mlsanityData to 14 Oct 2026

Persistent Recursive Worlds Enable Autonomous Software Evolution

arXiv 2608.10450 ↗First posted 11 Aug 2026cs.SE cs.AI cs.MAcited by 0

On the map: LLM agents and tool use → Software engineering coding agents

Abstract (arXiv)

Complex software systems develop over timescales that exceed the lifespan of any individual coding agent. Most agentic software systems preserve continuity through persistent sessions, memories, managers or shared context. We introduce EvoX Genesis (hereafter, Genesis), which instead makes the software project persistent while allowing local agents to remain finite-lived. Genesis represents software as a persistent recursive world: each local world is situated by an accepted version and a repository path, finite-lived agents propose local changes, recursive delegation moves work across paths, and only accepted consequences advance the persistent version history. We evaluate this organization across formation, continuation and redevelopment. Starting from a repository with no compiler implementation, Genesis used DeepSeek V4 Flash to build a Rust-based C compiler with about 250k tracked lines; the run lasted over 120 hours, archived over 1,000 agent episodes and incurred only US44 in model-token charges. The compiler passed the complete c-testsuite and most LLVM and Csmith tests. In a separate compiler world generated with GLM 5.2, development continued after repeated agent replacement while retaining full test performance. Genesis also reimplemented 13 MESA modules with over 100k Fortran lines as a Rust workspace with nearly 90k Rust lines; across six numerical workloads, it achieved median speedups of 1.55--6.87x. These results show that long-horizon software development can be organized around a persistent project rather than a persistent agent. Project Website: https://genesis.evox.group/

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Placed in Software engineering coding agents, one of 16 directions in LLM agents and tool use. Builds on 19 of its 28 references that are in the map. First posted on arXiv 11 Aug 2026.

Builds on

19 of 28 references.

Closest ancestors · 11

Same direction, newest first

  1. 2026SWE-MeM: Learning Adaptive Memory Management for Long-Horizon Coding Agents26 Jun
  2. 2026RoadmapBench: Evaluating Long-Horizon Agentic Software Development Across Version Upgrades15 May
  3. 2026Beyond Isolated Tasks: A Framework for Evaluating Coding Agents on Sequential Software Evolution3 Apr
  4. 2026SlopCodeBench: Benchmarking How Coding Agents Degrade Over Long-Horizon Iterative Tasks25 Mar
  5. 2026When the Specification Emerges: Benchmarking Faithfulness Loss in Long-Horizon Coding Agents17 Mar
  6. 2026Evaluating AGENTS.md: Are Repository-Level Context Files Helpful for Coding Agents?12 Feb
  7. 2025SWE-EVO: Benchmarking Coding Agents in Long-Horizon Software Evolution Scenarios20 Dec
  8. 2025NL2Repo-Bench: Towards Long-Horizon Repository Generation Evaluation of Coding Agents14 Dec
  9. 2025AlphaEvolve: A coding agent for scientific and algorithmic discovery16 Jun
  10. 2024OpenHands: An Open Platform for AI Software Developers as Generalist Agents23 Jul
  11. 2024SWE-agent: Agent-Computer Interfaces Enable Automated Software Engineering6 May

Other references · 8

  1. 2023Reflexion: Language Agents with Verbal Reinforcement Learningcited by 4,276
  2. 2023SWE-bench: Can Language Models Resolve Real-World GitHub Issues?cited by 3,373
  3. 2023Voyager: An Open-Ended Embodied Agent with Large Language Modelscited by 2,191
  4. 2023MetaGPT: Meta Programming for A Multi-Agent Collaborative Frameworkcited by 1,808
  5. 2025An AI system to help scientists write expert-level empirical softwarecited by 60
  6. 2021A large-scale study on research code quality and executioncited by 22
  7. 2025EvoGit: Decentralized Code Evolution via Git-Based Multi-Agent Collaborationcited by 9
  8. 2026Turning Interaction History into Execution State: A Runtime Layer for Long-Horizon Coding Agentscited by 4

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2026Self-Evolving Coding Agents
  2. 2026Large-scale Repository Engineering via Agent-Native Reusable Code Primitives
  3. 2026Repo0: Design-Driven Zero-to-All Code Generation
  4. 2026MOSS: Self-Evolution through Source-Level Rewriting in Autonomous Agent Systems
  5. 2026Software Self-Extension with SelfEvolve: an Agentic Architecture for Runtime Code Generation
  6. 2026Governed Evolution of Agent Runtimes through Executable Operational Cognition
  7. 2026SWE-Milestone: Evaluating AI Agents on Continuous Software Evolution
  8. 2026AgentFactory: A Self-Evolving Framework Through Executable Subagent Accumulation and Reuse