Mark Zuckerberg’s Meta ‘Supertive’ EU boot and keyways from the brand of large tenants

• Establishment of Meta Super Suffer Labs (MSL):
  Meta, LLAMA models, products teams, including its fundamental AI research (fair) team, including teams, including teams, as well as a new unit, integrate all the efforts aimed at developing the next generation of the MSL. MSL, former CEO and Meta’s “Chief AI Officer” and the former Github CEO and the AI ​​investor became together with a former scale CEO and Cofoundan Alexandr Wang.
• Strategic Goal:
  The open goal is to build a personal super environment for everyone, “Zukerberg said. The superstition would be AI that can perform outside the human level in the most cognitive tasks.
• Do not get aggressive talent:
  Meta, the best AI talent from competitors such as Openai braconyer, Googleand anthropic. To do this, according to the comments made by Openai CEO SAM Altman, the unprecedented signature announcement bonuses – offers up to $ 100 million. In memory, Zuckerberg announced the hiring of 11 respected AI researchers from other AI laboratories. Meta has invested in billions of investments to ensure the meta movement of the Corps of the Corps of the AI-centered ventral capital of the EU to bring the EU to Meta.
• Failed Purchase Attempts:
  Hiring Spree is followed by the former Openai General Scientist Ilya Sutskever company’s attempts to obtain the immediate reservoir and confusion AI.
• Llama fight and competitive pressure:
  The last Llama EU model family, called Meta Llama 4, appreciated expectations. The company, designed for Llama 4, was criticized by the Models, which was performed more competitive to appear more competitive than in fact. The release of 4 Behemoth, the largest and in metha, repeatedly postponed the most powerful model produced. Meta, anthropic, Google, Google, Deepseek and Alibaba’nın qwen lost their rivals and “justification” have not yet debuted. This caused an internal dispute about the EU direction of the meter and led to an increase in the ambulance to revive the AI ​​portfolio.
• Holding and penetrating problems:
  Meta has increased the loss of meta to increase the need to attract and maintain world-class AI talent. The company is also facing legal and ethical research on information practices,
• Mass capital obligation:
  Meta invests in meta infrastructure, information centers and special devices to provide a leading role in the AI ​​era.

Those acquired are the full text as the memory of Zuckerberg Fortune correspondent Sharon Goldman:

As the AI ​​speeds up speed, developing super-management. I believe that this will be the beginning of a new era for mankind, and I fully work full to fulfill what is guided for meta. I want to share some details about how we are built to see our AI efforts today: individual super leadership for everyone.

We will call a general organization Meta Supertive Labs (MSL). This includes a new laboratory aimed at developing all our foundations, products and fair teams, as well as our models.

Alexandr Wang, our Chief AI officer and MSL participated in Meta to serve as our leadership. Alex and I have been working together for several years and I consider it the most effective builder of his offspring. He built a rapidly growing company in the development of almost all leading models of industry as the historical significance of the super headline and established a co-founder and CEO’s skalean.

Nat Friedman, EU products and application, was also connected with Alex with Alex to lead the MSL, which began our study. NAT will work with Connor to determine the role. Escaped Gitub MicrosoftAnd the latest recently launched one of the leading AI investment companies. Nat served in our meta consulting group in the last year, so it is already a good feeling of our map and what we need to do.

Today there are several new team members that I am happy to share in several strong new team members or in recent weeks:

• In a hurry – Pionered RL in the chain of O-Series models in Openai and the chain of combination.
• Shuchao bi – GPT-4O sound mode and a creator of O4-Mini. Multimodal post-training in Openai before.
• Huiwen Chang – Muse-to-to-to-architecture of GPT-4O’s image generation joint creator and previously invented maskgit and Google research
• Ji lin – O3 / O4-mini, GPT-4O, GPT-4.1, GPT-4.5, 4O-IMGEGEN and Operating Stack helped.
• Joel Pobar – Result in anthropic. META, HHVM, hack, streaming, redex, performance tools and machine learning is 1 year in learning.
• Jack Rae – Training before training for justification for Gemini and Gemini 2.5. LED Gopher and Chinchilla Early LLM Efforts Depth.
• Hongyyu Ren – GPT-4O, 4O-mini, O1-mini, O3-mini, O3 and a creator of O4-mini. Previously a group of leading for post-training in Openai.
• Johan Shalkwyk – Early contributor of technical lead for the former Google companion, Sesame and Maya.
• Pei sun – After training, coding and justification for twins in Google Deprmind. Previously, the width of Wayon created the last two generations of the perception models.
• Jiahui Yu – O3, a creator of O4-Mini, GPT-4.1 and GPT-4O. Previously, he headed the perception group in Openai and multi-ethnic multimodal in twins.
• Shengjia Zhao – ChatGPT, GPT-4, a creator of all mini models, 4.1 and O3. Previously synthetic data in Openai.

I am pleased with the progress we plan for 4.1 and 4.2. These models increase the number of agents increasing and growing number of agents, which helps us to develop more than 1 billion monthly activity and product and technology in our apps. We continue to build these models.

In parallel, we will start an investigation in our next generation models to be on the next year or more. In the last few months, for these small talented efforts, I am gathered in the top of the past few months to bring together the founding group and gathered a place. We still form this group and we will ask a few people to join this laboratory in AI org.

Meta is unparalleled to lead the world. We have a strong job that supports more than small labs. There are deeper experience and growing products that reach billions of people. We lead the category of AI glasses and clothing that grow in advance and very rapidly. And the structure of our company allows us to act very large in prison and courage. I am optimistic that this is a parallel approach to this new talent and model development, will build us to deliver individual super management promises to everyone.

In the coming weeks, we have even greater people in all levels of this audio, so stay. I am pleased to dive and start working.

***

DISCLAIMER: For this story, the generation of generation used in luck will help a preliminary draft. An editor confirmed the accuracy of the information before publication.