The AI Legend is Leaving Meta

Yann Lecun is one of the pioneers in AI. Photo: Tuan Anh .

Yann LeCun, Meta’s chief AI scientist , is planning to leave the company to start his own startup, according to the Financial Times . Over the course of more than a decade, he became one of the three “fathers” of modern machine learning.

LeCun, a professor at New York University, a senior researcher at Meta, and a winner of the prestigious AM Turing Award, is expected to leave the company in the coming months. He has now begun talks to raise capital for his startup, which aims to continue developing its research on “ world models.”

It is an AI system that develops an innate understanding of its environment, simulating cause-and-effect scenarios to predict outcomes. Leading labs and startups like Google DeepMind and World Labs are also pursuing this line of research.

LeCun’s departure comes at a pivotal time for AI competition in the market. Meta recently changed its approach to AI development in response to concerns about being left behind by rivals like OpenAI, Google, and Anthropic.

The company has begun restructuring its AI organization, rebranding it as Meta Superintelligence Labs (MSL), after hiring more than 50 engineers and researchers from rival companies. Notably, in June, Meta invested $14.3 billion in data labeling company Scale AI, bringing in its CEO, Alexandr Wang, to lead the new division.

According to TechCrunch in August, those decisions left Meta’s AI division in greater disarray than ever, with new hires frustrated by the cumbersome bureaucracy of a large corporation, while the previous team had its scope of operations significantly reduced.

MSL is currently undergoing a round of layoffs of around 600 employees under the leadership of Alexandr Wang. The layoffs are mainly focused on groups such as Fundamental AI Research (FAIR), AI Infrastructure/Products, while the superintelligence group is said to continue hiring.

CEO Mark Zuckerberg’s decision to restructure the lab overshadowed LeCun’s long-term work in FAIR, after the company’s previous AI model, Llama 4, failed to keep up with its competitors. Unlike MSL, FAIR was founded with a focus on long-term AI research, developing techniques that could be applied in the next 5-10 years.

LeCun has long been skeptical of the way AI technology, particularly large language models (LLMs), is being promoted as a panacea for humanity. He even posted on X that AI development is still a long way off.

“In my opinion, before we urgently seek to control AI systems that are superior to humans, we need at least a preliminary idea of the design of a system smarter than a house cat,” he wrote.

Source: https://znews.vn/huyen-thoai-ai-sap-roi-meta-post1602037.html