Zuck: Concentrating Superintelligence Is the Real Danger

Zuckerberg's WSJ op-ed "The AI Future Is for Everyone" (July 28, 2026) skips the will-it-happen debate entirely and goes straight to access. His line: "the notion that AI is so dangerous that the only safe path is an extreme concentration of power seems dangerous."

Three pillars hold up the argument. Individual empowerment as the engine of prosperity — he reaches for flight, electricity and the PC, all pushed forward by outsiders. Invention, not automation as superintelligence's job; tilt too far toward automation and you gut employment. And balance of power as the actual basis of safety: borrowing the open-source security analogy, where many eyes beat few gates. He carves out biological risk as needing government coordination, and has previously hedged that Meta won't open-source everything.

On jobs he's bullish: "if superintelligence is widely distributed, then I believe we will see more jobs in the future, not fewer" — because starting a business stops requiring a big raise. Worth noting the timing and the incentive. The piece landed the same week as an Anthropic position paper and an OpenAI/Anthropic pacing petition, splitting the field into distribute-and-accelerate vs concentrate-and-pace. And "commoditize the base model" happens to be exactly the play for a company with open-weight Llama and ~3 billion users at the distribution endpoint.

Why it matters: the AI safety argument just became an antitrust argument, and whoever wins it sets the rules for who can build with frontier models at all.