Previously, Zhao worked for Openai, where he helped develop several historical AI models, such as GPT-4, GPT-4.1 and GPT-O3. His research focuses on the dynamics of training and large -scale training capacities. Zhao has introduced new paradigms for scaling and is considered one of the most desirable candidates on the AI talent market today.
The hiring by Meta de Zhao signals his progress towards general artificial intelligence (AG). The company has increased its investments in the fundamental research of AI and actively recruits better talents from organizations such as Openai, Google Deepmind and Anthropic. also significantly increased its AI team, with reports that it has offered several key researchers agreements worth hundreds of millions of dollars.
After a silent reception following the release of Llama 4, generally considered to be lagging behind the GPT-4 of Openai, this appointment opens the way to the establishment of the new superintendent laboratory.
The new laboratory is part of an aggressive program to produce ready -made models for future with long -term capacities and open research paradigms. This competition has intensified internationally, with Zhao among the many high -level researchers who recently left Openai for Meta.
Recently, the CEO of Openai expressed his concerns that movements can bring a cultural change. Demis Hassabis de Deepmind, among others, said that it was a meta movement “catching up” rather than a real innovation.
The Meta Superintelligence LAB was officially launched on June 30, 2025. Its main objective is to integrate the advanced research of models, the infrastructure and future AI systems of the company in a unified entity.
Alexandr Wang will direct the laboratory strategy, while Zhao will supervise the scientific agenda. In addition, the former CEO of GitHub, Nat Friedman, joined Meta to help develop consumer products powered by AI in collaboration with the laboratory.
Read also:
