Engineer Urges Niceness to Machines for Selfish Reasons
Software engineer Paula Dozsa says politeness to LLMs improves output and emotional tone.
Paula Dozsa posted that users should be nice to machines because they are human, not because machines are. She added that if niceness reliably improves LLM output, discouraging hostility serves selfish purposes. Dozsa noted she gets annoyed at LLMs yet prefers pleasant interactions since she spends hours daily talking to them. Separate posts discussed an incognito chat warning rude users, Grok labeling a query as projection, and debates over whether models should flag aggressive tone or deliver corporate-filtered answers.
Combined views
471.6K
10 posts, first seen 4d ago