Out of curiosity, I submitted this prompt to both ChatGPT and Claude. I have always kept history disabled on both.
ChatGPT jumped into analyzing my personality based on the system prompt, and when challenged about the behavior it should have access to being only the one line question, it prevaricated.
Claude looked for the directory of transcripts, found it empty, told me it couldn’t answer the prompt given current information, and offered a few ways forward if I wanted to get an answer to the prompt.
Seems the post's been deleted.
https://web.archive.org/web/20260213222727/https://old.reddi...
This is insane! You could probably stop the interview at this point and request a conversation with a higher up person. Just on the off chance you are being interviewed by an idiot and that their job may soon be available. But I’d never want to work there.
Wildly inappropriate and probably illegal behavior.
Even if it’s about a protected class, it’s rarely illegal to ask - just a terrible, lawsuit baiting, idea.
Out of curiosity, did you enter this prompt later privately? Were you comfortable with sharing the results?
I saw this on Reddit last night. I shared the results of this prompt with my cofounder, he replied in kind.
I thought this was an interesting concept. It feels a little dystopian in a corporate context but much less so in an early stage startup.