team focused on developing "superintelligence". This strategic pivot, coupled with recent AI job cuts at Meta, suggests a change in priorities that may have influenced LeCun's decision to pursue his independent research path.
Despite venturing out on his own, LeCun has stated that Meta will be a partner in his new company. The collaboration suggests that some of the research conducted by the startup will align with Meta's commercial interests, while other aspects will remain independent. LeCun's vision for AMI involves moving beyond purely language-based learning, emphasizing models grounded in perception and interaction. He aims to build AI systems that can learn from video and spatial data to construct intuitive "world models," simulating physical environments and predicting outcomes in a manner similar to human intuition.
LeCun's departure highlights a growing tension within the AI community between open research and product-oriented development. His focus on AMI reflects a belief in the importance of foundational research and the exploration of alternative approaches to achieving advanced AI. As AI continues to evolve, LeCun's new venture could potentially lead to breakthroughs in areas such as robotics, computer vision, and natural language understanding. He also remains a part-time professor at New York University, affiliated with the NYU Center for Data Science.
