Yes, LeCun Reports to 28-Year-Old Alexandr Wang! Inside Meta’s New AI Team Revealed}

LeCun reports to Alexandr Wang? Inside Meta: Yes.

After recruiting a host of AI experts, many wonder if Meta can handle the scale. Recently, Meta’s aggressive talent acquisition, including offering $100 million salaries and targeting competitors like OpenAI, has shaken the tech world.

      ^{According to Bloomberg, Meta’s bid to recruit Apple’s AI head Pang Ruoming has reached a staggering $200 million.}

Following the initial leadership setup, Meta’s new AI research division, led by former Scale AI CEO Alexandr Wang, known as the “Superintelligence Lab” (Meta Superintelligence Labs, MSL), has become a focal point of attention regarding its structure and research directions.

There are many exaggerated rumors, prompting some Meta employees to clarify the facts.

Meta FAIR scientist Zhu Zeyuan clarified on X that: first, FAIR is a long-term basic science research institution, different from the Llama team; second, FAIR conducts open research on open data, and its data, code, and GPUs are not shared with GenAI or MSL.

Our sources also revealed some personnel movements behind the formation of MSL.

The most dramatic question: Is Yann LeCun, the Turing Award winner and pioneer in deep learning, reporting to Alexandr Wang, who was born in 1997?

The answer from insiders: Yes, but it’s not important.

Meta insiders told us that Yann LeCun’s FAIR has reported to Reality Labs since 2022, and has undergone several changes in its reporting hierarchy. Before this, FAIR was under the Chief Product Officer overseeing Family of Apps like Facebook and Instagram.

It’s worth noting that Yann LeCun founded FAIR in 2013, which was later renamed Fundamental AI Research, an independent research organization with about 600 members including visiting scholars, interns, and postdocs.

This independence means it cannot report directly to Zuckerberg, so “affiliation” is inevitable. The recent change to report to MSL does not imply any major strategic shift for FAIR.

Due to its independent research nature, FAIR’s GPU resources have always been limited—only about 5% of those allocated to the GenAI department (and not sharing the same cloud provider). The two divisions also do not share data or code.

For a long time, the distinction between Meta AI and FAIR was unclear. To clarify: the Llama 1 team originated from FAIR, but after its success, the team was spun off into the GenAI product group.

GenAI’s mission is to develop product-level large language models (LLMs), while FAIR focuses on exploring future AI directions and坚持开源。FAIR’s members are independent of GenAI and MSL.

This separation explains why Meta’s Llama-4 release, which caused controversy over leaderboard cheating, was criticized—LeCun was unfairly blamed.

In the wave of generative AI, many prominent scholars, such as Tian Yuan Dong, have shifted to the GenAI division.

The entire visual research team from FAIR has also been transferred to GenAI.

After Zuckerberg announced the establishment of MSL, the original members responsible for Llama will join it, though specific responsibilities are still being defined. Additionally, under the high salaries offered to recruit talent, layoffs are expected in both GenAI and FAIR divisions.

This raises another question: Will internal competition and integration occur as teams are built and research directions are set? The high-level executives recruited by Zuckerberg often include clauses allowing them to recruit their own teams, hinting at potential internal conflicts.

It’s clear that Zuckerberg’s pursuit of superintelligence has just begun, and whether these reforms will succeed remains uncertain.

We may only find out when Meta’s next-generation large models are released in a few months.

Interestingly, The Information recently reported that internal turmoil at Meta is intensifying. Some employees complain that Zuckerberg’s restructuring of the AI research teams is eroding the company culture.

                ^{This has led engineers to question Meta’s AI mission.}

The report states that Tijmen Blankevoort, a researcher about to leave Meta, wrote a lengthy article criticizing the internal culture and organizational issues hindering AI development, including:

• Culture of Fear: Performance reviews and layoffs (5% staff cut in February 2025) create high-pressure environments, where motivation stems from fear of being fired rather than mission-driven work.
• Lack of Vision and Strategic Confusion: The nearly 2,000-strong Llama team is unclear about the company’s goals; resources are scattered across multiple competing projects like Llama, Meta AI Assistant, and social media AI features.
• Talent Drain and High Salaries: Retention of top external talent is low; many elites recruited from OpenAI and Google leave shortly after joining, amid ongoing aggressive hiring.
• Reputation Crisis: The controversy over the Llama 4 leaderboard cheating incident damaged credibility, as the model’s performance was artificially inflated by uploading optimized versions.

What do you think about Meta’s high-paying talent acquisition strategy? Feel free to leave your comments and discuss.