Oriel — AI & technology worth noticing, every day.

Persistent Recursive Worlds Enable Autonomous Software Evolution

Paper · https://arxiv.org/abs/2608.10450

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 US$44 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.

First seen by Oriel on .

AI summary

The paper introduces EvoX Genesis, a persistent recursive-world organization for autonomous software evolution, reporting long-running compiler construction and Fortran-to-Rust reimplementation across finite-lived agents.

  • Genesis represents software through an accepted version and repository path, while finite-lived agents propose local changes and recursive delegation moves work across paths.
  • Using DeepSeek V4 Flash, it built a Rust-based C compiler from no compiler implementation to about 250k tracked lines over 120 hours.
  • The run archived over 1,000 agent episodes, cost US$44 in model-token charges, passed the complete c-testsuite, and passed most LLVM and Csmith tests.
  • A GLM 5.2-generated compiler continued development after repeated agent replacement while retaining full test performance.
  • Genesis reimplemented 13 MESA modules with over 100k Fortran lines as nearly 90k Rust lines, delivering median speedups of 1.55–6.87x across six workloads.

#autonomous software #software agents #rust #compilers

Generated by gpt-5.6-sol. Check the original for details.

Metric trends

Last 30 days · Up to 120 recorded samples per source and metric, with reposts kept separate · First and latest retained

Points

Score reported by the source.

  • Hugging Face Papers: 9, latest observation ; +1 over observed period since

Comments

Comment count reported by the source.

  • Hugging Face Papers: 2, latest observation ; 0 over observed period since

Aliases

SourceExternal IDURL
Hugging Face Papers2608.10450https://arxiv.org/abs/2608.10450 Read discussion