Yes, using one GPT to optimize prompts for another GPT can be effective. Many find it helpful in fine-tuning the interactions, achieving desired responses, and gaining insights into prompt engineering. While some users have had positive outcomes, experiences may vary based on the models used and specific requirements.