Meta CEO Mark Zuckerberg has started to take serious steps to move on to the next stage of artificial intelligence. A new team created within the company aims to develop an artificial intelligence superzet. This objective means the establishment of systems that exceeding it, not only approaching human intelligence. In this process, Zuckerberg has taken the initiative and carries out recruitment himself.
Zuckerberg, WhatsApp through the managers through the private talks shared the candidate lists said. In this chat group “Recruiting Party ,, evaluations are made about carefully selected researchers and engineers. Zuckerberg is also trying to convince potential candidates to dinner at his home in California. These personal contacts are concentrated in the last month.
Meta plans to invest billions of dollars in Scale AI
One of the names expected to be in this new team is Alexandr Wang, the founder of Scale AI. Scale AI draws attention as an important initiative that provides data for the training of artificial intelligence models. It is claimed that Meta plans to invest billions of dollars in this company. In addition to all these, some of the SCALE AI employees are transferred to Meta.
However, it is not yet clear whether these transfers will take part in the new team. On the other hand, it is seen that Meta has made great efforts to transfer experts from other technology giants. Many artificial intelligence experts working in companies such as Google and OpenAI were shared with seven and nine -digit salary offers. It is also among the information that some names that accept these proposals provide an agreement with Meta.
Behind the direct involvement of Mark Zuckerberg is another reason. According to Bloomberg, Meta’s large language model Llama 4 did not meet expectations of the management front caused disappointment. Criticisms from both the sector and the public showed that Llama 4 does not offer the promised performance. These developments accelerated Zuckerberg’s new team directly to manage.
Meta continues to work on a larger Llama 4 model named “Behemoth”. The company claimed that this model would pass its competitors such as GPT-4.5, Claude Sonnet 3.7 and Gemini 2.0 Pro. In spite of everything, uncertainties continue because the release date of this model is postponed. It is predicted that such projects will be restructured with the newly created Superzekâ team.
Meta’s superzekâ target aims to go beyond the target of artificial intelligence at the classical network. This involves not only the thinking of machines as human beings, but also showing much more advanced cognitive skills. In order to achieve this goal, both powerful software and hardware infrastructures need to be developed. Zuckerberg emphasizes that these investments can be financed by Meta’s own resources.
In the world of artificial intelligence, where competition is increasingly hardened, such initiatives are carefully monitored. In this environment where companies compete with billions of dollars of budgets, the road to superzekâ brings both technological and ethical debates. The process, which is carried out with the participation of Meta’s aggressive recruitment policies and direct CEO participation, reveals how serious the company is. On the other hand, how such projects will be met in the public is still a question mark.