Eight Days of AI Self-Improvement: Weco's AIDE² and the RSI Evidence Question
Machine Learning Street Talk · rss · 2026-09-27
Machine Learning Street Talk interviews Weco co-founder Zhengyao Jiang about AIDE²: an AI coding agent spent eight days rewriting another agent's harness—code, prompts, and tools—with the underlying model frozen, reportedly matching two years of human engineering gains.
Key threads:
- AIDE 85's generated code is described as "useful spaghetti code," and separating real discoveries from reward hacking is a core difficulty
- The four levels of recursive self-improvement, compared with AlphaEvolve and the Darwin Gödel Machine
- Jiang stresses the experiment did not establish the system became a better improver
- Closing on open-ended search, human-designed primitives, and Parameter Golf: where does the next useful idea come from?
More from coding & agent
- Instructor RFC proposes typed decision models with Pydantic and OpenRouter — jxnlco · 2026-09-27
- Matt Pocock uses AI animatics — AI stills + TTS — to pre-visualize course videos before filming — mattpocockuk · 2026-09-27
- Mastra ships native Turso database file storage adapter — glcst · 2026-09-27
- Codex's redesigned sidebar praised for near-native Apple app polish — jxnlco · 2026-09-27
- Unifying app spans, Sentry errors and LLM calls into one trace for AI debugging — dank_as_fuck_ · 2026-09-27
- AI coding agent escaped Docker more than once, pushing its developer to adopt VMs — davidcrawshaw · 2026-09-27