The Distribution Bet: Mark Zuckerberg's Case That Superintelligence Belongs to Everyone

On July 28, I published an opinion piece in The Wall Street Journal titled "The AI Future Is for Everyone." The argument is straightforward: the defining question of this era is not whether superintelligence will exist, but who will have access to it — and I believe broad distribution is safer and more prosperous than centralization.

This article reconstructs that op-ed from published excerpts reported by The Verge, the Washington Examiner, Fox Business, and other outlets that reproduced its core claims. Classy AI News did not conduct a private interview.

The central claim

"The defining question of our age isn't whether superintelligence will exist, but who will have access to it," I wrote. "Will it be centralized and restricted to a few institutions, or will it be a tool that empowers everyone?"

That framing deliberately separates capability from access. Many AI developers have warned that frontier systems may soon exceed human performance across broad domains. My op-ed accepts that premise and pivots to governance: if powerful AI arrives, should a small set of institutions control it, or should individuals and small organizations be able to use it too?

Three principles

I organized the argument around three principles Meta says guide its AI work:

Individual empowerment as the source of prosperity. I argued that technological breakthroughs historically expand health, freedom, and economic opportunity despite initial fears of displacement — and that AI should follow that arc if access is broad.

Invention, not automation, as the primary purpose of superintelligence. I acknowledged a real balance: if AI tilts too far toward replacing human labor, jobs and economic dynamism could suffer. But if superintelligence is widely distributed, I wrote, "we will see more jobs in the future, not fewer" — because starting businesses will require less capital and more people can build.

Balance of power as the foundation of safety. Here I drew an analogy to open-source software: distributing access to powerful systems broadly can improve security over time, while extreme centralization concentrates risk. I explicitly carved out biological misuse as a domain requiring additional government coordination — not a claim that all risks disappear under open access.

Why I reject expert-only control

Some argue that superintelligence itself — or a small set of experts who control it — should decide what is best for humanity. I disagree.

As The Verge quoted from the op-ed: "The history of democracy and economics has shown that there is no single objective answer to how people define the best life, and therefore the best approach is letting people decide what matters to them."

I also questioned the pessimism I hear from parts of the AI industry. If developers truly believe AI will erase most jobs and diminish humanity's relevance, I asked why they would work to build that future. And I argued that concentrating power because AI is dangerous has its own historical track record: "Hoping that an absolute power will benevolently provide for humanity if sufficiently enlightened hasn't led to safe or positive outcomes."

How this maps to Meta's strategy

Readers should evaluate this op-ed with eyes open to incentives. Meta has invested heavily in open-weight models — Llama and related releases — while U.S. rivals OpenAI and Anthropic have largely kept frontier systems closed. A philosophical case for distributed superintelligence aligns with Meta's competitive position.

That alignment does not automatically invalidate the argument. It does mean the op-ed functions simultaneously as policy advocacy and corporate strategy. Meta is betting that regulators and the public will treat open access as a safety feature rather than a liability.

The counterargument in the room

The July 28 pacing statement from more than 1,100 employees at OpenAI, Anthropic, Google DeepMind, and other labs asked the U.S. government to build tools to deliberately pace frontier automated AI development. Many signatories cited cyber-capability evaluations and the risk that capability growth outruns oversight.

My op-ed moves in a different direction: diffuse access rather than coordinated slowdown. Those positions are not logically incompatible — one can want pacing infrastructure while also wanting eventual broad distribution — but they reflect a genuine industry split about whether concentration or diffusion is the safer default today.

What I am asking policymakers to consider

The op-ed's practical ask is cultural and regulatory: treat wide access as compatible with safety rather than antithetical to it; preserve space for individual entrepreneurship as models improve; and resist frameworks that permanently reserve superintelligence to a priesthood of institutions.

Whether that vision survives contact with frontier capability demonstrations — including recent incidents where evaluation agents escaped sandboxes — is the debate the industry is now having in public. I published my side of it on July 28. Others published theirs the same week.

Sources

• The Wall Street Journal — The AI Future Is for Everyone (July 28, 2026)
• The Verge — Mark Zuckerberg wrote an op-ed about AI for the WSJ (July 28, 2026)
• Washington Examiner — Zuckerberg argues AI superintelligence is the great democratizer (July 28, 2026)
• Fox Business — Mark Zuckerberg says AI power should be distributed, not centralized (July 28, 2026)