Imagine trying to find the perfect settings for an algorithm, like tweaking knobs and dials to get the best performance. That's the challenge of hyperparameter tuning in evolutionary algorithms. Traditionally, it's been a manual, time-consuming process, but what if AI could take over? New research explores using large language models (LLMs) like Llama2 and Mixtral to automate this tricky task. Instead of relying on human intuition, these LLMs analyze the optimization logs, looking for patterns and insights to recommend better settings in real-time. The initial findings are promising, showing that LLMs can compete with traditional methods, and even surpass them in certain scenarios. This opens exciting possibilities for optimizing complex systems, automating tedious tasks, and potentially discovering even better optimization strategies with further LLM advancements. The study also reveals that not all LLMs perform equally, highlighting the need for further research into specialized AI models fine-tuned for optimization tasks. As LLMs become more sophisticated, we might see a shift towards more autonomous and efficient optimization processes in fields like machine learning and evolutionary computation.