Mark Zuckerberg has a new perspective on the AI race — literally.
The Meta CEO now has a desk among the company’s AI researchers that gives him a front-row seat to the company’s overhaul of its AI efforts. He’s also coding again.
“Mark has actually moved his desk and is seated in the AI lab with Alex Wang and Nat Friedman, and he’s coding all day long,” Meta President Dina Powell McCormick said during an appearance at Semafor’s World Economy Summit in Washington, D.C.
McCormick joked that, “On the one hand, they are sort of like, ‘Oh great, more advice from you Mark, as you’re coding.'”
Striking a more serious tone, she said that Zuckerberg feels a need to be this engaged in the company’s AI efforts.
“I think he feels so strongly that he has to understand it at that level to really, you know, think about how our our model the strongest it can be,” McCormick said.
McCormick’s comments follow previous reporting that has detailed the extent to which Zuckerberg has thrown himself deep into Meta’s AI development as competition with OpenAI’s ChatGPT, Anthropic’s Claude, and Google’s Gemini heats up.
In June, Bloomberg reported that Zuckerberg had “rearranged desks at the company’s Menlo Park headquarters so the new staff will sit near him.” On Monday, the Financial Times reported that Zuckerberg “has spent five to 10 hours a week coding on different AI projects” and reviewing the work of others.
Meta has spent billions building up its new AI team, called Superintelligence Labs, including investing roughly $15 billion in Scale AI and to recruit cofounder Alexandr Wang. Zuckerberg has been closely involved in hiring efforts for the division.
McCormick also mentioned former Github CEO Nat Friedman, who joined Meta in July to help lead the company’s development of AI products.
Externally, Meta has been viewed as falling behind its fellow hyperscalers and frontier model makers in the generative AI race, a perception the MSL team is working to combat with new models.
OpenAI has dominated the AI chatbot market through ChatGPT. While Google has been able to successfully deploy its “full-stack” advantage to keep OpenAI on its toes. Anthropic has received widespread praise for its AI advances in the last 6 months, including Claude Code and Cowork. Anthropic recently announced it would not publicly release its latest model, Claude Mythos, over concerns it is too good at finding cybersecurity vulnerabilities.
There are signs that Meta’s steep investment in building out the MSL team is starting to pay off. Meta’s stock jumped 8% after it released Muse Spark, its first new AI model created by Superintelligence Labs. Zuckerberg called it the division’s “first milestone.”
Read the full article here
