Meta raids Google DeepMind and Scale AI for its all-star superintelligence team

In a significant move to enhance its artificial intelligence capabilities, Meta has aggressively recruited talent from Google's DeepMind division and Scale AI, securing at least ten researchers since July, as per LinkedIn insights reviewed by Business Insider. The new additions to Meta's superintelligence team include experts who played crucial roles in developing some of Google's advanced AI models. Among them is Yuanzhong Xu, a seasoned researcher whose contributions were pivotal in the creation of LaMDA, Google’s conversational AI model, and PaLM 2, which underpins the Gemini generative AI models. Xu's presence at Meta hints at the company's ambition to further its AI research and development under its Superintelligence Labs, where he is now listed as working on a project dubbed "TBD." In addition to Xu, Meta has also brought in Tong He, recognized for his contributions that helped Google secure gold at the International Mathematical Olympiad, along with several other notable researchers from DeepMind. Mingyang Zhang, another recent hire, has extensive experience in optimizing large language models for effective information retrieval. Xinyun Chen, also from DeepMind, specializes in AI systems designed for autonomous code generation and logical reasoning. Meta's recruitment strategy appears to be a direct response to the fierce competition for AI talent among major tech firms. The company has made headlines for its lucrative offers and strategic poaching of personnel from rivals such as OpenAI. Alongside its new hires from DeepMind, Meta has also invested heavily in Scale AI, acquiring nearly half of the company for $14 billion earlier this year. This investment aims to bolster Meta’s safety and evaluation efforts, crucial for ensuring that AI systems align with human values and do not propagate harmful content. Scale AI's SEAL (Safety, Evaluations, and Alignment Lab) team has been integral in overseeing AI's ethical guidance, and Meta has recruited several of its leading researchers, including Ziwen Han and Nathaniel Li, who co-authored a rigorous test for AI models known as Humanity's Last Exam, aimed at challenging AI capabilities. Despite the intense talent acquisition race, Google’s CEO Sundar Pichai has downplayed concerns regarding employee retention, suggesting that the company's workforce remains stable. As Meta continues to expand its AI research team, the tech landscape is watching closely to see how these strategic hires will influence the future of AI development and deployment.