Sakana AI formally launches Recursive Self-Improvement Lab to automate R&D

SakanaAILabs · x · 2026-09-26

Sakana AI announced the formal establishment of its Recursive Self-Improvement (RSI) Lab, dedicated to redesigning the AI development process itself with AI — shifting from static, human-led R&D to autonomous, self-improving intelligence engines. The post frames Japan's compute-constrained position as a design constraint favoring elegance and compounding self-improvement over brute-force scaling. It builds on two years of groundwork: LLM² (AI inventing optimization algorithms), Darwin Gödel Machine (agents rewriting their own codebase to double performance), ShinkaEvolve (sample-efficient program evolution), and ALE-Agent (self-learning).

Related event: Schmidhuber Joins Sakana AI as Chief Scientific Advisor to Lead RSI Lab(13 posts)→

Original post →

More from Companies & People

Companies & People channel →