In plain words: A system that repeatedly rewrites its own code, tests each change on real coding problems, and keeps an archive of variants to branch new versions from. Its success rate on a coding benchmark rose from 20% to 50%, beating versions without self-improvement or branching.
Abstract
Today's AI systems have human-designed, fixed architectures and cannot autonomously and continuously improve themselves. The advance of AI could itself be automated. If done safely, that would accelerate AI development and allow us to reap its benefits much sooner. Meta-learning can automate the discovery of novel algorithms, but is limited by first-order improvements and the human design of a suitable search space. The Gödel machine proposed a theoretical alternative: a self-improving AI that repeatedly modifies itself in a provably beneficial manner. Unfortunately, proving that most changes are net beneficial is impossible in practice. We introduce the Darwin Gödel Machine (DGM), a self-improving system that iteratively modifies its own code (thereby also improving its ability to modify its own codebase) and empirically validates each change using coding benchmarks. Inspired by Darwinian evolution and open-endedness research, the DGM maintains an archive of generated coding agents. It grows the archive by sampling an agent from it and using a foundation model to create a new, interesting, version of the sampled agent. This open-ended exploration forms a growing tree of diverse, high-quality agents and allows the parallel exploration of many different paths through the search space. Empirically, the DGM automatically improves its coding capabilities (e.g., better code editing tools, long-context window management, peer-review mechanisms), increasing performance on SWE-bench from 20.0% to 50.0%, and on Polyglot from 14.2% to 30.7%. Furthermore, the DGM significantly outperforms baselines without self-improvement or open-ended exploration. All experiments were done with safety precautions (e.g., sandboxing, human oversight). The DGM is a significant step toward self-improving AI, capable of gathering its own stepping stones along paths that unfold into endless innovation.
Jenny Zhang, Shengran Hu, Cong Lu, Robert Lange, Jeff Clune
arXiv:2505.22954 · cs.AI · submitted May 29, 2025 · updated Mar 12, 2026
abstract · pdf · html · Code at https://github.com/jennyzzt/dgm
It's a helpful analogy to understand the contrast between today's gradient descent vs open-ended exploration.
[1] First half of https://www.youtube.com/watch?v=T08wc4xD3KA
More notes from my deep dive: https://x.com/jinaycodes/status/1932078206166749392