In a significant revelation that has sent shockwaves through the tech world, Yann LeCun, one of the founding fathers of modern artificial intelligence and a long-standing pillar at Meta, has publicly detailed the reasons behind his dramatic departure from the social media giant. The 65-year-old Turing Award winner cited mounting pressure from CEO Mark Zuckerberg and a deep philosophical rift over the company's strategic pivot towards large language models (LLMs) as the core reasons for his exit after more than a decade.
The Breaking Point: A Fudged Benchmark and Lost Confidence
In a candid interview with the Financial Times, LeCun disclosed that the immediate trigger was Zuckerberg's reaction to the performance of Llama 4, Meta's generative AI model released in April 2025. LeCun stated that the generative AI team had "fudged" benchmark results for the model, an act that severely disappointed the CEO. "Mark was really upset and basically sidelined the entire GenAI organisation," LeCun revealed. He further predicted a wave of resignations, saying, "a lot of people who haven't yet left will leave."
This incident, according to LeCun, led Zuckerberg to "basically lose confidence" in the existing generative AI leadership. Subsequently, Meta made a high-profile hire, bringing on 28-year-old Alexandr Wang, founder of Scale AI, to lead its superintelligence efforts after investing a staggering $14.9 billion in his company.
Philosophical War: LeCun Declares LLMs a 'Dead End'
The departure underscores a fundamental clash of visions for the future of artificial intelligence. While Zuckerberg is betting billions on developing ever-larger language models in pursuit of artificial superintelligence, LeCun has been a vocal sceptic of this approach. He minced no words in his criticism, telling the FT, "I'm sure there's a lot of people at Meta, including perhaps Alex, who would like me to not tell the world that LLMs basically are a dead end when it comes to superintelligence."
LeCun argues that LLMs, which are trained primarily on text, are inherently limited because they lack a true understanding of the physical world. His new startup, Advanced Machine Intelligence Labs, will instead focus on developing "world models" that learn from videos and spatial data. He believes this approach is crucial for creating AI that can genuinely comprehend and interact with reality.
On Alexandr Wang: "Young and Inexperienced"
LeCun also offered a blunt assessment of his successor, Alexandr Wang. While acknowledging that Wang "learns fast" and "knows what he doesn't know," LeCun pointedly highlighted his lack of research experience. "There's no experience with research or how you practice research," LeCun said, adding a firm principle of academic freedom: "You don't tell a researcher what to do. You certainly don't tell a researcher like me what to do."
Despite the professional tensions and public airing of grievances, LeCun maintained that he remains on good personal terms with Mark Zuckerberg. He also confirmed that Meta has agreed to partner with his new AI venture, suggesting the door for collaboration remains open even as their strategic paths diverge.
This exit marks a pivotal moment for Meta's AI ambitions, losing one of the field's most respected and original thinkers. It raises critical questions about whether the industry's current obsession with LLMs is the definitive path to advanced machine intelligence or if visionaries like LeCun are pointing toward a more promising, yet challenging, road ahead.
