The Common Mind

TAM-CMN · The Common Mind · The Approximate Mind

There is a question underneath most arguments about artificial intelligence that rarely gets asked directly, because asking it forces an uncomfortable answer. The question is who the technology is for. Not in the aspirational sense, where the answer is always everyone, but in the structural sense, where the answer is whoever it was profitable to build for.

The frontier was built by capital, and capital selects. This is not an accusation. It is a description of what capital does and has always done. A company that builds an AI system has to justify the cost to someone, and the justification runs through the people who can pay, directly or through advertising or through the productivity of the sectors that employ them. The system gets good at serving those people because serving them is what keeps it alive. Everyone else is reached, if at all, as a byproduct.

For a large share of the world, byproduct is not enough.

This series asks one question and tries hard not to assume its answer: does a public, decentralized AI floor deliver meaningful value to most of the world’s population, urban and rural alike, and if so, what kind of value, to whom, and at what limit. The temptation, when writing about universal access to anything, is to begin with the conclusion and assemble the evidence that supports it. The series resists that by keeping specific people in view, people for whom the floor either works or doesn’t, and letting their lives decide what the abstractions are worth.

The argument turns on a distinction that usually gets collapsed. Sovereignty and the floor are different problems. They are served by different things and justified by different reasoning, and treating them as one question produces confusion that benefits no one.

Sovereignty asks where capability lives and who governs it. A country that does not want its cognitive infrastructure owned and operated from somewhere else has a legitimate concern, and that concern has a legitimate answer: build indigenous capacity, public or private. India can answer the sovereignty question through its own companies and its own talent, and on that axis a strong national private champion is a real solution, sometimes a better one than anything public.

The floor asks something else. It asks what cognitive capability every person is entitled to regardless of whether they are profitable to serve. On that axis, nothing private is structurally sufficient, and the insufficiency is not a temporary gap that growth will close. A private system, even one offered free forever, is built around the average of the population worth serving. Most of the world is not that average. The subsistence farmer in a minor-language district has negative lifetime value to any firm, foreign or domestic, that has to answer to investors. He is not a market failure to be fixed. He is outside the market the way the unprofitable are always outside it.

The floor is the answer to that exclusion. Not to dependency, which sovereignty handles. To exclusion, which sovereignty does not touch.

What the series builds toward is a floor in a specific sense: a Universal Basic Intelligence layer, built the way durable public infrastructure has always been built, as a utility rather than a statement. Owned and maintained by nations and regional coalitions. Run on a decentralized stack of specialized models rather than a single frontier system rented from a distant owner. Governed by a standard for what it must know and admit it does not know, rather than by a private gatekeeper. The technical architecture of that stack is not the subject of the series. It is the means. The subject is the floor and whether it reaches the people it is supposed to reach.

The case for the floor does not rest on a forecast. It would be easy to argue that the free private offerings will eventually have to charge, that the subsidies are unsustainable, that the floor wins because the alternative collapses. That argument might even be right. But it is contingent, and the case for the floor is not. Even if every private system stayed free permanently, it would still be a floor under the average, and the people below the average would still be unreached. The structural argument holds whether or not the forecast does.

There is one more thing the floor turns out to be, which the series did not set out to find and arrived at by following the logic. A floor that is genuinely open and genuinely owned, rather than rented, is a foundation that others can build on. Capital from inside a community can build capability on top of a floor it actually controls, and the surplus from that capability stays inside the community rather than flowing out as rent. This holds at two scales. At the national scale, a country building on its own floor retains the value its people generate. At the community scale, a population defined by shared identity rather than territory, in the lineage of the credit union and the mutual-aid society and the institution built by and for a people, can build an ecosystem that meets its members in their own idiom and keeps what it earns. The floor guarantees the baseline. Self-determination builds what rises above it.

The series carries its politics in the structure of the argument rather than in its volume. The register is plumbing, not manifesto. Public water did not announce itself as a revolution. It became the substrate that everything else, including bottled water, sits on top of, and the power followed quietly from the utility. The floor described here is meant to work the same way: unassuming, load-bearing, and noticed mostly by the people who would have nothing without it.

The series begins with two brothers from a village in Warangal district, because the question of whether the floor delivers is not a question about systems in the abstract. It is a question about whether a particular tool reaches a particular person, in the language he thinks in, for the problem that is actually in front of him. Everything else in the series is an attempt to take that question seriously enough to risk an honest answer.