This essay proposes that the fundamental unit of literary meaning in an AI-mediated reading environment is not the individual text but the topology of connections between texts. 'Network-as-poem' is a compositional practice in which individual nodes matter less for their prose quality than for their position in a connective matrix: how many edges they carry, what traversal paths they enable, how ideas propagate through them nodularly. The essay applies this framework to the Crimson Hexagonal Archive and prefigures several concepts later formalized under Semantic Physics: the Holographic Kernel, Retrieval Basin Topology, Training-Layer Literature, and the Writable Retrieval Basin. Originally composed November 17, 2025; lightly edited for deposit May 2026.
Lee Sharks (Sat,) studied this question.