Resource 1
Raw Slop
Abundant generated possibility with variable coherence, relevance, provenance and mutation risk.
From Tiberium to Slop: a resource theory for real-time strategy games in the generative era.
Real-time strategy games have historically organized play around scarce material resources: spice, Tiberium, minerals, gas, wood, food, gold, stone, mass and energy. This paper proposes a different RTS paradigm for the generative era: not merely AI as an opponent, asset generator or tool, but generative abundance as the primary economic substrate.
The model defines Raw Slop as abundant but unreliable generated possibility. Players harvest it and convert it through coherence, computation, energy, state, provenance, doctrine and embodiment into dependable strategic capability.
Classical RTS games harvest scarce matter. An AI-era RTS can harvest generated possibility - converting abundant, unreliable output into coherent, embodied and strategically situated capability.
The AI-era RTS shifts the strategic bottleneck from material extraction to the conversion of unstable abundance into usable capability.
This extends the Slop Resource Principle from a content-layer argument into an economic and strategic grammar: Slop is not merely what appears in a game. It becomes what the player gathers, processes, risks, weaponizes and turns into organized capability.
| Game family | Primary resource logic | Strategic bottleneck |
|---|---|---|
| Dune II | Spice harvested from exposed fields | Territorial access and transport |
| Command & Conquer | Tiberium harvested and refined | Field control, harvester safety and expansion |
| Warcraft / Age of Empires | Multiple physical materials | Worker allocation and age progression |
| StarCraft | Minerals and vespene gas | Expansion timing and asymmetric conversion |
| Supreme Commander | Mass and energy flow | Throughput, storage and production balance |
The AI-era proposal does not reject this lineage. It asks what happens when the foundational input is not naturally scarce matter but an excessive stream of generated possibility.
The proposed economy contains five linked resources. They should not behave as interchangeable currencies; each represents a different stage or constraint in converting generated possibility into field action.
Abundant generated possibility with variable coherence, relevance, provenance and mutation risk.
The refinery function: classification, filtering, alignment, deduplication, verification and binding to doctrine or use case.
Allocatable cognition produced by datacenters and consumed by training, inference, detection, coordination and verification.
The physical substrate that powers datacenters, foundries, shields, transport, the Sentinel Radius and Corebreaker logistics.
Earned situational knowledge acquired through observation, scouting, continuity, verification, survival and preserved residue.
Slop is harvested. Coherence is produced. Compute is allocated. Energy is supplied. State is earned. Capability is embodied.
The Sentinel is the architectural departure from a conventional RTS base. It is not a town hall with an AI theme placed on top. It is the intelligence infrastructure from which the base derives its ability to perceive, interpret, coordinate, assimilate and develop.
Issues global priorities and coordinates production.
Represents terrain, threats, routes, resources and uncertainty.
Higher states unlock new forms of classification and capability.
Identifies nuclear signatures, stealth, poisoned resources and false signals.
Distributes processing capacity across training and operation.
Destroying or overriding the Sentinel collapses the intelligence architecture of the base.
The Sentinel Radius is operational awareness: the spatial extent of sensing, inference, power, state exchange and command reliability.
Production remains legible but becomes compositional. The player produces a capability template and embodies it in a chassis.
| Model | Chassis | Embodied result |
|---|---|---|
| Observation model | Flying micro-drone | Survey Swarm |
| Route model | Tracked utility chassis | Pathfinder |
| Siege model | Heavy walker chassis | Breach Walker |
| Coherence model | Relay chassis | Field Stabilizer |
| Repair model | Distributed microbot body | Residue Repair Swarm |
This preserves RTS readability while making unit quality, doctrine, provenance, state and infrastructure dependence strategically visible.
Traditional RTS information warfare includes fog of war, stealth, decoys, scouting denial and false positioning. An AI-era RTS extends this into attacks on the enemy's world model.
False observations: phantom armies, duplicate convoys, synthetic resource fields or fabricated nuclear signatures.
Corrupted classification or routing: interceptors chase a decoy or a bridge is misread as passable.
Temporary unstable structures or terrain: false bridges, copied walls, incomplete turrets or collapsing road surfaces.
False information enters persistent state and contaminates route predictions, enemy profiles or unit doctrines.
The design challenge is fairness: hallucination warfare should create uncertainty, not arbitrary interface lying. Confidence indicators, replay tools and provenance trails make epistemic combat legible.
The Corebreaker replaces the instant superweapon with a slow catastrophic convoy. Once launched, it must traverse the map toward the enemy Sentinel. The attacking player clears routes and defends the vehicle; the defender attacks its route, escort, energy supply and world model.
Models can generate plans rapidly. Datacenters can evaluate routes. Hallucinations can fabricate alternatives. Yet the vehicle still moves through physical distance. Exponential intelligence meets stubborn matter.
Sentinel Radius is presented as a reference implementation of this resource grammar. The minimal vertical slice includes one Sentinel per side, one Slop Field, a Coherence Plant, Fusion Reactor, Datacenter, Model Forge, Robot Foundry, one hallucination ability and one slow Corebreaker with destructible obstacles and two viable routes.
The reference implementation matters because the theory must be tested as play: whether coherence is understandable, compute allocation creates readable trade-offs, generated units remain predictable, hallucination warfare is fair and the convoy produces strategic tension rather than frustration.
The novelty claim is intentionally narrow. AI in RTS is not new. Procedural unit generation, datacenter games, convoy missions, deception mechanics, assimilation systems and sentient bases all have precedents. The contribution proposed here is the integrated grammar.
No prior game or design framework was found in the review for this preprint that combines generative abundance as a harvestable RTS resource, coherence as an economic conversion layer, Sentinel-centered AI infrastructure, generated embodied units, hallucination warfare, epistemic assimilation and a slow physical superweapon journey into one cohesive real-time strategy model.
This is a limited literature and market review, not proof that undiscovered or unpublished work does not exist.
The central proposition is plain: classical RTS games harvest scarcity; an AI-era RTS can harvest abundance. Its strategic problem is not simply how to obtain more output, but how to make possibility coherent, trustworthy, embodied and effective within a contested field.
Sentinel Radius is intended as a reference implementation of that proposition. The theory establishes the grammar; the prototype must test whether the grammar creates a game worth playing.