Mark Zuckerberg wants an AI-powered future for all — and he doesn’t understand what the problem is.
In a 6,500-word manifesto titled “The Future is for Everyone,” the Meta CEO laid out the company’s philosophy around AI as one rooted in “individual empowerment as the source of prosperity,” arguing the introduction of artificial superintelligence will transform humanity, that the general population, rather than select AI companies, should harness its potential and that the U.S. should not restrict access to open-source models.
But such a future means embracing collaboration with the government and reducing, not enhancing, training guardrails that could impede technological progress, Zuckerberg wrote. On those points, he questioned why those AI companies generated discourse around AI development that appeared “so filled with doom.”
“I do not understand why anyone who believes that AI will eliminate most jobs and much of humanity’s relevance would rush to build that future,” Zuckerberg wrote. “The notion that AI is so dangerous that the only safe path is an extreme concentration of power seems inherently problematic. Historically, hoping that an absolute power will benevolently provide for humanity if sufficiently enlightened has not led to safe or positive outcomes.”
The manifesto came alongside Meta’s release of an open-source version of its most advanced AI model, “Muse Glimmer,” which came weeks after it released “Musek Spark.” But the essay — and castigation — came weeks after Anthropic CEO Dario Amodei outlined why Anthropic did not join Meta, OpenAI and other tech companies in signing a letter urging lawmakers not to restrict access to foreign open-source models, or those that can be downloaded and changed without paying for them.
Anthropic’s Amodei argued that while he didn’t believe in a federal ban on such models, he feared that China and “authoritarian governments” could use them to develop superior AI products, highlighting the ongoing tension in Silicon Valley over how AI is developed. On Monday, Sen. Bernie Sanders urged AI companies to proactively pause AI development or risk congressional intervention over fears that humans have lost control of such work.
In his lengthy AI post, Zuckerberg claimed that U.S. companies were hamstrung in the developmental process due to restrictions on training data and argued that policymakers must “reduce this additional friction if we want American open source models to lead over time.” He also criticized those, such as Amodei, who took issue with distillation, a process where an AI overwhelms another model with prompts to understand how it works.
In Meta’s view, Zuckerberg wrote, distillation marked “an important principle of how the open source ecosystem works.”
“All AI models are derived from human knowledge,” he wrote. “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. This is how the world works, and the US will not be able to lead if we restrict ourselves on this front.”
That should also mean companies should work with the government to improve cyber infrastructure with early access to AI models throughout their training process, trading such access to models for specific use cases to prevent “slowing the pace of model innovation or meaningfully delaying individuals’ access to this technology” once the models are ready for public release.
“Looking at the big picture, falling behind in AI overall would almost certainly be a larger and longer term national security issue than any specific issue we’ll face in its development,” Zuckerberg wrote.
