Turing Post Maps 9 Research Paths Toward Recursive Self-Improvement in AI
TheTuringPost · x · 2026-09-14
- Turing Post surveys 9 research directions toward Recursive Self-Improvement (RSI), noting no system yet achieves full RSI — autonomously building the next generation of more capable AI.
- Grouped by which part of the RSI loop they improve:
- Strategy: Metaⁿ, MetaSkill-Evolve
- Memory & skills: Recuris, SkillGLoW
- Policy: Q-Evolve, RISE
- Self-modifying code & evaluators: Mendel Gödel Machine (MGM), Red Queen Gödel Machine (RQGM), Darwin Gödel Machine (DGM)
- The piece stresses "self-improvement" is not one thing: a system can improve answers without improving the process, change policy with fixed memory, evolve code with frozen weights, or improve the evaluator that defines "better." Per the 2026 RSI survey literature, full open-ended RSI remains unsolved; the nine systems each cover one corner of the loop, together forming a research map.
More from AGI Musings
- A fiction of the "Department of Superintelligence": nationalization, open-source bans, and AI capex dependency — sterlingcrispin · 2026-09-14
- TSMC paces the frontier by accident whenever it underestimates chip demand — dan_s_becker · 2026-09-14
- Why humans may still beat frontier LLMs at open-ended AI research — flowersslop · 2026-09-14
- Google AI x Econ team finds field evidence that prior expertise drives learning from AI-assisted work — soumitrashukla9 · 2026-09-14
- Debate: don't ban open source — ensure aligned models out-compute misaligned ones — basedjensen · 2026-09-14
- Consumer agents: open questions on agentic commerce, incentives and multi-agent patterns — illscience · 2026-09-14