XNews

Google and DeepMind's Dream-RSI Cuts Agent Calls by 162x via Recursive Self-Improvement

Google and DeepMind's Dream-RSI Cuts Agent Calls by 162x via Recursive Self-Improvement

Google and DeepMind demonstrated Dream-RSI, an AI recursive self-improvement loop that replays past discovery processes to test thousands of exploration strategies at low cost. It improved or maintained result quality in algorithm design, math optimization, and GPU kernel engineering, reducing agent calls by up to 162x.

Source posts