Yann LeCun, one of the “AI godfathers” and head of Meta’s AI lab since 2013, has just announced that he will leave the company at the end of the year to found a new startup.

The information was revealed in a memo sent to all Meta employees, obtained by Bloomberg News . However, Meta plans to cooperate deeply with the new company and “will have access to the startup’s innovations,” according to the internal memo.

xyt188yc.png
Yann LeCun is chief AI scientist at Meta. Photo: Bloomberg

Professor LeCun said he is building a startup focused on Advanced Machine Intelligence (AMI) – an approach to AI based on the ability to learn from visual and sensory data, allowing machines to understand the physical world , remember long-term, reason and create complex sequences of actions. He described this as “the next big revolution in AI”, and a continuation of research he and his colleagues at Meta and New York University have been pursuing for many years.

“The startup’s goal is to create the next big AI revolution: systems that understand the physical world, have persistent memories, and are capable of reasoning and complex action planning,” LeCun wrote in the memo, emphasizing that the technology will have broad applications across many economic sectors.

Despite being considered an icon in the AI ​​research community, Yann LeCun has struggled in recent years to secure resources for long-term projects, especially as Meta shifted its focus to large language models (LLMs) to compete with OpenAI, Google, and Anthropic.

He has also repeatedly criticized Meta’s LLM-focused strategy, arguing that AI needs a world model—like JEPA—to achieve human-level intelligence. This view has led to repeated clashes with internal teams, and he has been accused of “no longer representing Meta’s AI strategy.”

After Meta launched Llama 4 but failed to make a splash and began pivoting toward “hyperintelligence,” some executives wanted LeCun to limit his public appearances for fear that he would continue to oppose LLM or the company’s new perspective, insiders said. He is also a strong proponent of open source, while Meta has been quietly moving away from that strategy.

Meta spent billions of dollars recruiting talent when it set up a new AI center this summer, but chose an outsider to lead the new team, a sign that LeCun is no longer as central as before.

Professor LeCun’s departure from Meta marks a significant shift for one of the most influential figures in AI. With a startup centered around AMI, he is betting on a path that is far different from the current LLM race, while Meta, while remaining a partner, is also facing a new era of AI without LeCun at the forefront.

Professor Yann LeCun visited Vietnam to attend the VinFuture 2024 Science and Technology Week in December 2024. He won the VinFuture Grand Prize for his pioneering research on convolutional neural networks for computer vision (CNNs), an important model in the development of image recognition and deep learning technology. CNNs have now become the standard in AI applications used by billions of people every day, playing a central role in the development of technologies such as facial recognition and medical image processing.

(According to Bloomberg, VinFuture)

Prominent Chinese scientist criticizes Meta Meta 's controversial decision to lay off around 600 employees in its AI research unit has exposed internal tensions. Prominent Chinese scientist Tian Yuandong is among those who have spoken out against Facebook's parent company.

Source: https://vietnamnet.vn/bo-gia-ai-yann-lecun-roi-meta-2464590.html