Has the 'Windfall' for AI Talent Recruitment Ended? Meta Completes Superintelligence Development Organization

Wang CAIO: "Superintelligence Is Near"

Meta has finalized its artificial intelligence (AI) organizational restructuring after months of talent recruitment. It is reported that the company is now composed of four organizations: large language models (LLM), superintelligence, consumer products, and infrastructure.

According to Bloomberg News, on the 19th (local time), Meta announced this organizational restructuring in an internal message. Previously, until the announcement of the new Superintelligence Lab AI organization in June, Meta operated three teams: the FAIR (fundamental AI research) basic research organization, the AGI Foundation developing LLMs such as LLaMA, and an AI product team. The structure was reorganized into four groups by creating a separate team focusing on infrastructure such as data centers and power.

Meta’s Chief Artificial Intelligence Officer (CAIO), Alexandr Wang, leads the TBD (to be determined) Lab, which focuses on LLM development. The TBD Lab is expected to handle the development of short-term AI models ranging from a few months to one year.

Long-term research is assigned to the FAIR team. Rob Fergus, who co-founded the FAIR team with Meta’s Chief Scientist Yann LeCun in 2013, manages the team. After working at Google DeepMind for five years, he returned to Meta around this spring. The team’s main tasks include mid-to-long-term research such as superintelligence. Recently, Nat Friedman, former CEO of GitHub recruited by Meta, leads the product team responsible for consumer products like Meta Glass. In addition, Aparna Ramani, Vice President overseeing Meta’s Data and Developer Infrastructure division, leads the MSL Infrastructure Team.

Wang CAIO announced the restructuring, saying, "Superintelligence is approaching, and to take this seriously, we must organize around the core areas of research, products, and infrastructure." With this restructuring, Meta’s series of talent recruitment and internal adjustments appears to be concluding. Recently, Meta restructured its AI organizations four times in the past six months. It is reported that there is internal criticism regarding excessive spending on AI talent during this process. Some speculate that, as OpenAI's latest ChatGPT model, GPT-5, faces performance controversy, Meta will focus even more on superintelligence development to secure AI leadership. The development of AI with superintelligence is called the hidden card that could deflate the 'AI bubble'.

Silicon Valley = Inyeob Kim, Correspondent inside@hankyung.com