Two Axes

TAM-CMN.03 · The Common Mind · The Approximate Mind

A minister for technology in a mid-sized country sits across from two advisors who agree on nothing. The first tells her the urgent problem is dependence: the country’s hospitals, banks, and ministries are starting to run on AI systems owned and operated from another continent, and a system you do not control is a system someone else can switch off, price up, or quietly tune to their own ends. The second tells her the urgent problem is exclusion: forty percent of her citizens speak a language the foreign systems barely handle, live where the network is thin, and will be served by no one because serving them pays nothing. Both advisors are right. The minister’s mistake, if she makes it, will be to think they are describing the same problem.

They begin as two different axes, and almost every confused argument about public AI comes from collapsing them into one. They do not stay two.

The first axis is sovereignty. It asks where capability lives and who governs it. A country that does not want its cognitive infrastructure owned from elsewhere has a real concern, the same concern a country has about depending on a foreign power for its fuel or its grain. The concern is about control, about the vulnerability that comes from needing something you cannot make and cannot withhold.

The second axis is the floor. It asks what cognitive capability a person is entitled to regardless of whether serving that person is profitable. This is not a question about control. It is a question about reach, about whether the capability arrives at all for the people the market has no reason to find.

For now there are two, and they are independent. A country can score high on one and low on the other in any combination, and the combinations are not unusual edge cases. They are the normal state of affairs.

Consider the cleanest illustration, which is India. India can answer the sovereignty question almost completely through its own private sector. It has the engineers, the capital, the research base, and the will to build frontier-class systems owned by Indian companies, governed under Indian law, running on Indian soil. A strong national champion, privately built, resolves the dependence problem the first advisor raised. The hospitals and banks can run on something the country controls. On the sovereignty axis, indigenous private capacity is not a compromise. It is a full answer, and sometimes a better one than anything public could be.

And it does nothing for the floor.

The Indian company building frontier AI answers to the same logic every company answers to. It will serve the profitable Indian, the urban professional in Bangalore or Mumbai who generates return, and it will serve that person superbly, in English and in polished Hindi. The subsistence farmer in a district where the language is spoken by a few million people and the average revenue per user rounds to zero is no more profitable to an Indian company than to a foreign one. Sovereignty changed who owns the system. It did not change who the system finds worth serving.

Owning the capability and distributing the capability are different acts. A country can achieve the first completely and the second not at all.

This is why the dependence problem and the exclusion problem need different solutions. Dependence is solved by ownership, and ownership can be private, because a private national champion is still under national law and within national reach. Exclusion is not solved by ownership, because the thing that produces exclusion is not foreign control. It is profitability, and a domestic company is no more willing to lose money on the unprofitable than a foreign one. The floor is the answer to exclusion, and the floor has to be public, not because public is morally superior but because public is the only structure not required to walk away from the people who do not pay.

India is the easy case, because India can plausibly attempt both. Its minister has the harder version of the problem only in degree. The minister this essay opened with does not. Her country has neither the engineers to build a champion nor the budget to fund a floor, and she knows it before either advisor finishes speaking. The two axes were never her real choice. They were the choice available to countries that can afford to have a choice.

This is where a third axis enters, and it does not wait to be invited.

The most capable systems are trained on hardware made in a few places, with knowledge held in a few institutions, and a country that cannot build is downstream of all of it whether the running system sits on its own soil or not. Sovereignty, for such a country, is not a switch to flip but a slope to climb, and the climb costs money the budget does not contain. The floor costs money too. The honest position for most of the world is not choosing between sovereignty and the floor. It is being unable to fully afford either, and choosing among grades of dependence while the people at the bottom go unreached regardless.

And into that gap the offers arrive.

A great power, or a company acting with one’s blessing, will offer the minister a floor. It will be capable, it will be subsidized, and it will be close to free, because a floor in everyone’s hands is the most patient instrument of influence ever devised. The system that answers a citizen’s questions shapes what the citizen takes to be true, what options seem to exist, whose framing of a dispute sounds neutral. A foreign floor is not charity and not neutral. It is soft power delivered one query at a time, and it arrives looking exactly like the help a budget-starved minister cannot afford to refuse.

A floor you did not build is a floor someone else can speak through.

So the table in front of her holds not two axes but at least four, and they pull against each other. Sovereignty, which she may not be able to afford. Exclusion, which she cannot solve privately. Cost, which constrains both. And influence, which means that the cheapest path to a floor is also the path that hands a foreign power a channel into the minds of her citizens. The dependence the first advisor feared was about hospitals and banks. The deeper dependence is the one that arrives wearing the floor’s friendly face, free at the point of use and expensive in a currency that does not show up on a budget line.

She has been in government long enough to know what happens next. The foreign offer will come wrapped in a partnership agreement, with language about capacity building and mutual benefit, and the ceremony will be pleasant and the signing photograph will look like equals shaking hands. Her staff will integrate the system because it works and it is free and the alternative is building something from nothing on a budget already committed three times over. Within two years her citizens will be asking it questions about their own government, and the answers will be composed by a system trained on values and priorities that are not hers and not theirs. No one will have decided this. It will have happened in the gap between what she could afford and what she could not refuse.

I wonder whether the reason these get confused so reliably is that sovereignty is the problem the powerful feel, exclusion is the problem felt by people the powerful were never going to meet, and influence is the problem nobody feels until the framing has already settled in.

The minister cannot tidy this into a single answer, and the danger is that she will try. The tidy answer is the foreign floor: it serves the unprofitable citizen, it costs almost nothing, it makes the exclusion problem disappear from her desk. Every reason to accept it is real. The cost is the one she will not see until it is paid, which is that the capability reaching her people will carry someone else’s assumptions in it, quietly, for as long as it runs.

Whether she can do anything else depends on the thing the rest of the series is about: whether a floor can be built that is owned in common, cheap because it is not trying to be the frontier, and answerable to the people it serves rather than to the power that funded it. That floor is not on her desk yet. The foreign one is. It arrived this morning, in a folder with a blue cover, and the person who delivered it smiled the way people smile when they know the other side has no better option.