Meta’s chief AI scientist Yann LeCun reportedly plans to leave to build his own startup

Yann LeCun, a prominent figure in artificial intelligence and chief scientist at Meta, is reportedly planning to depart the company to embark on his own startup journey. According to a report from the Financial Times, which cites unnamed sources, LeCun intends to leave within the next few months and is currently in discussions to secure funding for a new venture focused on advancing his research in world models. World models represent a significant stride in AI development, creating systems that possess an internal comprehension of their surroundings to simulate various scenarios and predict outcomes. This innovative area is also being explored by leading organizations such as Google DeepMind and World Labs. LeCun’s exit would come during a critical juncture for Meta as the firm shifts its AI strategy amid rising competition from entities like OpenAI, Google, and Anthropic. The company has been restructuring its AI division, recently hiring over 50 engineers and researchers from rival firms to bolster its new AI unit, named Meta Superintelligence Labs (MSL). In a notable strategic move, Meta invested $14.3 billion in the data-labeling company Scale AI earlier this year and appointed its CEO, Alexandr Wang, to lead the new division. However, this rapid expansion has led to increasing disarray within Meta’s AI team, with new recruits expressing frustration over the challenges of navigating the complexities of a large corporation. Furthermore, the capabilities of Meta's previous generative AI team have been curtailed. LeCun’s extensive research under Meta's Fundamental AI Research Lab (FAIR) has been somewhat overshadowed by CEO Mark Zuckerberg’s initiatives to revamp the company’s AI framework, particularly following the underperformance of its Llama 4 AI models against competitors. Unlike MSL, FAIR concentrates on long-term AI research, focusing on methodologies that may be applicable in the next five to ten years. In recent statements, LeCun has voiced skepticism regarding the overzealous promotion of AI technologies, especially large language models (LLMs), as solutions to every global challenge. He has emphasized the need for more foundational advancements, asserting, “Before ‘urgently figuring out how to control AI systems much smarter than us,’ we need to have at least a conceptual design for a system smarter than a house cat.” Meta has not provided any comments regarding LeCun’s potential departure as of yet.