Introducing Sakana AI's Recursive Self-Improvement (RSI) Lab

Sakana AI announces the formal establishment of its Recursive Self-Improvement Lab in Tokyo, positioning Japan as a leader in autonomous, self-improving AI systems. The initiative builds on two years of research breakthroughs including LLM-Squared, Darwin Gödel Machine, and The AI Scientist, emphasizing progress through ideas rather than raw computational scale. The lab seeks to demonstrate that frontier AI development can advance through sample-efficient optimization, enabling democratized AI accessible to nations and institutions beyond hyperscale clusters.