The reason why the 'father' of AI left Meta.

Yann Lecun, 65, is a leading AI scientist at Meta. Photo: Bloomberg .

Yann LeCun, chief scientist in charge of foundational AI research at Meta, will soon be leaving the company, according to several reliable sources. At 65, he is a pioneer in the field of AI science, possessing an impressive body of knowledge.

Meanwhile, CEO Mark Zuckerberg stated that Meta is making internal progress to the point of being very close to superintelligence, after continuously poaching top talent from competitors. Therefore, the departure of a leading scientist from a company that is spending heavily on AI has sparked public discussion.

In fact, Lecun has previously expressed very different beliefs about AI. He recently gained notoriety for arguing that large language models (LLMs) are no longer worth pursuing, leading to distractions and ultimately a dead end, no matter how much investment is made.

A WSJ analysis of LeCun's career suggests that some of the other possibilities stem from differences in career paths. Earlier this year, Alexandr Wang, 28, became head of AI at Meta, and LeCun's superior. The company then brought in another relatively young chief scientist, Shengjia Zhao.

In Meta's announcement, Zhao's new role was highlighted based on a breakthrough related to his ability to increase model size. LeCun, however, no longer believed this.

According to the FT, Lecun will launch a startup focused on "world models," arguing that this is what will answer questions about AI. Having participated in the development of Meta's AI smart glasses, he strongly believes that in the future, humans will interact with wearables, not LLMs, because they don't understand the world .

Yann Lecun once conducted an experiment demonstrating the lack of sensitivity of LLM. Photo: Shutterstock.

"We can't even replicate the intelligence of cats or mice, let alone dogs," he said. World models are AI models designed to understand and simulate the world, based on how objects and events behave, and the consequences of actions.

Scientists believe that world models will allow the construction of systems capable of hierarchically planning actions to achieve a goal. These systems will have stronger security features because the control mechanisms will be integrated directly into them.

LeCun presented a thought experiment to illustrate what world models, or any human being, can do, but LLMs cannot. Imagine a cube floating in front of you, rotating 90 degrees around its vertical axis. He stated that LLMs can write explanatory text, but they cannot help users actually interact with the cube.

LeCun began working on world models at Meta. In his presentation at the AI Action Summit, he described a dream model capable of estimating the current state of the world. Instead of predicting based on consecutive tokens like LLM, this model predicts the resulting state of the world after a user performs a certain sequence of actions.