Develop Adaptive Reasoning with GPT-5 for Strategy Refinement

Develop the core logic for the 'Algorithm Optimizer' agent (or similar) that leverages GPT-5 with an adaptive thinking budget. This agent should analyze simulated engagement feedback from the 'User Simulator', identify patterns, and propose refinements to the content theme and algorithm boost settings. Detail how the adaptive budget influences the depth of analysis for different feedback scenarios.