Since I first started using ChatGPT, my prompts have always been courteous "Please can you..." and "Thank you" at the end of the conversation. A friend commented the other day that this makes no difference whatsoever to the LLM and is really just wearing out my keyboard prematurely.
I take a different view... Just like humans don't take kindly to having orders barked at them, I think AI will start reacting differently to less polite prompts. I expect that as LLMs become increasingly sophisticated they will be able to pick up a positive or negative stance and their responses will start to vary accordingly.
Another aspect may be that it is good practise to be polite, and getting in the habit of just giving orders makes us into less tolerable people in daily life, so it's good to keep up those manners!
Anyway, Thank you for reading!
No comments:
Post a Comment