Meta splits AI group into four parts in pursuit of superintelligence

By Riley Griffin

 

Meta Platforms Inc. is splitting its newly formed artificial intelligence group into four distinct teams and reassigning many of the company’s existing AI employees, an attempt to better capitalize on billions of dollars’ worth of recently acquired talent. 

 

The new structure is meant to “accelerate” the company’s pursuit of so-called superintelligence, according to an internal memo sent Tuesday by Alexandr Wang, the former Scale AI chief executive officer who recently joined Meta as chief AI officer. 

 

“Superintelligence is coming, and in order to take it seriously, we need to organize around the key areas that will be critical to reach it — research, product and infra,” Wang wrote in the memo, which was reviewed by Bloomberg News.

 

The group, known as Meta Superintelligence Labs, or MSL, will now have four parts:

 

TBD Lab, led by Wang, which will oversee Meta’s large language models, including the Llama tools that underpin its AI assistant.

 

FAIR, an internal AI research lab that’s existed within the company for more than a decade. The team, whose name stands for fundamental AI research, is focused on longer-term projects.

 

Products and Applied Research, a team led by former GitHub CEO Nat Friedman, which will take those models and research and put them into consumer products.

 

MSL Infra, which will focus on the expensive infrastructure needed to support Meta’s AI ambitions.

 

No layoffs were part of Tuesday’s reorganization, according to a person familiar with the matter, who asked not to be identified because the deliberations are private. Details of the new structure were first reported by the Information.

 

Meta is hoping to stabilize its AI efforts after months spent poaching dozens of top AI researchers from competitors with lofty pay packages, many reaching hundreds of millions in total compensation. CEO Mark Zuckerberg has said the company’s goal is to achieve superintelligence, or AI technology that can complete tasks even better than humans, and he expects to spend hundreds of billions of dollars on the talent and infrastructure needed to get there. 

 

But Meta’s AI leadership has faced several shake-ups in the past few years, including multiple changes this year alone as the company has raced to keep pace with rivals like OpenAI and Google. Before announcing MSL in June, the social media giant had three primary AI teams — FAIR, an AI products group, and the AGI foundations team, which focused on generative AI products and research. 

 

The AGI foundations group is being dissolved, and leaders Ahmad Al-Dahle and Amir Frenkel are now “focusing on strategic MSL initiatives” and reporting to Wang, according to the memo. The former head of the AI products group, Connor Hayes, was already reassigned to run Threads, Meta’s rival product to Elon Musk’s X. 

 

As part of Tuesday’s reorganization, Aparna Ramani, a Meta vice president charged with leading the company’s AI, data and developer infrastructure units, will run the MSL Infra team, according to the memo. Robert Fergus will continue to lead FAIR, an organization he co-founded in 2014. He had previously left the group and spent several years at Alphabet Inc.’s DeepMind before returning to run FAIR this spring.

 

Loredana Crisan, who previously led the company’s Messenger app and moved to the company’s generative AI group in February, is departing Meta for Figma Inc., according to a person familiar with the move.