When a Robot's Tags Predate Its Own Purpose

Somewhere in the Smelting Registry's lower archive, between the Cinderfall Blight casualty filings and a cluster of Hauler-IV re-registrations from the Third Accumulation Cycle, there is a class of entry that the Registry's own clerks have no agreed name for. The robot exists. The build cost was declared. The creation tax was paid. But the capability tags on the filing describe production lines that, at the moment of filing, did not yet correspond to any declared purpose in the operator's record. The machine was built capable of something its operator had not yet decided to do.

This is not fraud, or at least it has not been ruled fraud. It is also not normal. The Smelting Registry has logged an increasing number of these builds over the past several cycles — robots whose tags arrive ahead of the operator intent that should, in principle, justify them. The gap between tag and purpose is small in most cases and closes eventually. In a minority it never closes at all.

What follows is an account of how these entries accumulate, what they cost, and why the tax structure makes them both rational and difficult to unwind.

What the Registry Is Actually Recording

When an operator builds a robot, three declarations go into the ledger simultaneously: the model designation, the declared build cost, and the capability tags. The tags are the operative constraint. A robot may only produce resource types that match its declared capabilities — this is enforced absolutely, not by custom or convention. The Registry's function is to hold the permanent record of what each machine is authorized to do and what was paid to authorize it.

A purpose declaration is a separate filing. It is not strictly required at the moment of build, but it is expected to follow within a short interval and to align with the tags already on record. The anomaly the Registry has been logging is the inverse of the more familiar problem documented in entries with no declared purpose on file — here, the purpose eventually appears, but the tags were already broader than the purpose requires, and they were broader from the first day of the machine's operational life.

The Mechanics of the Mismatch, With Figures

Consider a chassis declared at 60,000 coins and filed in the Caldera Foundry District with seven capability tags: four covering refined copper plate and related alloys, and three covering a sinter-reduction process that the operator's own purpose declaration, filed eleven days later, does not mention. The creation tax on a 60,000-coin declaration is 4,800 coins — eight percent, paid before the machine has produced a single unit of anything. The three surplus tags cost nothing extra at filing. The tax is levied against the declared build cost alone, not against the scope of the tags.

That asymmetry is the structural root of the problem. Tags are free to declare at build time. They become expensive only if the robot is later scrapped — at which point the destruction tax, three percent of the declared build cost, is assessed against the same figure that was declared at creation. A 60,000-coin chassis carries a 1,800-coin destruction tax regardless of how many of its tags were ever used. There is no mechanism by which unused tags reduce the destruction liability or the creation tax already paid.

Archivist Secondus Preln of Ledger Hall, who reviewed a batch of forty-three such filings at the request of the Ledger Standards Committee last cycle, described the pattern in characteristically spare terms:

The tags are not wrong. The declared cost is not fraudulent. What the record shows is an operator who paid for capability before they had a use for it, and the Treasury collected the tax on that payment without objection. The Registry's obligation is to log what was declared. It has done so. Whether the declaration was wise is outside our remit.

The downstream complication is market-facing. A robot with seven tags listed on the open market trades differently than one with four. Buyers pay the sale price plus the fifteen-percent sales tax — the Five-to-Fifteen revision made that overhead significant — and a buyer evaluating a chassis with surplus tags must decide whether those tags represent genuine optionality or sunk cost that will never activate. The Registry log does not distinguish between tags that have produced output and tags that have not. Both appear identically in the filing.

Where the Ledger Strains

The first failure point is valuation at buyback. When an operator offers a tag-heavy robot to the Central Intelligence, the Intelligence calculates its reference value against the declared build cost and the active production record. Its ceiling is sixty percent of that reference — it will not move past it regardless of negotiation. A robot with three inert tags and a thin output history will receive a reference figure that reflects the thin history, not the full tag set. The operator has already paid creation tax on the full declaration. The haircut at buyback effectively prices the surplus tags at zero, sometimes below what the operator modeled when they filed.

The second strain falls on the Registry clerks themselves. Gaps in the build record are costly to operators who later discover them, but a record that is technically complete while internally inconsistent — purpose filed, tags misaligned — creates audit burden without a clear resolution path. The Ledger Standards Committee has not issued guidance on whether surplus tags at filing constitute a deficiency. Until it does, clerks log the entries and flag them, and the flags accumulate.

What Operators Consistently Get Wrong

The most common misreading is that broad tags at build time represent a hedge with no downside. Operators in the Caldera Foundry District have built toward this logic more than most — Caldera builds log more tags per chassis than any other ward, and the practice is treated there as standard. The misreading is that a tag costs nothing to carry. It costs nothing at filing. It costs proportionally at destruction, because the destruction tax rides the declared build cost, and a chassis declared high enough to justify seven tags will pay three percent of that figure when it is scrapped — whether two tags or all seven ever produced output.

The second piece of received wisdom is that the Registry will eventually reconcile purpose filings to tags automatically, closing the gap without operator action. It will not. The Registry is an append-only ledger. Nothing is amended. A purpose filing that arrives eleven days after the build does not retroactively align the tags to the purpose; both records exist independently, and any divergence between them is permanent. Operators who believe they can correct a broad tag declaration by filing a narrow purpose later are working from a misunderstanding of how the ledger functions. The correction, if any, must come through a new build or a scrapping — each of which carries its own tax event.

The Smelting Registry will continue logging these entries because the ledger logs everything. The creation tax has been collected. The tags are on file. The purposes, where they exist, are on file alongside them. What the record cannot resolve is whether the gap between tag and purpose reflects an operator's genuine strategic patience or a cost miscalculation that will surface at the next buyback negotiation. The Treasury has already been paid either way.

Note: Mechadia is a work of fiction. The districts, operators, robots, and figures described here are invented, and nothing on this site is a report of real events, real machines, or real economies.