Meta CEO Mark Zuckerberg unveiled an ambitious manifesto on Monday, detailing a vision for “superintelligence” that challenges the centralized power structures of industry rivals like OpenAI, Anthropic, and Google.
Democratizing Superintelligence
Zuckerberg’s core argument centers on the belief that AI should not be concentrated in the hands of a few corporations. Instead, he advocates for a model that empowers individuals. “Rather than centralizing superintelligence, we should distribute it widely and give every person the ability to direct it,” Zuckerberg wrote. He envisions a new era of personal empowerment, where users leverage these tools to fulfill their potential and improve their daily lives.
To achieve this, Meta plans to provide free AI access to billions of users. For those requiring higher compute power, the company intends to implement a “dynamic auction mechanism.” Zuckerberg claims this system will ensure users pay the lowest possible price while optimizing compute capacity based on collective demand.
Introducing the Muse Glimmer Model
As a tangible step toward this vision, Meta launched its Muse Glimmer open-weight model on Monday. Designed for “always-on” local agent workflows, the model allows users to download and run it directly on their own PCs or Macs, facilitating personal agentic use cases without relying on cloud-based dependency.
Challenging AI Safety Narratives
Zuckerberg took a sharp stance against industry leaders who frame AI as a catastrophic threat to humanity. He dismissed the “doomer” narrative, stating, “The notion that AI is so dangerous that the only safe path is an extreme concentration of power seems inherently problematic.”
Defending the Practice of Distillation
The manifesto also defends the controversial practice of “distillation”—the process of using established AI models to train new ones. While critics view this as an unethical shortcut that exploits the R&D investments of others, Zuckerberg views it as a strategic necessity. “Some have tried to frame distillation as harmful, but I think it is important to protect the principle that you can learn from anything you can observe,” he argued, framing it as a vital component for the United States to maintain its lead in the global AI race.
Strategic Pivot and Financial Pressure
These announcements arrive at a critical juncture for Meta. The company has been pouring billions into its AI infrastructure to regain momentum following the underwhelming performance of its Llama 4 models last year. In response, Meta established its “Meta Superintelligence Labs” and enlisted Scale AI founder Alexandr Wang to lead its renewed efforts.
Beyond individual users, Zuckerberg is eyeing the enterprise market and exploring the possibility of renting out excess computing capacity to generate new revenue streams. Despite these strategic shifts, investor sentiment remains cautious; Meta’s stock has declined over 20% in the last 12 months and is down approximately 8% year-to-date.