cs.AINov 17, 2026

Think Once, Act Fast: Cost-Effective Recursive Self-Improvement for Decision Model Harness

Authors: Siqi Lai, Haitao Li, Hao Liu

Organizations: AI Thrust, Hong Kong University of Science and Technology (Guangzhou) · LIGHTSPEED

Abstract

Fast System-1 models enable responsive actions but struggle with complex or unfamiliar tasks. Large language model (LLM)-based System-2 provides strong reasoning capabilities, yet repeated invocation incurs substantial latency and computational costs. To bridge this gap, we propose JevRSI, a System-1/System-2 collaboration framework that transforms LLM reasoning into transferable decision support through recursive harness improvement. A Jev-style System-1 model enables low-latency action selection, while an asynchronous System-2 LLM develops a modular harness. The harness integrates procedures, state memory, and world models to provide decision contexts spanning the past, present, and future. Utility-driven retrieval selects modules based on both contextual relevance and their utility on previous tasks. Environmental feedback guides targeted revisions and enables recursive self-improvement. Experiments across three datasets show that JevRSI consistently improves decision quality, matching or surpassing LLM agents in several settings. It preserves low-latency execution with favorable performance-cost trade-offs. Moreover, System-2 development interventions decrease as the harness evolves, amortizing System-2 costs through continuous improvement. Our project is available at https://github.com/usail-hkust/JevRSI.

Figures & tables

Explore similar work

CardsList
  1. Thinking Fast, Anywhere: One System-1 for the Decision Points of Any Agent Workflow

    Nov 18, 2026Rui Li, Liyang He, Zheng Zhang +2LLM Agent Workflow OptimizationLLM Agent Efficiency

  2. System-One RSI: Building Muscle Memory for Interactive Agents through Guided Practice

    Nov 19, 2026Qingxu Fu, Xuanfa Jin, Ziqian Chen +2Self-Improving AgentsRecursive Self-Improvement

  3. JevAgentHarness: Benchmarking, Deploying and Training Decision Models That Share Control with LLMs

    Oct 15, 2026Kunlun Zhu, Tao Feng, Xu Han +5LLM Agent HarnessesLLM Agents