Meta’s aggressive push toward Artificial General Intelligence (AGI) has taken a decisive leap with the hiring of Jason Wei and Hyung Won Chung—two highly respected researchers from OpenAI. Their departure, reportedly marked by the deactivation of their internal OpenAI Slack profiles, signals a strategic coup for Meta, which is rapidly assembling a powerhouse team within its newly formed Superintelligence Lab.
Wei, known for his foundational work on OpenAI’s o3 model and earlier chain-of-thought prompting research at Google, brings deep expertise in large language model alignment and interpretability. Chung, meanwhile, has been instrumental in developing reasoning agents and contributed to the development of OpenAI’s o1 model. Their history of collaboration, dating back to their time at Google, offers Meta not only technical prowess but also a pre-established synergy—something Zuckerberg appears to be intentionally leveraging.
This “acquihire in pairs” strategy mirrors successful talent acquisitions from the past, emphasizing team cohesion over isolated brilliance. According to Wired, Zuckerberg’s intent is to replicate the startup-style effectiveness of small, tight-knit research groups who already share mutual trust and a shared vision—critical factors in advancing fast-moving, experimental research.
The Wei-Chung recruitment is only part of a broader poaching spree. In recent weeks, Meta has also lured researchers like Ruoming Pang (Apple), Trapit Bansal (OpenAI), and a series of luminaries from Google DeepMind including Lucas Beyer, Jack Rae, and Johan Schalkwyk. The Meta Superintelligence Lab is now led by ex-GitHub CEO Nat Friedman and former Scale AI CEO Alexandr Wang, further solidifying its position as a central hub for next-gen AI development.
Meta is pairing these hires with massive infrastructure investments. The upcoming Prometheus AI supercluster (expected by 2026), the Hyperion project, and temporary tent-based data centres reflect the company’s urgency in creating flexible, scalable compute environments. These are likely aimed at training the next generation of frontier models—potentially rivaling OpenAI’s GPT and Google’s Gemini.
While Zuckerberg has denied reports of paying up to $200 million (₹1,600 crore) per researcher, he hasn’t disputed the high-value compensation packages that are clearly part of Meta’s aggressive courtship. Regardless of the figures, the strategy is yielding results: Meta is positioning itself as not just a player, but a serious contender in the AGI race.
By securing researchers who have worked on foundational models and reasoning agents, Meta is gaining the intellectual toolkit required for AGI—language understanding, complex reasoning, and efficient learning. Wei and Chung’s addition could mark a turning point in the company’s AI trajectory, accelerating its capabilities and reshaping the competitive dynamics of global AI leadership.
