Post #2346356
2026-05-07 13:05 UTC
@wronglang@bayes.club @strypey@mastodon.nzoss.nz
It happened to me many times in the past as well. To be completely transparent, in this case I am even using suggestions from an AI with prompts such as “please do not make me sound rude or offensive”.
And honestly, it works surprisingly well for emails, long-form discussions, or social posts like these, where there is enough space to add nuance and clarify intent. I am less sure it would work equally well in a fast real-time chat.
The problem is that when you try to express technical opinions very clearly and precisely, especially in architecture or protocol discussions, people can sometimes perceive that clarity itself as being harsh, confrontational, or dismissive, even when no hostility is intended at all. Unfortunately, for reasons, I am not able to renounce precision or directness.
Replies (2)
-
@wronglang@bayes.club 2026-05-07 13:12
@uriel@x.keinpfusch.net @strypey@mastodon.nzoss.nz "make me sound more vanilla" is one of the few uses of LLMs that's reasonable and within design parameters of the models, assuming you still check the text for meaning.
-
@strypey@mastodon.nzoss.nz 2026-05-08 11:59
@uriel@x.keinpfusch.net > when you try to express technical opinions very clearly and precisely, especially in architecture or protocol discussions, people can sometimes perceive that clarity itself as being harsh, confrontational, or dismissive, even when no hostility is intended at all OMG this! ☝️ Although I find this when trying to express *anything* clearly and precisely in a text-only medium. Not just technical ones. @autistics@fedigroups.social @wronglang@bayes.club