Meta has appointed Shengjia Zhao, co-creator of ChatGPT and GPT-4, as the Chief Scientist of its newly established Superintelligence Labs. This announcement was made by CEO Mark Zuckerberg on July 25, 2025. Zhao, previously a senior scientist at OpenAI, played a pivotal role in developing GPT-4 and its variants, including 4.1 and o3, and led synthetic data initiatives at OpenAI.
In her new role, Zhao will collaborate closely with Alexandr Wang, former CEO of Scale AI and now Meta’s Chief AI Officer. Together, they aim to drive Meta’s ambitious efforts in developing artificial general intelligence (AGI) through the Superintelligence Labs. This initiative is part of Meta’s broader strategy to enhance its AI capabilities and compete with other tech giants in the rapidly evolving AI landscape.
Meta’s aggressive recruitment strategy includes significant investments, such as a $14.3 billion stake in Scale AI, and attracting top AI talent from competitors like OpenAI and DeepMind. These moves underscore Meta’s commitment to leading in the next generation of AI development.
🔍 Background on Shengjia Zhao
Role at OpenAI: Before joining Meta, Shengjia Zhao was a senior scientist at OpenAI. She played a pivotal role in developing GPT-4 and its variants, including 4.1 and o3, and led synthetic data initiatives at OpenAI.
🧠 Meta’s Superintelligence Labs Initiative
Leadership: Zhao will collaborate closely with Alexandr Wang, former CEO of Scale AI and now Meta’s Chief AI Officer. Together, they will shape the lab’s research agenda.
Objective: Meta’s Superintelligence Labs aim to unify the company’s advanced AI research efforts and drive progress toward artificial general intelligence (AGI). This initiative is separate from Meta’s FAIR division, led by Yann LeCun.
💰 Meta’s Investment in AI
- Financial Commitment: Meta has made significant investments to bolster its AI capabilities, including a $14.3 billion stake in Scale AI.
- Talent Acquisition: The company has aggressively recruited top AI talent from competitors like OpenAI and DeepMind. Notably, four OpenAI researchers—Shengjia Zhao, Shuchao Bi, Jiahui Yu, and Hongyu Ren—have joined Meta’s Superintelligence team
🌐 Global Talent Strategy
- Diverse Hiring: Meta’s Superintelligence Labs have attracted a diverse group of AI experts, including several from China. Of the 11 publicly disclosed hires, seven are from China, reflecting Meta’s global approach to assembling a top-tier AI research team.-
t the 2025 International Mathematical Olympiad (IMO) held in Queensland, Australia, five teenagers achieved perfect scores of 42 out of 42, outperforming advanced AI models from Google and OpenAI, which scored 35 points each.
This marks the first time AI models have reached gold-level scores at the IMO, showcasing significant advancements in AI’s mathematical capabilities. However, the perfect scores by the human contestants highlight that, despite AI’s progress, human problem-solving skills remain unparalleled in certain complex tasks.
The AI models, including Google’s Gemini chatbot and OpenAI’s experimental reasoning model, solved five out of six problems within the competition’s 4.5-hour time limit. Their solutions were noted for being clear and precise. Nonetheless, the IMO organizers cautioned that the extent of computational resources used by these AI models and the possibility of human assistance could not be verified.
This event underscores the rapid progress of AI in tackling complex mathematical problems while also emphasizing the enduring excellence of human intellect in the field.
