The announcement that Yann LeCun, a prominent figure in the AI field, is departing from Meta and intends to launch his own venture sent shockwaves through the market. Before the market even opened, Meta's pre - market value plummeted by 1.5%, which translates to a staggering loss of over $20 billion (equivalent to more than 1.4 trillion yuan). And the downward trend of the stock price persisted after the market opened.

There were early telltale signs of LeCun's impending departure. Following Meta's adjustment of its AI strategy, the strategic direction kept shifting frequently, and the organizational structure was in a state of constant flux. LeCun's standing within the company gradually diminished. He found himself in a position where he had to report to Alexander Wang, who was relatively less experienced. Moreover, his published research achievements were subjected to stringent review restrictions.

In addition, the FAIR (Facebook AI Research) team, which LeCun founded, was hit hard by massive layoffs.

Technical differences also played a significant role in driving a wedge between LeCun and Meta. LeCun was a staunch advocate of the world model architecture and held a skeptical view of large language models. In contrast, Meta placed its bets on large language models. When it came to the debate between open - source and closed - source, LeCun firmly supported open - source initiatives. Meanwhile, within Meta, there were ongoing discussions in favor of closed - source approaches.

LeCun joined Meta back in 2013 and was the driving force behind the establishment of FAIR. As a recipient of the prestigious Turing Award, his influence in the AI community was immense. The FAIR model he championed was once regarded as a hallmark of Meta's AI capabilities. However, his departure now signals the end of an era—a golden age of "academic" research at Meta.