Meta layoffs: Chinese scientist Tian Yuandong responds to being "laid off".
Meta is correcting its hiring spree by laying off approximately 600 employees at its Super Intelligence Lab. This layoff does not affect Meta's recent AI talent hires.
According to a Meta employee memo passed to The New York Times, Meta announced on Wednesday that it would lay off approximately 600 employees at its Super Intelligence Lab. The lab, established in July to accelerate research in artificial general intelligence (AGI), is led by Alexandr Wang, the 28-year-old former CEO of data labeling startup Scale AI, and has approximately 3,000 employees.
According to CNBC, this layoff is part of Meta's efforts to further streamline departments and solidify Alexandr Wang's leadership in the company's AI strategy by reducing layers and improving operational efficiency.
In a memo to employees, Alexandr Wang wrote, "By reducing team size, there will be fewer communication steps required for decision-making, and everyone will have greater responsibility, a wider scope, and more influence."
It is understood that the TBD (Total Depth Design) department, led by Alexandr Wang and responsible for developing superintelligence and managing Meta's large language model, was unaffected by the layoffs. Sources familiar with the matter indicated that the TBD department includes several top AI talents recruited by Meta this summer. The fact that these employees, managed by Alexandr Wang, were not laid off highlights that Meta CEO Mark Zuckerberg values these highly paid hires more than existing employees.
Meta executives emphasized that the layoffs do not mean the company will reduce its investment in AI; super intelligence remains one of its core priorities. Two sources familiar with the matter revealed that the layoffs are aimed at addressing organizational bloat. Meta's AI business expanded too rapidly over the past three years, ultimately leading to this situation. The layoffs are intended to help Meta develop AI products more quickly.
For the past three years, Meta has been grappling with the rapid rise of artificial intelligence technology. Following the emergence of ChatGPT in 2022, OpenAI, Google, and Microsoft have been aggressively hiring to develop next-generation AI chatbots and other products. Meta initially developed the Llama open-source AI model, but subsequent progress stalled. In the past 18 months, Meta has restarted its hiring spree, but strategic missteps have ultimately led to frequent product development problems.
After a turbulent first half of this year, Zuckerberg relaunched the AI business, launching a massive hiring spree and offering lucrative salaries to attract top AI talent. In June, he invested $14.3 billion in Scale AI, subsequently bringing top Scale AI talent, including Alexandr Wang, into the Meta superintelligence lab.
Zuckerberg has also spent billions of dollars poaching top researchers from other AI labs and companies such as OpenAI, Google, Microsoft, and Apple, offering some of them salaries of hundreds of millions of dollars.
In August of this year, Zuckerberg split Meta's Superintelligence Lab into four departments. The Foundation for Artificial Intelligence (FAIR) department is responsible for AI research, while the other three departments are responsible for superintelligence development, product development, and infrastructure construction, respectively. Within Meta, research teams like FAIR often compete for computing resources with product-focused teams.
Two sources familiar with the matter said that after the reorganization, FAIR employees eagerly sought to join Alexandr Wang's team. While Wang's core team members are primarily recruited externally from companies like OpenAI and Google, he has recently also recruited dozens of AI researchers with specific expertise from other Meta departments.
It is understood that the layoffs will affect employees in the FAIR, product, and infrastructure departments of the Super Intelligence Lab, while the TBD department, responsible for developing super intelligence, is unaffected. Laid-off employees have received email notifications, and Meta plans to find alternative positions within the company for affected staff. On October 23, Tian Yuandong, a Chinese AI scientist at Meta, a FAIR researcher, and senior manager, posted on social media that he and several members of his team were affected by the layoffs. Tian Yuandong's research focuses on efficient training and inference of large models, representation learning, intelligent decision-making, and optimization.
Subsequently, The Paper contacted Tian Yuandong, who confirmed the news. He revealed that he had been looking for external opportunities before this, and the company leadership was aware of it. Meta's layoffs resulted in him receiving eight months' salary as compensation, and he is now actively seeking external job opportunities.
