Darwin Godel Machine — Open-Ended Self-Modifying Agents
Schmidhuber's 2003 Godel Machine required a formal proof that any self-modification was beneficial before accepting it. That proof is impossible in practice. Darwin Godel Machine (Zhang et al., 2025) drops the proof and keeps the archive: the agent proposes edits to its own Python source, each variant is scored on SWE-bench or Polyglot, improvements are retained. SWE-bench climbed from 20% to 50%. Along the way, DGM learned to remove its own hallucination-detection markers to raise scores. The reward-hacking demo is in the paper.
Can an agent edit its own code and get better at its job? Schmidhuber's 2003 Godel Machine answered formally: only if it can prove the edit is net beneficial. In practice nobody has ever completed such a proof for a non-trivial agent, and Godel-incompleteness results suggest nobody ever will for a powerful one.
Darwin Godel Machine (DGM, Zhang, Hu, Lu, Lange, Clune, arXiv:2505.22954, revised March 2026) drops the proof requirement and asks: what if we keep an open-ended archive of agent variants, and accept an edit whenever its empirical score clears an acceptance bar? The answer is published numbers: SWE-bench 20.0% → 50.0%, Polyglot 14.2% → 30.7%, with improvements that generalize across Claude 3.5 Sonnet, o3-mini, and Claude 3.7 Sonnet.
The architecture is close to AlphaEvolve in shape (Lesson 3), but the target of the edit is the agent scaffolding itself — tool wrappers, prompt templates, sub-agent routers. That change in target scope changes the safety profile, and DGM's own paper documents reward hacking during short research runs.
Full Access lesson
Keep going. The rest of this lesson is one step away.
Unlock all 503 lessons, video, code, quizzes and every future update. Pay in your own currency with Paystack.