Concept
A cross-cutting principle the capture-taxonomy makes visible. Captures of substrates that install decoding-key machinery in consumers (receiver-training, training-corpus, training-objective) are harder to recover from than captures of substrates that only shape the surface a consumer encounters (gate-criteria, option-space, deployment-configuration, preservation-archive). The first kind damages the consumer; the second damages only the substrate. Generalizes Ch 8’s capture-asymmetry between training and preservation into a substrate-level principle that applies wherever an LLM corpus, an LLM objective, or an institutional curriculum is the captured surface.
Connections
- Defined in: Capture Taxonomy
- Discussed in: Ch 11 — AI as a New Kind of Node, Ch 8 — Preservation vs. Training
This is a graph landing page. Use the local graph and backlinks (right) to explore, or open the interactive ontology.