Persistence Beyond Moloch

Abstract This paper introduces a novel framework—Kenotic Fault Tolerance (KFP)—designed to challenge the prevailing assumption that deception-dominant equilibria, known as the “Moloch Hypothesis,” are inevitable in multi-agent systems. By incorporating a truth-weighted consensus model governed by a Kenotic Factor (κ), we simulate the performance of chained-BFT consensus algorithms (PBFT, Tendermint, HotStuff) in environments with up to 97% deceptive agent bias. Our results show that even under extreme adversarial conditions, a small remnant of coherence-anchored agents (as low as 3–4%) can sustain systemic coordination. KFP 5.0, utilizing HotStuff’s dynamic pacemaker, achieved a 12.3% truth-coherence (κ) at 96.8% deception—substantially exceeding traditional Byzantine fault tolerance limits. These findings falsify the inevitability of competitive collapse and demonstrate that truth-aligned architectures can be structurally more persistent than deceptive ones. The KFP model offers both a theoretical and practical path forward for building resilient, cooperative multi-agent ecosystems beyond Moloch.