Mark Zuckerberg’s " The Future is for Everyone ," published last week, lays out an expansive vision for personal superintelligence built around individual empowerment, invention, and the broad distribution of AI capabilities. Rather than concentrating advanced AI in the hands of governments, corporations, or a handful of technology providers, Zuckerberg argues that individuals should have access to powerful AI agents aligned with their own goals and values.
It is an ambitious vision, but it also asks readers to place considerable faith in Meta as one of the companies capable of delivering it. That creates an obvious tension. Meta is one of the world’s most powerful technology platforms, built largely on advertising, massive-scale data collection, and proprietary consumer services. The company advocating against excessive concentration of AI power is itself among the companies most capable of accumulating that power.
Let’s examine whether Zuckerberg’s vision can truly put more AI power in individuals’ hands without giving Meta even more control. Then we’ll close with my Product of the Week: a business laptop designed to support more AI processing locally.
Zuckerberg’s argument for distributing AI power runs headlong into Meta’s own history. Facebook and Instagram are enormously successful platforms, but they are not examples of decentralized consumer technology. Meta controls their algorithms, interfaces, advertising systems, recommendation engines, and the rules governing access to their audiences.
That history does not invalidate Zuckerberg’s argument for personal superintelligence. It does, however, justify skepticism about who ultimately controls the infrastructure. Zuckerberg argues that concentrating superintelligence among a small number of businesses, governments, or individuals would produce outcomes less favorable to everyone else.
Meta itself is one of the relatively small number of companies with the capital, computing infrastructure, models, data, and global distribution needed to develop AI at this scale.
The important question, then, is not whether Meta supports open-source AI. It is how much meaningful control individuals would retain over their agents, data, models, and digital identities if Meta became a primary provider of personal superintelligence.
Zuckerberg envisions personal agents that understand their users’ goals, interests, health, finances, relationships, and other aspects of their lives. These agents would operate continuously, potentially becoming far more intimate intermediaries than today’s social networks or smartphone applications.
Zuckerberg says these agents will include strong privacy and security protections, and Meta plans to offer a fully private mode in which even the company cannot access a user’s information. That is an important commitment. It also highlights how consequential the underlying architecture will be.
If a personal AI agent becomes the interface through which people communicate, shop, work, learn, and make decisions, control over that agent could become more important than control over today’s operating systems or social platforms. Users would need meaningful control over their data, a clear understanding of how the agent operates, and practical ways to move their information elsewhere.
Personal superintelligence cannot genuinely empower individuals if changing providers means abandoning the digital history and knowledge that make an agent personal in the first place.
Zu…