Zuckerberg''s Case for AI You Personally Own
On August 10, Mark Zuckerberg published a document called "The Future is for Everyone." It''s a manifesto about what Meta thinks artificial intelligence is for, and the core claim is interesting enough to take seriously even if you don''t trust the source: that the point of superintelligence isn''t a better institution, it''s handing extraordinary capability to individual people.
He calls it personal superintelligence. The reception has been a good deal frostier than the title.
The three-part argument
Zuckerberg builds Meta''s AI philosophy on three claims. Individual empowerment is "the source of prosperity." Invention is "the primary purpose of superintelligence." A balance of power is "the foundation of safety."
That third one carries most of the weight. The standard worry about advanced AI is that a handful of labs end up controlling something enormously powerful. Zuckerberg''s answer isn''t to slow it down — it''s to spread it out. On his telling, the safest world is the one where the capability is everywhere rather than locked in a few buildings.
It''s a real position in the safety debate. It also happens to be extremely convenient for a company that bet its strategy on releasing open models. Both can be true at once.
What''s actually being promised
The concrete piece is a personal agent, free to users, that can code, design and work through complex multi-step tasks. Around it sits the bigger pitch: AI tutors, coaches, even lawyers, free or cheap, for anyone.
Hold onto that last one. A free legal assistant is either a real redistribution of professional access or an overpromise that hurts people who act on its advice. Which it turns out to be comes down entirely to execution, and no manifesto settles that.
The open-source commitment is more tangible. Meta has put out Muse Glimmer, a 30-billion-parameter agent model, under the Apache 2.0 licence — permissive enough that developers can build commercial products on it without asking anyone. That''s the balance-of-power argument written as a licence file instead of a paragraph.
The bill
The infrastructure budget behind all this runs up to $145 billion in 2026, mostly data centres.
Sit with that for a second. Here is a vision of individual empowerment that needs one of the largest private capital programmes in corporate history to arrive. The AI in your pocket runs on buildings the size of small towns, pulling from grids that ordinary households also depend on. "Personal" describes the interface. It doesn''t describe the infrastructure.
Why the pushback
TechCrunch''s response was blunt: this manifesto is exactly why people don''t like AI. The complaint isn''t that the goals are bad. It''s the framing — sweeping civilisational language from a company plenty of people already eye warily, promising to fix education, health and legal access while the business underneath it still runs on advertising.
There''s a fair version of that critique and a lazy one. The lazy version waves off the technical work because of who did it. The fair version points out that Meta has published ambitious visions of transformative platforms before, and the distance between the language and the shipped product has usually been considerable.
What to watch instead
Manifestos are cheap. Three things would tell you more than the document does.
Whether the free agent is still free once it has real users. Whether the open-weight releases keep pace at the frontier, or start landing a generation behind whatever Meta keeps for itself. And whether personal superintelligence means an assistant working for you, or one working for the ad business and aimed at you.
That last distinction matters more than any of the three pillars. An agent that books your travel and drafts your emails sits far closer to your life than a feed ever did. Whose interests it serves isn''t a philosophical question — it''s a product decision, and it''ll show up in what ships.
Image: panumas nikhomkhai, via Pexels
