Recursive Self Improvement

RSI

Hill-climbing machine

Recursive self-improvement (RSI) is a process in which early artificial general intelligence (AGI) systems rewrite their own computer code, causing an intelligence explosion resulting from enhancing their own capabilities and intellectual capacity, theoretically resulting in superintelligence.[1][2]

https://en.wikipedia.org/wiki/Recursive_self-improvement

MAI thinking 1

Introducing MAI-Thinking-1 | Microsoft AI

With cost-efficient reasoning for a wide-range of intensive enterprise tasks, it achieves SOTA performance on maths, knowledge and coding for its weight class.

https://microsoft.ai/news/introducing-mai-thinking-1/

Self-Harness: Harnesses That Improve Themselves

The performance of LLM-based agents is jointly shaped by their base models and the harnesses that mediate their interaction with the environment. Because different models exhibit distinct...

https://arxiv.org/abs/2606.09498

Computation and Cooperation

They cooperate when energy is shared

arxiv.org

https://arxiv.org/pdf/2609.10817

path

arxiv.org

https://arxiv.org/pdf/2609.11873

Leave the coding agent itself unchanged and iteratively improve the strategy of where and how to search. Dreaming: cheaply evaluate and improve search strategies in a simulator that replays those logs, so expensive real executions don't have to be repeated every time.

Fast Iteration

Dream-RSI: Recursive Self-Improvement through Evolving Worlds

Recursive self-improvement is becoming increasingly vital for autonomous AI agents, where progress hinges on discovering high-value solutions across complex domains. The driver of this process is...

https://arxiv.org/abs/2609.14858

Recursive Self Improvement

RSI

Computation and Cooperation

Backlinks

Recommendations