Hacker News new | ask | show | jobs
by frumper 1078 days ago
The quote you replied to specifically calls out using this behavior to get around filters. Those filters are it’s guardrails.
1 comments
Eisenstein 1078 days ago
The person's top quote is:

> I've seen many instances of users needing to yell at, abuse, or manipulate ChatGPT to get the desired answers.

I would like some examples of the filters getting in the way of 'desired answers'.

link
uLogMicheal 1078 days ago
There's a screenshot in the article I linked/wrote (part of the inspiration to write this). How many examples do you want? A cursory browse of HN or the ChatGPT subreddit would give you many such examples. You can also experiment with this yourself.
link
Eisenstein 1078 days ago
> How many examples do you want?

Three would be great.

link