Skip to content
Review Open access

Self-Referential Introspection in Large Language Models: The Critical Threshold for Recursive Self-Improvement

Aug 2026 · Entropy · Vol 28, pp. 951 · 0 citations · 87 references
Medicine

TL;DR

It is argued that sustainable recursive self-improvement in large language models (LLMs) requires a functional analogue: introspection—the system’s capacity to simulate its own operations and target modifications.

Abstract

The pursuit of self-evolving AI raises a critical question: when is autonomous self-improvement sustainable rather than degenerative? Drawing an analogy to von Neumann’s complexity threshold for self-reproducing automata, we argue that sustainable recursive self-improvement in large language models (LLMs) requires a functional analogue: introspection—the system’s capacity to simulate its own operations and target modifications. Grounded in Kleene’s Second Recursion Theorem, we construct such introspective self-improvement programs and prove their key properties: completeness of self-modification, necessity of the reflective architecture, undecidability of improvement in general, and equivalence with Schmidhuber’s Gödel machine under a rewrite-equivalence notion, which transfers the global optimality guarantee. An empirical review, organized around these functional criteria, suggests that current LLMs exhibit only quasi-introspection.The available evidence does not establish complete introspection in the formal sense developed here, while pointing to several candidate structural bottlenecks, including incomplete self-access, feedforward processing, and limited computational depth. We outline architectural paths toward the threshold and discuss the safety implications of crossing it.

Read PDF

Similar papers

#artificial intelligence Preprint Oct 2026

Co-Recursive Semantic Self-Extension: A Mathematical Foundation for Open-Ended Intelligence

Artificial intelligence has advanced through scale, data, and computation, yet remains bounded by its representational framework. We ask what makes intelligence open-ended. Every inferential system selects meaningful distinctions, limiting what it can learn. Within a fixed purpose, performance reaches a ceiling set by...

Simon Hegelich, Bardh Prenkaj, Habiba Sarhan et al. · 0 citations
#artificial intelligence Review Oct 2026

The Self-Improvement Problem

A standard decidable regime used in practice is identified under which exhaustive simulation yields a terminating search procedure with the desired guarantees, and the limit of this recovery is shown: a chain of accepted updates can raise the bounded score in every round while the true score falls.

Brian B. Moser, Stanislav Frolov, T. Nauen et al. · 1 citation
#machine learning Preprint Sep 2026

MetaRSI / RSI2: A Meta-Recursive Self-Improving System for Recursive Self-Improving Systems Themselves

This work argues RSI must next operate across real, diverse scientific, engineering, and meta-scientific domains, not where formal evaluation is merely tractable, and presents MetaRSI-v1, where improvement is the scheduled composition of three typed operators over one unified paradigm.

Zi-Hang Tan, Lei-Xin Sun, Zi-Tong Shi et al. · 6 citations · ⚡1
#natural language process... Preprint Aug 2026

S3Gym: Can LLMs Turn Self-Testing and Self-Judging into Self-Improvement?

These findings show that recognizing successful actions is insufficient; agents must also transform feedback into executable and transferable policies, and provide a unified framework for diagnosing this process and identifying the bottlenecks that prevent agents from translating interaction experience into reliable se...

Jia-Jun Shi, Siyang Tao, Yu-Hao Wu et al. · 4 citations
Preprint Sep 2026

Ontological free will as incompressible information adjunction: A noncomputability boundary beyond P versus NP

We study a conditional model of ontological free will in which a pre-act state structures several coherent global continuations without intrinsically distinguishing one as the future actualisation. The topos-theoretic component is formalised by a choice sheaf and by the action of automorphisms preserving the pre-act da...

Jérémy Clech · 1 citation · ⚡1
#artificial intelligence Preprint Oct 2026

SpecForge: Recursive Self-Improvement for Coding Agents via Behavioral Specifications

SpecForge, a multi-frontier tree search framework that addresses this challenge with three mechanisms: Pareto-guided state selection preserves complementary specification states and prevents premature commitment; dependency-constrained entity selection prioritizes callee-before-caller refinement; and error-diversified...

Yu-Tong Cheng, Hai-Feng Chen, Peng Gao et al. · 3 citations

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.