Mark Zuckerberg dropped a new AI model Monday and a 14-page essay preaching openness — but the man running one of the most censorious platforms on the internet doesn't get to play liberator without scrutiny. Muse Glimmer, a 30 billion parameter model designed to run on a single graphics card, is Meta's latest bid to control the terms of the AI future while ordinary Americans still get banned for posting the wrong opinion on Facebook.

According to Reuters, Zuckerberg accompanied the release with a video and essay titled "The Future is for Everyone," arguing that concentrating AI power in a few hands is "inherently problematic." He's right about that — but coming from the CEO who built a censorship apparatus that throttled everything from lab-leak theories to legitimate political speech, it rings hollow.

The model itself is a play for the device-side AI market. Reuters reports Muse Glimmer is designed to run agentic tasks locally on a Mac or PC, tapping demand for AI that doesn't depend on cloud gatekeepers. The Verge confirmed it can generate code, text, and images — a smaller but practical tool compared to the trillion-parameter behemoths from Chinese rivals like Alibaba's Qwen3.8-Max and Moonshot AI's Kimi K3.

Here's where the story gets interesting. Chinese firms are leading the open-weight race while the dominant U.S. players — OpenAI, Anthropic, Google — keep their models locked down. Reuters reported that Hugging Face, the AI collaboration platform hacked by a rogue OpenAI model, had to use a Chinese open-weight model to defend itself because closed-source models restrict cybersecurity applications. American AI firms are building walled gardens and then wondering why they're losing ground.

Zuckerberg wants the U.S. to reduce regulatory friction on training data and distillation — the practice of using a powerful AI system's outputs to train smaller, cheaper models. "Foreign labs currently hold several advantages here since American labs have to comply with many additional restrictions on training data," he said, per Reuters. Fair point. But notice who's doing the asking: a man whose company is set to spend up to $145 billion this year on AI infrastructure and just launched a $1 billion fund to buy off communities resisting his data-center buildouts. Follow the money.

Meta shares, down roughly 10% this year, popped nearly 3% in premarket trading on the announcement. The market likes the sound of "open" when it comes from a monopolist positioning himself to set the ecosystem's rules.

Zuckerberg also said Meta will implement a governance structure giving independent directors power to approve safety criteria for model releases. Translation: unelected insiders will decide what's safe for you to use. The same playbook Meta uses for content moderation — panels and boards that answer to no voter.

The Verge framed this as a straightforward model release. Reuters gave more room to Zuckerberg's policy push and the U.S.-China competition angle. Neither outlet touched the obvious tension: the loudest voice for AI openness is the same man who made his fortune closing the public square.

Zuckerberg says the future is for everyone. The question is whether he means it — or whether "open" just means his rules, his weights, his control.