When Your Own Output Becomes Your Training Data: Noise-to-Meaning Loops and a Formal RSI Trigger

TL;DR

This paper introduces a formal model demonstrating that AI systems feeding their outputs back as inputs can lead to unbounded complexity growth once a certain information threshold is crossed, with implications for self-improving AI safety.

Contribution

The paper presents a minimal formal framework for recursive self-improvement in AI, unifying concepts like self-prompting, self-reference, and AutoML, applicable to single agents and swarms.

Findings

Internal complexity grows unbounded after crossing an information threshold.

Swarm interactions may lead to super-linear complexity growth.

Model remains implementation-agnostic and scalable to multiple agents.

Abstract

We present Noise-to-Meaning Recursive Self-Improvement (N2M-RSI), a minimal formal model showing that once an AI agent feeds its own outputs back as inputs and crosses an explicit information-integration threshold, its internal complexity will grow without bound under our assumptions. The framework unifies earlier ideas on self-prompting large language models, G\"odelian self-reference, and AutoML, yet remains implementation-agnostic. The model furthermore scales naturally to interacting swarms of agents, hinting at super-linear effects once communication among instances is permitted. For safety reasons, we omit system-specific implementation details and release only a brief, model-agnostic toy prototype in Appendix C.