Sakana AI announced the launch of RSI Lab, a dedicated research group focused on recursive self-improvement technology for artificial intelligence. The organization believes that the next stage of AI’s future lies in open-ended systems that collectively improve themselves under constraints, contrasting with the current paradigm of scaling individual models with massive computational resources. Through initiatives spanning 2024-2026 including LLM-Squared, The Darwin Gödel Machine, and The AI Scientist, Sakana AI has built toward this vision while emphasizing progress through ideas rather than computational scale.