LeCun responds to Zhao Shengjia's appointment as "Chief Scientist".

Artificial intelligence is a young cause, and it is also a cause for young people.

A post - 90s Chinese scientist, Shengjia Zhao (an alumnus of Tsinghua University and a core member of ChatGPT to o3) has taken up a new position as the Chief Scientist at Meta Superintelligence Labs (MSL).

Meanwhile, the topic of whether Yann LeCun, a Turing Award winner born in the 1960s, another Chief Scientist at Meta, will stay or leave has sparked intense public discussion.

In response, LeCun has publicly stated: He still serves as the Chief Scientist at Meta FAIR (Fundamental AI Research) and looks forward to collaborating with Shengjia.

The new architecture of Meta AI is thus beginning to take shape.

From Meta to Meta FAIR

As LeCun responded:

As the Chief Scientist at FAIR, I have always been focused on long - term research in artificial intelligence and building the next - generation AI paradigm. My responsibilities and FAIR's mission remain unchanged.

Mark Zuckerberg and Alexandr Wang, the current leader of Meta Superintelligence Labs, have also confirmed this on their respective social media platforms:

LeCun's role remains unchanged, and he will continue to serve as the Chief Scientist at FAIR.

However, netizens don't seem to be convinced. People are curious about the real situation of this Turing Award winner at Meta.

This has to start with the positioning of FAIR within Meta.

Some netizens said that FAIR is still an independent department, and LeCun doesn't need to report to the post - 90s leader.

But in Zuckerberg's latest memo about the "Superintelligence Lab", FAIR has been incorporated into Meta Superintelligence Labs.

We will name our entire organization "Meta Superintelligence Labs" (MSL). It encompasses all our foundations, products, and the FAIR team, as well as a new lab focused on developing the next - generation models.

Before this, FAIR had also undergone several rounds of strategic reorganizations:

On December 9, 2013, FAIR was founded by LeCun, who served as its head. As the core institution for Meta AI research, it has made significant breakthroughs in areas such as GANs, computer vision, natural language processing, and self - supervised learning. It promoted the birth and popularization of the PyTorch framework and released high - impact open - source results such as Segment Anything, SeamlessM4T, and Toolformer.

In 2022, FAIR was integrated into Meta Reality Labs and began to gradually transform from a pure academic institution into a research and development institution for services and products (such as the metaverse), thus bridging the gap between basic research and practical applications.

With the explosion of large language models, FAIR has been regarded as a sister department parallel to the team responsible for developing GenAI applications (responsible for developing the Llama model), focusing on basic research. It mainly conducts research in areas such as vision, audio, video, and world models (such as the classic JEPA world model).

At the beginning of 2024, Zuckerberg said that the FAIR team would closely cooperate with the GenAI team to work towards the realization of AGI. FAIR will focus its long - term research agenda on world models and embodied intelligence, thus continuing Meta's focus on areas such as the metaverse, embodied intelligence, and AR.

However, with the major changes in the Llama team this year (78% of the original research team members left Meta) and the departure of Joëlle Pineau, a core figure at FAIR, the role of the FAIR team has changed subtly.

In May this year, Meta split its AI organization into two teams, one focusing on consumer products and the other on AGI research.

It can be said that although the name of the FAIR team still exists, independent product and research teams have been separated, and the role of the FAIR team has started to become marginalized.

After Zuckerberg established Meta Super Artificial Intelligence Lab in July to take overall control, the status of FAIR and LeCun is self - evident.

As the former Chief Scientist at Meta, the outside world often blames LeCun for Meta's lag and setbacks in the field of large language models.

However, some netizens said that everyone can see LeCun's contributions to Meta (LeCun himself also reposted something, you can think about it).

Since LeCun is still at Meta, his contributions to Meta will have to be evaluated by history.

But judging from Meta's recent actions, the more extravagantly Zuckerberg spends, the gloomier LeCun's face becomes.

Shengjia Zhao, the New Chief Scientist

As we mentioned at the beginning, Shengjia Zhao has been appointed as the Chief Scientist at Meta Super Artificial Intelligence Lab and will closely collaborate with Zuckerberg and Alexandr Wang, the head of the AI team, to lead Meta's AI strategy.

After graduating from Tsinghua University with a bachelor's degree in computer science, Shengjia went to Stanford University for further studies in 2016 and obtained a doctorate in computer science in 2022.

During this period, he published several academic papers and won honors such as the ICLR Outstanding Paper Award, Google Excellence Scholarship, Qualcomm Innovation Fellowship, and JPMorgan PhD Fellowship.

After graduating with his doctorate, Shengjia Zhao joined OpenAI as a researcher and became a core member of the ChatGPT, GPT - 4, and several Mini models (such as GPT - 4.1, o3) series.

In addition, he is also a pioneer in the field of chain - of - thought reasoning models and is responsible for leading the synthetic data generation strategy for training large language models.

His work has not only promoted the development of reasoning models but also brought about a technological paradigm shift in the industry.

What kind of sparks will this young man create with Meta? Let's wait and see!

Reference Links

[1]https://www.businessinsider.com/yann-lecun-meta-chief-ai-scientist-shengjia-zhao-fair-superintelligence-2025-7

[2]https://x.com/OpenSui/status/1948895981308969443

[3]https://www.linkedin.com/in/yann-lecun/recent-activity/all/

[4]https://player.me/metas-agi-ambitions-fair-aims-to-rival-open-source-agi/?utm_source=chatgpt.com#Meta_FAIRs_Integration_and_Impact

This article is from the WeChat official account "QbitAI" (ID: QbitAI), author: henry, published by 36Kr with authorization.