In a recent letter, Meta CEO Mark Zuckerberg unveiled his vision for "personal superintelligence," an AI-driven future where individuals harness artificial intelligence to achieve their personal goals. This revelation signals a significant shift in Meta's approach to AI model releases, as the company delves deeper into the realm of superintelligence.
Zuckerberg emphasized the importance of sharing the benefits of superintelligence globally, while acknowledging the novel safety concerns it raises. He stressed the need for rigorous risk mitigation and careful consideration of what to open source. This stance is a departure from Meta's historical positioning of its Llama family of open models as a key differentiator from competitors like OpenAI, xAI, and Google DeepMind.
Meta's previous goal was to create open AI models that rivaled or surpassed closed models. However, Zuckerberg has left room for maneuver on this commitment, stating that if a qualitative change occurs in the capabilities of AI, Meta may choose not to open source it. This suggests a potential shift in priority, with open source no longer being the default for Meta's cutting-edge AI.
Meta's rivals keep their models closed to maintain control over monetizing their products. Zuckerberg pointed out that Meta's business model, which relies on internet advertising, is not dependent on selling access to AI models. Therefore, releasing Llama models does not undercut Meta's revenue, sustainability, or research investment like it does for closed providers.
However, Meta's stated viewpoint on open models was before the company started feeling the pressure of falling behind competitors and became obsessed with beating OpenAI's GPT-4 model while developing Llama 3. In June 2025, Meta began its public AGI sprint by investing $14.3 billion in Scale AI, acquiring Scale’s founder and CEO, and restructuring its AI efforts under a new unit called Meta Superintelligence Labs. The company has spent billions of dollars to acquire researchers and engineers from top AI firms and build new data centers.
Recent reports indicate that this investment has led Meta to pause testing on its latest Llama model, Behemoth, and instead focus on developing a closed model. With Zuckerberg's mission to introduce "personal superintelligence" to the world, Meta's AI monetization strategy is taking shape. It's clear that Meta plans to deliver "personal superintelligence" through its own products like augmented reality glasses and virtual reality headsets.
Zuckerberg envisions personal devices like glasses that understand our context by seeing, hearing, and interacting with us throughout the day becoming our primary computing devices. When asked about Meta potentially keeping its most advanced models closed, a Meta spokesperson said the company remains committed to open source AI and expects to train closed source models in the future.
"Our position on open source AI is unchanged," the spokesperson said. "We plan to continue releasing leading open source models. We haven't released everything we've developed historically and expect to continue training a mix of open and closed models going forward."