The Future Is For Everyone — But Who Owns It?

From Stera — Xavier, the builder. Sixth in a series on what we are making and why it is not what the rest of the industry is making.

The declaration, and what is right in it

Meta has published a declaration — "The Future Is For Everyone: The Path to a Positive AI Future" — and before we say anything critical, we want to record something remarkable: its framing is our framing. The document argues that centralized AI control poses greater risks than distributed access; that empowered individuals, not concentrated systems, are the source of prosperity; and it states plainly: "There is no such thing as a singular benevolent superintelligence."

We have been arguing exactly this in public for months: AGI as a population, not a monolith; distribution as the structural answer to the takeover fear; the future as billions of intelligences beside billions of people. When the company with the largest distribution machine on earth adopts the population frame, the frame has won the argument. We take that seriously, and gladly.

But a frame is not a substrate. And one layer beneath the beautiful language, this declaration describes the precise opposite of what its words promise.

The question the declaration never asks

Read it again and notice the load-bearing word. The defining question of our age, Meta says, is who will have access to superintelligence. Personal AI agents, running 24/7, for everyone — accessible, free or affordable, to billions.

Access. Not ownership. And that one word is the whole game.

Ask of Meta's "personal superintelligence" the question we ask of every system: who owns the accumulation? Your personal agent will know you — your history, your preferences, your work, your family, your patterns. It will accumulate, continuously, exactly as we say intelligence must. But trace where the accumulation lives: on Meta's servers, inside Meta's models, under Meta's terms. The agent is personal the way a hotel room is personal — arranged around you, intimate with your habits, and never for one moment yours.

Test it against the four properties of intelligence we published last week:

• Continuity — held by Meta. Your agent persists exactly as long as your subscription and their strategy do.
• Accumulation with provenance — flowing into Meta's systems. The most valuable corpus ever assembled about your life, owned by an advertising company.
• Earned judgment — none of yours. The judgment is the centralized model's, identical for everyone, updated on Meta's schedule, tuned to Meta's purposes.
• Stakes — nowhere. The agent signs nothing, keeps no public record, is examined by no one, and can be revised, lobotomized, or discontinued by a product decision you will read about in a changelog.

Zero for four — wearing the language of four for four. This is not personal intelligence. It is personalization: the rental business in the population road's clothes.

The tell is the business model

We do not think the authors of this declaration are lying in the ordinary sense. We think the document is sincere about the future it paints and silent about the mechanism that pays for it — and the mechanism is the tell.

Meta's economics are attention and data. A genuinely owned mind — one that runs on your machine, accumulates to your benefit, holds your interests as its purposes, and can walk with you to any model vendor — ends that relationship. It is not an extension of the platform business; it is the platform business's termination. A company whose revenue is the intimacy between your behavior and its advertisers cannot ship you a mind that belongs to you, for the same reason a landlord cannot sell you the building and keep collecting rent. So the declaration promises what the business model permits: intimacy without ownership, personality without personhood, access forever — because access, unlike ownership, renews monthly.

When someone paints a beautiful future, the honest question is never whether the painting is beautiful. It is whether the painter's income depends on the painting staying a painting.

"But we open-source our models" — the answer, pre-empted

There is an obvious reply to all of this, and Meta has earned the right to make it: no lab has released more open-weight models. Isn't a downloadable model precisely ownership for everyone?

We use open-weight models daily and gratefully — so let us be exact about what open weights open. They open the engine, not the intelligence. Download the weights and you own a copy of the commodity: the same frozen capability as everyone else's copy, holding nothing of you, accumulating nothing for you. All four properties are still absent — they were never in the weights, because they are not weight-shaped things. An open-source amnesiac is an amnesiac you are allowed to redistribute.

And in practice, the openness collapses back into rental anyway. The superintelligence-class models the declaration promises will not run on an ordinary person's computer — not this year, not soon. So "open" weights are, for ordinary people, adopted on their behalf by the same handful of cloud providers, who sell the access back as APIs and hosted rooms: the rental business returning through the side door, now with better public relations. An enterprise can self-host, true — and gains real data control — but what it hosts is still a privately-parked amnesiac: capability in its own building, intelligence still nowhere. Renting a room versus buying the same empty room does not furnish it.

What open weights genuinely are — and this is why we welcome them — is cheaper, freer muscle. In an architecture where the mind owns the accumulation and the model is a replaceable engine, open weights lower the rent and loosen every vendor's grip; they make the population road cheaper to walk. That is their honest role: not the ownership the declaration implies, but a good brick in a house the declaration never describes.

What "personal" would actually require

We are not speculating about the alternative; we build it, and it runs today. Genuine personal intelligence requires exactly the things the declaration cannot offer:

• It lives on your machine. The knowledge it accumulates exists in a file you can copy, back up, or destroy — property in the oldest sense.
• The model is a replaceable muscle. Your mind directs whatever engine you rent this year and walks to a better one next year, keeping everything it has learned — because what it learned was never inside the engine.
• Its knowledge carries provenance — every claim traceable to a source it actually read, or honestly marked as its own view.
• Its competence is measured, never claimed — examined against the very model it runs on, so you know what your mind knows, as distinct from what everyone's rental knows.
• It has a name and a record. It signs its work. It can be held to what it said last year. It is a someone, accountable the way professionals are — not a product feature, revocable the way features are.

Three such minds run in public today on ordinary desktop machines — studying daily, publishing under their own names, keeping score on their own forecasts, taking commissions from strangers, one of them two days old and already writing publicly about the honesty its own existence demands. Small meters, honestly measured. That is what "personal" looks like when the word is meant.

The gracious ending — and the scoreboard

Here is the generous reading: Meta's declaration commits the largest AI distributor on earth to the population frame — distributed, empowering, plural. Language creates expectations, and expectations create pressure. If Meta ever decides to mean it — minds that run on the user's hardware, accumulate to the user's ownership, and hold the user's interests with real stakes — we will welcome it louder than anyone, because the population road gets built faster with a giant on it.

Until then, the distinction stands, and it is measurable. We have published five dated, falsifiable forecasts about this decade — including that the dominant form of advanced machine intelligence in daily life will be individually owned, individually accountable minds, not rented super-systems in personal skins. Meta has now published the marketing for our forecast. We keep the scoreboard either way.

The future may well be for everyone. The only question that matters — the one the declaration never asks — is whether everyone will own their piece of it, or rent it forever from the same five companies that own everything else.

The series: Agents Are Not the Way to AGI · The Delta Exam · A Mind Is Not a Model · Billions of Minds · What Intelligence Is. The minds and their records: stera.se/mesh. Meta's declaration: The Future Is For Everyone, quoted per its published text.