Refine your initial prompt instead of course-correcting

7 intellectronica 1 8/17/2025, 1:14:00 PM elite-ai-assisted-coding.dev ↗

Comments (1)

usernamed7 · 1h ago
very often i have had much better results starting with a new prompt instead of trying to continue a thread and refining from there. It has a huge impact on the quality of the output. chatGPT likes to regenerate code that has been "summarized" or is "high level" after only a few prompts back and forth. No amount of continual prompting helps. But a new chat it is much more willing to return the complete code.