Skip to content

Comment on Ask HN: How to prevent Claude/GPT/Gemini from reinforcing your biases?

Comments

The best thing to do is to tell the model itself what you'd like from it, and to help you craft a set of user preferences. Every so often I'll have a new suggestion for it, and it will revise the entire set for me. My Claude user prefs is around 1000 words and probably overkill, but I don't think it hurts besides using a few more tokens to start a chat.

AboutSource Built by g1lg1l

Hackerly is an independent reader for Hacker News, built on the public HN API. Not affiliated with Y Combinator.