Are you using GPT to author or edit your responses?
GPT can be useful for editing purposes, esp. for people writing in their second language or trying to find a more concise way of expressing themselves. (I'd often throw a phrase at GPT to see if it feels natural/idiomatic, instead of bothering my better half - an editor)
Having said that, if your goal was to make the responses here (on HN) more personalised, you achieved the opposite. Perhaps this is a sign of my personal paranoia, but I'm still not sure if I'm talking to a person or a bot. I don't mean to sound harsh, just thought it might be useful for you to know.
(FYI, I meant the responses here on HN. )
Your comments lack empathy and you aren't "reading the room" at all.
This is totally off-brand and sleazy if you're selling mental health advice. Particularly the fact that your responses don't demonstrate natural emotional variation, when you are literally selling a tool to help with emotion modulation.
From another language: “Translate this to English. Provide any clues you can get from idiosyncrasies in spelling, grammar, or word choice which might indicate which specific region or sub-culture the writer might be from and use this to provide notes on deeper or alternate meanings specific to that region or sub-culture. Note any potential ambiguities or uncertainties in translation and provide multiple options where appropriate. Use context clues to explain the tone, tenor, and emotion of the writing/writer, and what that commonly means in this culture vs. American culture.”
To another language: “Translate this to $LANGUAGE. Please make it sound natural to a native speaker, use local idioms or phrases wherever possible. Rather than strictly transliterating it, focus on conveying the specific tone of the original, but note where there may be a conflict between cultures and inform me if there’s a risk of being misinterpreted.”