pull down to refresh
reply
tbh I think this is a rosy rationalization, ie what I hope I'm doing. So in an effort to fight that, here's a less charitable guess:
- have I seen this nym before
- is the content deeper than average but in a kind of generic, illusory way
- is the content eerily consistent with their past content
- can I find human failings that should cause me to question my judgement so far
I saw you all get caught. I think what I look for is how derivative something sounds. Human or bot I don't like derivative content. If it's something I can imagine many people saying - I don't need you to say it because I can imagine it. It's hard to generate novelty so bots tend to only produce derivative tokens, no matter how many tokens they produce. Yet it's relatively easy to verify novelty.
This will probably change but it's true now: bots, without very long chains of reasoning prompted to discover novelty and criteria to verify it (which is not easy to come by), say derivative things.
What I try to be very sensitive to when reading content now: surprise. My surprise in content, grammar, tone, formatting, etc.