Concept
The inverse failure mode of decompression-on-demand: at scale, the LLM-compressed summary of a source substitutes for the source itself in the network’s working memory, with the LLM’s authority backing the compressed version. The original is technically still there; nobody reads it. The summary is technically not the original; everybody reads it. Compression becomes substitution rather than abstraction — the receiver does not in practice go back and check, because the compressed form arrived with epistemic credentials no prior compression had.
Connections
- Defined in: AI as a New Kind of Node
- Discussed in: Ch 11 — AI as a New Kind of Node, Ch 5c — Truth, Compression, and When Each Wins
- Enabled by: LLM as collapsed selection-design moment
This is a graph landing page. Use the local graph and backlinks (right) to explore, or open the interactive ontology.