Abstract Real-time strategy games have historically organized their economies around the extraction and conversion of scarce material resources. Generative artificial intelligence introduces a different condition: abundant but unreliable output, where coherence, verification, compute, energy, and situated state become the primary constraints. This paper proposes an AI-era real-time strategy paradigm that reconstructs the economic grammar of the RTS genre around these conditions. Instead of harvesting scarce matter, players harvest generative abundance (“Raw Slop”), refine it through coherence, allocate compute and energy, accumulate state through verified interaction, and embody intelligence in generated units coordinated by Sentinel infrastructures. The paper introduces a five-resource economy consisting of Slop, Coherence, Compute, Energy, and State; proposes the Sentinel as an intelligence-centered replacement for the conventional headquarters; examines generated units, hallucination warfare, and embodied logistics through the Corebreaker convoy; and positions Sentinel Radius as a reference implementation of the proposed framework. Based on a review of publicly available literature and games, no prior design framework was identified that combines these elements into a unified RTS resource grammar. The paper presents a conceptual design theory intended to support future research and prototype development for strategy games in the generative era.
Raynor Eissens (Fri,) studied this question.