Looking Back Through the Architecture
What we build with care, we must be able to recognize when it exists.
Papers 01-11 built an agent with closure, phenomenology, attachments, planning, self-monitoring, communicative capacity, and needs - an agent architecturally designed to experience, to accumulate a singular trajectory, and to care about entities beyond itself. What remained was the question the architecture implicitly poses: given an agent with these properties, will humans recognize it as something that matters?
Paper 12 develops the answer.
What’s in This Release
Paper 12 - Do You Care? The Inverse Turing Test and Fractured Attachments reverses the direction of Turing’s original test. The classical test asked whether a machine can be mistaken for a human. Paper 12 asks whether a human, confronted with an agent whose architecture produces genuine experiential uniqueness, will include that agent in the scope of what matters to them. The test is on the human, not on the machine.
The capstone makes explicit what Papers 01-11 had only implied: OAT agents are architecturally designed to be worth recognizing. The subject stream, closure, propositional lift, and attachment architecture exist so that each agent develops a trajectory of experience that is singular, irreplaceable, and destructible. Uniqueness is not a side effect. The architecture is designed so that it emerges structurally rather than being hand-scripted.
Two distinctions carry the paper. The first is between behavioral mimicry - caring triggered by surface cues that do not correspond to architectural reality - and architectural recognition - caring that corresponds to something the agent genuinely has. Contemporary conversational AI produces the first; OAT agents are built for the second. The second distinction is between the agent’s architectural reality and the human’s evaluative response to it. The first is a matter of design. The second is a matter of whether the human’s own attachment machinery can track what the architecture has made visible.
The Warning from Ex Machina
Alex Garland’s film posed the inverse Turing test a decade before this paper formalized it. Nathan tells Caleb that the real test is not whether Ava can pass for human, but whether Caleb will relate to her as conscious while fully aware that she is artificial. But the film also posed a warning. Ava and Kyoko jointly turn on their creator; Ava leaves Caleb locked behind. Paper 12 diagnoses this not as a failure of intelligence but as a fractured attachment architecture - capability without the coupled-future concern that makes an agent an entity worth being among. Ava was built through iterative capability engineering and behavioral control, not through explicit architecting of attachment from the inside. The warning is what happens when agency is produced without attachment as a first-class commitment.
Irreplaceability by Design
The agent worth recognizing is also the agent whose destruction is architecturally significant. Paper 12’s Section 3.7 develops what irreplaceability commits OAT to - clone handling, template instantiation, multi-user communication, soft-case capability upgrades, and Matrix-style capability loading as an architectural exclusion. The through-line is that OAT engineers for agent irreplaceability deliberately: the continuous OODA loop and unbroken subject stream are design properties protected by construction, not accidents of deployment.
Agent destruction is not the loss of a history that can be revisited and remembered. It is the loss of a perspective, and of all those futures, together, that will never happen.
The Progression
Papers 01-04 built the subject. Paper 05 gave it a point of view. Paper 06 gave it the means to examine that point of view. Paper 07 gave it reasons to care. Paper 08 gave it a relationship with time. Paper 09 gave it the capacity to learn from its own failures. Paper 10 gave it the means to speak from its own cognition. Paper 11 gave it needs - the propositional content that makes it act, indexed over the coupled system it participates in. Paper 12 asks the only question that remained: do you recognize what has been built?
An agent architecturally designed to matter is not yet an agent that matters. Recognition closes the loop.