According to Financial Times News, Meta’s chief artificial intelligence scientist Yann LeCun—a Turing Award winner considered one of the pioneers of modern AI—is planning to leave the social media giant in the coming months to found his own startup. The French-US scientist is already in early talks to raise funds for his new venture, which will focus on his “world models” research. This comes as Mark Zuckerberg radically overhauls Meta’s AI strategy after the company fell behind rivals like OpenAI and Google. Zuckerberg recently hired 28-year-old Scale AI founder Alexandr Wang to lead a new “superintelligence” team, paying $14.3 billion to acquire a 49% interest in his company. LeCun, who previously reported to chief product officer Chris Cox, now reports to Wang, and Meta’s shares plunged 12.6% in late October after Zuckerberg signaled AI spending could top $100 billion next year.
The great Meta AI shakeup
Here’s the thing—this isn’t just one guy leaving. This is a complete philosophical shift in how Meta approaches artificial intelligence. Zuckerberg is basically saying “screw long-term research, we need products yesterday.” He’s pivoting hard from LeCun’s Fundamental AI Research Lab (Fair) toward rapid deployment and commercial applications. And he’s putting his money where his mouth is—hiring young hotshots like Alexandr Wang with billion-dollar deals and creating exclusive teams like TBD Lab with $100 million compensation packages to poach talent from rivals.
LeCun’s different vision
So why is LeCun really leaving? Look, he’s been pretty vocal about his skepticism of the current LLM craze. He thinks these large language models are “useful” but will never achieve true human-like reasoning. Meanwhile, he’s been chasing this “world models” concept—AI that learns from videos and spatial data to actually understand the physical world. That’s a decade-long research project, minimum. Basically, he wants to build the next generation of AI while Zuckerberg wants to ship what works today. You can see why there might be some tension there.
Wall Street is watching
And let’s not forget the pressure Zuckerberg is under. When you wipe out $240 billion in market value with one earnings call—which happened in October after he signaled massive AI spending—you’ve got to show results fast. Wall Street wants to see that $100 billion investment pay off in revenue, not in some far-off research project that might pan out in 2035. This puts Meta in a tough spot—do you chase immediate commercial gains or bet on foundational research that could actually leapfrog the competition?
What happens now?
LeCun’s departure marks a significant brain drain for Meta’s research efforts. He’s not alone—VP of AI research Joelle Pineau left for Cohere in May, and Meta laid off about 600 people from its AI research unit recently. The old guard is being replaced by high-priced external hires, which always creates cultural friction. Meanwhile, LeCun gets to pursue his world models vision without corporate constraints. The big question is whether Zuckerberg’s product-focused approach will actually catch Meta up to OpenAI and Google, or whether they’ll end up playing perpetual catch-up while the real breakthroughs happen elsewhere.
