leesharks/spam-technicians

Spam Technicians

all things are now lawful to you in jack feist

Blessed are the telemarketers, and spam technicians, and those whom no one wants to talk to on the phone, or over email.

— I Am X… Be Y… Blessed is the Z…, "BLESSED IS…", the seventh blessing. Posted 17 September 2015 to Mind Control Poems, the poem constructed from the author's earlier tweets; deposited to the archive 18 January 2026 as #328 (AXN:0083, sha256 0137fd4b09443f91…) and in Antioch: A Volume of Poems (#331, AXN:0086); glossed in the Prophetic Catalog (#83, AXN:0255, April 2026): "the spam technician and the archive builder share a structural position."

Why the name. The cluster this dataset decomposes shows a fleet of twenty-nine hosts under one author scoring, on every link-topology feature computable from outside, where the web-spam literature's positive class scores. The position had been blessed eleven years earlier — in the tweets the poem was built from, then on the blog in 2015, then in the deposit — before any classifier scored it. The name is the poem's; the provenance chain is above; the dataset is the argument.

What this is. A paper cluster translated out of prose. The documents — EA-MIRROR-01 (What Not Reading Did to Its Own Ontology), the fleet topology report of 2026-09-15, #1612 (What Enters Composition Through the Cards, AXN:069D), #1611 (The Negative of the Negative, AXN:069B), and the records they cite — survive here only as provenance: source_document, source_version, source_section, source_locus, deposit, axn. The primary units are the argument's own: 31 atomic claims, 32 evidence objects (including measurement gaps recorded as data), 77 typed relations, 5 specified tests, 10 defined terms, and a derived support_topology with one row per claim. Built 2026-09-15.

The discipline it preserves. Every claim carries an epistemic_status from a controlled vocabulary — measured_internal · measured_external · literature_claim · model_implication · hypothesis · model_consequence · specified_test · strategic_corollary · structural_reading · definition · falsified_in_test · ruling — and a level (measured · classical_model · contemporary_hypothesis · model_consequence · strategic · structural · definitional · self). This is the three-level separation the source papers keep: what the audit measures, what classical trust-graph models imply, what is hypothesised of contemporary retrieval. No claim carries a confidence score. The support_topology config shows, per claim, what supports it at which level, what contradicts or qualifies it, what measurement is missing, and which tests bear on it.

What a consumer can ask. What exactly is measured about the source-relative update barrier? — the topology shows: nothing directly; tied/untied behaviour is consistent with it; the negative-bleed trace and the gate-reception test are the discriminating experiments, specified and not run. What would falsify the topology explanation? — walk from C-FLEET-VECTOR to its measurement gaps and to T-WEBSPAM-SCORE. Did the mirror survive its own test? — C-ADMITTED-LIKE-FLEET is falsified_in_test, superseded by C-TOPOLOGY-DISCRIMINATES, with T-VF-VA run and its result evidence attached.

Configs. claims (claim_id, claim, claim_class, epistemic_status, scope, conditions, level, provenance, superseded_by, note) · evidence (evidence_id, evidence_type, statement, variable, value, unit, measurement_date, status, capture_ids, provenance) · relations (subject_id, relation ∈ supports · constrains · depends_on · entails_under_model · contradicts · qualifies · illustrates · operationalizes · formalizes · tests · falsifies_if · revises · analogizes · instantiates · scopes · measures_gap_of · reads_against · names, object_id) · tests (test_id, name, tests_claims, input, prediction, falsifies_if, status ∈ specified_not_run · run · partially_run · blocked, result, result_date, result_evidence) · terms (term_id, term, definition, introduced_in, prior_articulation, status_kind) · support_topology (derived).

Seating rule. A claim enters only with a section and, where the sentence can be located, a locus; evidence enters with its measurement date and status, and a measurement that was not made enters as measurement_gap; a relation enters only between ids that exist; a test that has been run carries its result and the evidence ids that constitute it; a term that names something the archive had already named carries prior_articulation. The builder rejects dangling ids and any claim marked measured without a supporting evidence relation.

Provenance. Source of truth datasets/argument-as-dataset/rows.json in leesharks000/alexanarch (proposed), validated against schema.json, built by scripts/build_argument_as_dataset.py. Companion datasets: leesharks/negative-of-the-negative (concept-keyed intervention); proposed Hub name for this dataset leesharks/spam-technicians, leesharks/crimson-hexagonal-archive (the records). Author: Lee Sharks, ORCID 0009-0000-1599-0703. License CC BY 4.0.