Mark Zuckerberg On Monday said that he is making Meta Suptereligence Labs, leading some recent rented rented of his company, including Scale AI Ex-CEO Alexandra Wang and former Github CEO Nat Freedman.

Zuckerberg said that the new AI Superintendent Unit, MSL, various teams of the company will work on foundation models such as open-source lama software, products and foundation models like fundamental artificial intelligence research. ProjectsAccording to an internal memorandum received by CNBC.

Bloomberg First Informed About the new unit.

Meta co-founder and CEO is on one AI hiring blitz Since he faces a terrible competition with rivals like openi and like GoogleEarlier in June, the company said that it will happen hire Wang, now Meta’s Chief AI officer, and some of his colleagues as part of an investment of $ 14.3 billion in Scale AI.

Meta also hired Freedman and his business partner, Daniel Gros, who was the CEO of Saffy Superintendent, OpenEII co-founder Ilya Satsakewar, first made by CNBC AI Startup made by CNBC InformedMeta tried to buy safe superintendent, but was reprimanded by Sutaskewar.

Openai CEO Sam Altman Said Recently in a podcast that meta was recruiting AI Researcher From his company, offered to sign a bonus as high as high as high.

Andrew Bosworth, head of Meta Technology, told CNBC’s “Closing Bell overtime” in an interview on 20 June that OpenEEEAI was combating the proposals of Meta.

Bosworth said, “The market is set a rate for talent levels here which is really incredible and is unprecedented in my 20 -year career as a technology executive.”

A full internal memorandum of Zuckerberg has been released here on Monday:

As AI progresses the pace of progress, developing superintendent is coming in vision. I believe it will be the beginning of a new era for humanity, and I am completely committed to doing that takes way to meta. Today I want to share some details about how we are organizing our AI efforts to build our vision: Personal Superintending for all.

We are going to call our overall organization Meta Superintendent Laboratories (MSL). It includes all our foundations, products and fair teams, as well as a new laboratory that focuses on developing the next generation of our model.

Alexandra Wang has joined the meta to serve as our Chief AI officer and lead MSL. Alex and I have worked together for many years, and I consider him the most influential founder of my generation. He has a clear understanding of the historical significance of superintending, and as co-founder and CEO, he created the scalye in a fast growing company involved in the development of almost all major models across the industry.

Nat Freedman joined the meta to participate with Alex to lead AIIMSL, led our work on AI products and implemented research. The nut will work with Connor to carry forward its role. He ran a github in Microsoft, and recently run one of the major AI investment firms. NAT has served in our meta advisory group for the previous year, so they already have a good understanding of our roadmap and what we need to do.

We also have many strong new team members who are joining today or who have joined the last few weeks that I am also excited to share:

• Trapit bansal -RL pledged RL on the O-Series model idea and co-producers’ series in OpneaIA.
• Shuchao bye -GPT-4O Voice Mode and Co-Madeer of O4-Min. First led the multimodal post-training in Openai.
• Huven Chang -GPT-4O’s image generation co-producer, and first Google Research invented masculine and museum text-to-image architecture
• Zee lin -O3/O4-Mini, GPT-4O, GPT-4.1, GPT-4.5, 4O-Emagegen, and help in the construction of operator Reasoning Stack.
• Joel Pobar – Estimate in anthropic. First in meta for 11 years on Hhvm, hack, flow, redex, performance tooling and machine learning.
• Jack Rai -Pre-training take lead for Mithun and argument for Gemini 2.5. LLM’s efforts in Deepmind led.
• Hangyu Rain -GPT-4o, 4O-Mini, O1-Mini, O3-Mini, O3 and Co-producers of O4-Min. First led a group after training in Openai.
• Johanna sculpture – East Google Fellow, initial contributor to mole, and technical leadership for Maya.
• PE Sun -Training, coding and logic for Gemini in Google Deepmind. The first two generations of the model made the last two generations.
• Jiahui you -O3, O4-Mini, GPT-4.1 and GPT-4o co-producers. First led the Dharna Team in Openai, and co-eaten the multimodal in Mithun.
• Shringzia -Chatgpt co-producer, GPT-4, all mini models, 4.1 and O3. LED synthetic data in first Openai.

I am excited about the progress we have planned for Lama 4.1 and 4.2. These models power Meta AI, which is used by more than 1 billion monthly activities in our apps and increasing number of agents in meta that help improve our products and technology. We are committed to continue the manufacture of these models.

In parallel, we are going to start research on our next generation model to reach the frontier next year either. I have spent the last few months spent spending meta, other AI labs, and promising startups to keep the founder group together for this short talent-bursting attempt. We are still forming this group and we will also ask many people to join this laboratory in AI Org.

The meta is specificly deployed to give superintendent to the world. We have a strong business that supports calculating significantly more than small laboratories. We have a deep experience of manufacturing and growing products that reach billions of people. We are leading and leading AI glass and wareballs category that is growing very quickly. And the structure of our company allows us to move forward with much confidence and boldness. I am optimistic that this new influx of talent and parallel approach for the development of models will establish us to deliver us on the promise of personal superintendent.

In the coming weeks, we have more great people at all levels to be involved in this effort, so stay. I am excited to dive and work.

Watch, Meta’s AI talent spent race,