Here is the sharpest objection: how is this different from an ordinary dystopia, where a few corporations own the power, robots, resources, and infrastructure, and your access depends on your behavior, ability, and perceived value?
The answer: the difference is not guaranteed by the phrase “maximize relational geometry.” Used carelessly, that phrase could justify exactly that dystopia. A corporation could say we are optimizing each person’s relational fit, then score people by compliance, productivity, risk, conformity, and predicted future value, and ration housing, mobility, healthcare, and access accordingly. That is not the model. It is a captured version of it — the language of standing wrapped around concentrated power.
So the whole thing rests on one question: who gets to define legitimate participation? If the answer is the entity that owns the compute, robots, data, energy, and logistics, the system is a dystopia no matter how elegant its words. Three hard boundaries keep it from becoming one.
First boundary: standing before value
The principle stays humane only if standing is prior to value extraction. A participant does not earn the right to belong by proving usefulness to the system. The system may ask what participation is commensurate with someone’s place — but it cannot make basic standing contingent on market value, political obedience, or algorithmic desirability.
This is the order that matters. Standing establishes a rightful place; participation is then what that place legitimately asks. The dystopia reverses it: prove sufficient participation first, and then we may grant you standing. That reversal changes everything —
Ours: standing → obligation. Theirs: performance → permission to exist.
Second boundary: jurisdiction stays distributed
If one actor owns the AI, the robots, the infrastructure, the data, the money rails, and the access system, it has swallowed nearly every function at once. It can then decide what counts as value, what counts as truth, what counts as adequate contribution, who qualifies, what is viable, and what resources exist. That is not relational geometry. It is monopoly geometry.
A healthy system separates those functions structurally. The entity that computes cannot also hold unilateral authority over legitimacy. The one that owns infrastructure cannot be the sole judge of participation. The body that allocates resources cannot define truth. The body that sets standards cannot control every exit. This stops looking like “benevolent AI governance” and starts looking like a constitutional problem — which is the right frame.
Third boundary: exit, appeal, opacity, plurality
A participant must keep meaningful exit, appeal, privacy, and plurality. If someone is algorithmically categorized and cannot contest the classification, move to another community, keep a private life, choose different affiliations, or reject the system’s judgment, then their place is not being optimized — it is being imposed. A relational system has to let a person say “you have misunderstood my place.” Without that, “fit” becomes caste.
Why AI makes this urgent
AI is extremely good at estimating what you are likely to do, what you are good at, what you cost, how compliant you are, and what role you would fill efficiently. None of that answers the deeper question — what standing is legitimately yours? A model can predict that someone is a poor economic investment; that does not establish that they deserve less standing. That distinction has to stay inviolable: algorithms advise; they do not adjudicate worth.
The line, drawn
| Relational-geometry society | Dystopian optimization system |
|---|---|
| Standing is intrinsic | Standing is earned |
| Participation follows standing | Standing follows performance |
| Authority is distributed | Authority is concentrated |
| Algorithms advise | Algorithms adjudicate worth |
| Basic resources protect participation | Resources enforce behavior |
| Exit and appeal remain real | Exit is costly or impossible |
| Difference is preserved | Difference is ranked |
| Jurisdiction is bounded | Ownership becomes sovereignty |
The clearest warning sign is a single sentence:
The moment access to basic participation becomes contingent on obedience to the system that controls the resources, relational geometry has collapsed into domination.
The correction to the axiom
This is why the model cannot be primarily about optimization. “Maximize relational geometry” is too easy to weaponize. The safer formulation is less elegant and much harder to abuse:
Preserve the legitimate standing and viable participation of each participant, while distributing power so that no single actor can define, evaluate, and control all conditions of participation.
So the goal is not to maximize geometry. It is to preserve plural, contestable, non-captive geometry — because a perfectly optimized cage is still a cage.