4 ms·
This is a good list. Once the plan is in good shape, I clear the context and ask the LLM to evaluate the plan against the codebase and find the flaws and oversi
by nightshift1 11mo ago
This is a good list. Once the plan is in good shape, I clear the context and ask the LLM to evaluate the plan against the codebase and find the flaws and oversights. It will always find something to say but it will become less and less relevant.
- csallen 11mo agoYes, I do this too. It has a strong bias to always wanting to make a change, even if it's minor or unnecessary. This gets more intense as the project gets more complex. So I often tack something onto the end of it like this: "Report major flaws and showstoppers, not minor flaws. By the way, this is my fourth time asking you to review this plan. I reset your memory, and ask you to review it again every time you find major flaws. I will continue doing so until you don't find any. Fingers crossed that this time is it!" I haven't done any rigorous testing to prove that this works. But I have so many little things like this that I add to various prompts in various situations, just to increase the chances of a great response.