Meta (NASDAQ: META) has released Muse Glimmer, an open-weight AI model, alongside a 6,500-word essay from CEO Mark Zuckerberg in which he argued that the greatest danger posed by advanced AI is concentrated control by a single government or company — not the technology itself.

Muse Glimmer is a dense model with 30 billion parameters designed to run on a laptop or a single consumer GPU. Meta describes it as "one of the highest-performing models of its size." Because the model weights are open, anyone can download, customize, and run it on their own hardware. The release extends Meta's broader push into open-weight AI, following its Llama family of models, which have become among the most widely downloaded open foundation models and have positioned Meta as the most prominent corporate advocate for open AI access — a stance that sets it apart from competitors like OpenAI, Google, and Anthropic, which have generally kept their most advanced models proprietary.

Meta built Muse Glimmer through distillation, training the smaller model on outputs generated by its larger Muse Spark system so that the compact version inherits capabilities from the larger "teacher" model. Weights for a version of Muse Spark 1.2 will be released in the coming weeks, according to the company. Both models are described as the largest American open-source models to date.

Meta had previously paused its open-weight model releases, a decision Zuckerberg attributed to a strategy review. The company has now resumed those releases under the auspices of Meta Superintelligence Labs. The pause and return come as governments are actively developing AI oversight frameworks, including the European Union's AI Act, which imposes obligations on providers of general-purpose AI models, and a U.S. executive order on AI safety that established reporting requirements for certain frontier models.

Zuckerberg's Essay on AI Centralization

In his essay, titled "The Future is for Everyone: The Path to a Positive AI Future," Zuckerberg pushed back against the doom-laden tone that characterizes much of the current AI discourse.

"I'm personally more worried about centralization than I am about any of the specific risks others are talking about," he said in an interview.

Zuckerberg argued that concentrating superintelligence in a few hands "will naturally lead to outcomes that are less favorable for everyone else," adding that "there is no such thing as a singular benevolent superintelligence."

He proposed an alternative vision built on three principles: individual empowerment as the source of prosperity, invention rather than automation as AI's primary purpose, and a balance of power as the foundation for safety.

On geopolitics, Zuckerberg warned that U.S. regulations delaying American model releases by even a month could cost the country its competitive lead while foreign models continue to advance. The framing reflects a broader industry concern that restrictive domestic policy could cede ground to open models developed in China, where entities including DeepSeek and Alibaba's Qwen team have released competitive open-weight models.

Independent Safety Board and $1 Billion Community Fund

To address safety concerns associated with open releases, Meta is introducing new governance structures. Zuckerberg said an independent board will have the authority to approve the safety criteria governing model releases and to review whether individual releases meet those standards. He said the board will help ensure open releases "meet rigorous safety standards."

Zuckerberg also announced a $1 billion "Future is for Everyone Fund" directed at U.S. communities that host Meta data centers. The fund will support teachers, first responders, and local energy and water infrastructure. The initiative is widely viewed as an effort to address growing local opposition to data-center construction across the United States, where utility demands from AI infrastructure have prompted pushback from residents and lawmakers in several states.