Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

> and it's not too difficult to make an opinionated and challenging chatbot

Funnily enough, I've saved instructions for ChatGPT to always challenge my opinions with at least 2 opposing views; and never to agree with me if it seems that I'm wrong. I've also saved instructions for it to cut down on pleasantries and compliments.

Works quite well. I still have to slap it around for being too supportive / agreeing from time to time - but in general it's good at digging up opposing views and telling me when I'm wrong.



Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: