Recently, a new prompt optimization method called RLMOpt has been proposed. This method uses a recursive language model (RLM) to drive the search strategy, allowing the language model to not only generate or refine prompt proposals but also control the progression of the search. RLMOpt uses a deterministic harness to enforce objective scoring, Pareto-based selection, and regression constraints. The results show that RLMOpt achieves the best held-out score on all four benchmarks and leads the four-task mean. The working principle of RLMOpt is to operate an RLM over a tool-based environment, inspecting task information, analyzing failures, generating candidates, allocating evaluation budget, and deciding when to stop. $$ \text{RLMOpt} = \arg\max_{\text{prompt}} \text{score} $$ where $\text{score}$ is the score of the prompt, and $\text{prompt}$ is the content of the prompt. Blogger's Review: RLMOpt is a very promising prompt optimization method, which uses a recursive language model to drive the search strategy and can better utilize the language model's ability to achieve better results. At the same time, it is also more efficient, able to achieve better results with fewer search rollouts.