Source: Fortune
Summary
Meta launched Muse, a personal AI assistant, which became the most downloaded free app on Apple’s App Store within two weeks. The app is part of Meta’s broader AI push under Alex Wang, a Scale AI co-founder hired to lead the company’s Superintelligence Labs. Meta expects 2026 capital expenditures of $130 billion to $145 billion, nearly double last year’s spending. Wang, who has no formal AI training, was chosen to lead Meta’s AI efforts despite skepticism from some internal researchers. The company has invested heavily in recruiting top talent and developing new AI models, including Muse Wonder, which claims to be state-of-the-art.
Our Reading
The numbers tell one story.
Meta is pouring billions into AI, with a focus on personal assistants and frontier models.
Zuckerberg and Wang are betting on a small, elite team to drive breakthroughs.
Investors are skeptical, but the company is doubling down on its AI ambitions.
Wang’s unconventional background raises questions about the path to success.
Author: Evan Null
The 44-minute call and the memo
Zuckerberg and Wang’s partnership began with a 44-minute call and a detailed memo from Wang. The memo outlined strategies for rebuilding Meta’s AI team and improving model performance. Wang emphasized talent density and the need to start from scratch, which aligned with Zuckerberg’s concerns about the Llama 4 model’s reception. The memo convinced Zuckerberg to make Wang the head of AI, despite his lack of formal AI training.
Zuckerberg had been impressed by Wang for years, calling him the sharpest entrepreneur of his generation. However, the appointment raised eyebrows, as Wang had no experience training frontier models. Despite this, Zuckerberg insisted Wang was the right choice, clearing his calendar for months to support the AI initiative.
The memo marked the beginning of a major shift in Meta’s AI strategy. It led to a recruitment push that brought in top talent and reshaped the company’s approach to AI development. Wang’s leadership focused on building a small, elite team capable of delivering high-impact results.
The partnership between Zuckerberg and Wang has been intense, with both working long hours and making significant sacrifices. The results have included the launch of Muse Spark and Muse Wonder, which are positioned as cutting-edge AI models for personal and professional use.
While the models have shown promise, independent evaluations suggest they are competitive but not yet the best in the field. The success of Meta’s AI strategy will depend on whether these models can gain widespread adoption and deliver on the company’s ambitious goals.
The ‘supernova’ recruiting push
Zuckerberg described the recruitment effort as a “supernova,” a fast-paced and intense campaign to build a top-tier AI team. The push involved hiring from within Meta and poaching talent from rival companies. Meta paid $14.3 billion for a stake in Scale AI to bring Wang on board and offered large compensation packages to attract top researchers.
The team Wang assembled inside Superintelligence Labs was small, with fewer than 200 members, but highly specialized. He focused on creating a culture of “insecure overachievers” who consistently exceeded expectations. This approach aimed to drive innovation and deliver results quickly.
The recruitment effort was part of a broader strategy to rebuild Meta’s AI capabilities. It involved rethinking the company’s approach to model development and prioritizing speed and efficiency. The team worked closely with infrastructure and product teams to ensure the models could be deployed effectively.
Zuckerberg and Wang were deeply involved in the process, with both working long hours and making significant personal sacrifices. The effort paid off with the release of Muse Spark and subsequent updates, leading to the launch of Muse Wonder, which claims to be a state-of-the-art AI model.
Despite the challenges, the team’s work has positioned Meta as a major player in the AI space. The success of the recruitment push will be critical in determining whether Meta can maintain its competitive edge in the rapidly evolving AI market.
The ‘supernova’ recruiting push
Zuckerberg described the recruitment effort as a “supernova,” a fast-paced and intense campaign to build a top-tier AI team. The push involved hiring from within Meta and poaching talent from rival companies. Meta paid $14.3 billion for a stake in Scale AI to bring Wang on board and offered large compensation packages to attract top researchers.
The team Wang assembled inside Superintelligence Labs was small, with fewer than 200 members, but highly specialized. He focused on creating a culture of “insecure overachievers” who consistently exceeded expectations. This approach aimed to drive innovation and deliver results quickly.
The recruitment effort was part of a broader strategy to rebuild Meta’s AI capabilities. It involved rethinking the company’s approach to model development and prioritizing speed and efficiency. The team worked closely with infrastructure and product teams to ensure the models could be deployed effectively.
Zuckerberg and Wang were deeply involved in the process, with both working long hours and making significant personal sacrifices. The effort paid off with the release of Muse Spark and subsequent updates, leading to the launch of Muse Wonder, which claims to be a state-of-the-art AI model.
Despite the challenges, the team’s work has positioned Meta as a major player in the AI space. The success of the recruitment push will be critical in determining whether Meta can maintain its competitive edge in the rapidly evolving AI market.
The ‘supernova’ recruiting push
Zuckerberg described the recruitment effort as a “supernova,” a fast-paced and intense campaign to build a top-tier AI team. The push involved hiring from within Meta and poaching talent from rival companies. Meta paid $14.3 billion for a stake in Scale AI to bring Wang on board and offered large compensation packages to attract top researchers.
The team Wang assembled inside Superintelligence Labs was small, with fewer than 200 members, but highly specialized. He focused on creating a culture of “insecure overachievers” who consistently exceeded expectations. This approach aimed to drive innovation and deliver results quickly.
The recruitment effort was part of a broader strategy to rebuild Meta’s AI capabilities. It involved rethinking the company’s approach to model development and prioritizing speed and efficiency. The team worked closely with infrastructure and product teams to ensure the models could be deployed effectively.
Zuckerberg and Wang were deeply involved in the process, with both working long hours and making significant personal sacrifices. The effort paid off with the release of Muse Spark and subsequent updates, leading to the launch of Muse Wonder, which claims to be a state-of-the-art AI model.
Despite the challenges, the team’s work has positioned Meta as a major player in the AI space. The success of the recruitment push will be critical in determining whether Meta can maintain its competitive edge in the rapidly evolving AI market.
