Mark Zuckerberg Meta creates superintelligence labs. Read the note

Mark Zuckerberg said it is emerging on Monday Meta Superinteligence laboratories, aimed at his company’s latest recruitment, including AI AI AI ex-CEO Alexandr Wang and GitHub CEO NAT Friedman.

Zuckerberg said the new AI superintelligence unit, MSL, will perform various groups for the company’s company, such as open source Llama software, products and basic artificial intelligence research ProjectsAccording to an internal note obtained by CNBC.

Bloomberg first new about the new unit.

The Creator of Meta has been the founder and the General Manager AI Procurement Blitz While facing hard competition of rivals like Openai Google. Earlier in June, the company said hire Wang, now in charge of Meta’s Aiko and his colleagues within an investment of $ 14.3 billion, on a scale.

Meta also hired Friedman and his business couple, Daniel Gross, Director General of Surpek Superinteligencies, created by the AI ​​Startup created by CNBC new. He tried to buy a safe meta superintelligency, but Sutskever finished.

Openai CEO Sam Altman celurbill Meta was hiring in a recent podcast AI Researchers From his company, offering bonuses to sign up $ 100 million.

Andrew Bosworth’s Meta technology headed CNBC told the “Closing Bell Schedule” on June 20 to fix the offerings.

“The market is setting a level of talent here, my 20-year career is a kind of horrible and unpredictable way, as a technology executive,” Bosworth said.

Here is the full internal notice of Zuckerberg Monday released:

As the rhythm of advances accelerates, the development of superintelligence is sight. I think it will be the beginning of a new era for humanity, and I am fully committed to getting the way Meta needs. I want to share some details that we are organizing our AI efforts to achieve our vision today: personal superintelligence for everyone.

We will call the Meta Superintelligence Laboratories of our general organization (MSL). This includes all our foundations, products and fair groups, as well as a new laboratory focused on developing our next generation of models.

Alexandr Wang has introduced our Meta Main Officer and MSL lead. Alex and I have worked together for several years, and I think it is the most spectacular creator of his generation. He has a strong sense of historical importance of superintelligence and as a general creator and director, he participated in the development of almost important patterns in the industry that built the beggar.

Nat Friedman has also included a meta to assume MSL to join Alex, our work AI products and applied research. Nat will work with Connor to define his mission. He made GitHub at Microsoft, and has recently directed one of the most important AI investment companies. Nat has served our Meta Advice Group in the last year, so it has a good sense of our path map and what we need to do.

Today, we also have many powerful new team members in collaboration in collaboration in the weeks I am joining.

• TRAPIT BANSAL – Pioneering of the thought chain and creative of the O-Series models in Openai.
• Shuchao two – GPT-4O Voice mode and O4-Mini Creator. Before Multimodal LED Post Training in Openai.
• Huiwen Chang – Creator of GPT-4O image creation, and previously invented the architecture of mask and museum images in Google Research
• Ji Lin – O3 / O4-Mini, GPT-4O, GPT-4.1, GPT-4.5, 4o-Imagegen and operator.
• Joel Pobar – Anthropic inference. Before Meta 11 years old for 11 years HHVM, hack, flow, reduction, performance tooling and learning machine.
• Jack Rae – Technical training for Gemini and Gemini 2.5. LED Gopher and Chinchilla LLM initial efforts in Deepmind.
• Hongyu Ren – GPT-4o, 4o-mini, O1-Mini, O3-Mini, O3 and O4-Mini. Post-training team in Openai in advance.
• Johan Schalkwyk – Oral Google friend, Early Assistant of Sesame and Mayaka’s technical lead.
• Pei sun – Training, encoding and reasoning at Google Deepmind’s Gemini. In advance Waymo’s perception model created the last two generations.
• Jiahui Yu – O3, O4-Mini, GPT-4.1 and GPT-4O creator. He previously directed the Perception Group in Openai, and multimodal multimodal multimodal geminic.
• Shengjia zhao – Dog, GPT-4, all mini models, founder 4.1 and O3. Previously synthetic LED data in Openai.

I am excited about the progress we have planned by Llama 4.1 and 4.2. These models are used in Power Meta AI, more than a million assets in our applications are used more than a million assets and will help improve our products and technology throughout the meta. We are committed to continuing to build these models.

In parallel, next year or that is why we will start researching our next generation models. In recent months I have found the higher tops of Meta, AI, among some laboratories, and the expectations of putting together a group created for this intensive talent effort. We are still completing this team and will ask several people across AI Org to enter this laboratory.

The meta is very positioned to give the world superintelligence. We have a strong business that supports the construction of more pronounced computing than smaller laboratory. We have a deeper experience to grow products that get thousands of people. We are a very fast growing AI Glasses and the leader of the Wearables category. And our company’s structures allow us to think and move more courage. I am optimistic to provide this parallel development of talent and parallel approach to the development of the model development of the model for everyone’s personal superintelligency.

We also have a lot of people who make this effort in the coming weeks, so we have been attentive. I am excited to immerse me in and go to work.

Watch: Meta’s Ai Talent Spend Spree.