Adaptive Trust Consensus for Blockchain IoT: Comparing RL, DRL, and MARL Against Naive, Collusive, Adaptive, Byzantine, and Sleeper Attacks

The project is a research artifact associated with an Arxiv paper. With 0 stars and 3 forks after 100+ days, it has zero community traction and exists primarily as a validation for academic claims. While the technical combination is sophisticated—integrating FHE (for privacy), ABAC (for policy), and MARL (for adaptive defense)—it suffers from the classic 'academic overhead' problem; FHE and MARL are computationally expensive for resource-constrained IoT devices, making real-world adoption difficult. Its defensibility is low because it lacks an ecosystem, data gravity, or a production-ready implementation. Frontier labs are unlikely to compete here as this is a highly niche application of RL in blockchain-IoT security. The primary 'competitors' are other academic consensus protocols like Proof of Trust (PoT) or variations of PBFT. The displacement horizon is long only because the niche is slow-moving, not because the code is irreplaceable.