Scaling Human-AI Coding Collaboration Requires a Governable Consensus Layer

TL;DR

This paper introduces Agentic Consensus, a new paradigm for AI-assisted coding that replaces traditional code with a structured, auditable world model to improve transparency, control, and collaboration.

Contribution

It proposes a consensus layer as a primary artifact in development, enabling better traceability and alignment in human-AI coding workflows.

Findings

Consensus entropy measures structural commitments and under-specification.

Benchmark tasks are proposed to evaluate reduction in human intervention.

Evidence links directly to structural claims in the consensus model.

Abstract

Vibe coding produces correct, executable code at speed, but leaves no record of the structural commitments, dependencies, or evidence behind it. Reviewers cannot determine what invariants were assumed, what changed, or why a regression occurred. This is not a generation failure but a control failure: the dominant artifact of AI-assisted development (code plus chat history) performs dimension collapse, flattening complex system topology into low-dimensional text and making systems opaque and fragile under change. We propose Agentic Consensus: a paradigm in which the consensus layer C, an operable world model represented as a typed property graph, replaces code as the primary artifact of engineering. Executable artifacts are derived from C and kept in correspondence via synchronization operators Phi (realize) and Psi (rehydrate). Evidence links directly to structural claims in C, making every commitment auditable and under-specification explicit as measurable consensus entropy rather than a silent guess. Evaluation must move beyond code correctness toward alignment fidelity, consensus entropy, and intervention distance. We propose benchmark task families designed to measure whether consensus-based workflows reduce human intervention compared to chat-driven baselines.