Hacker News Reader

You can take a bunch of clues and decide if its a post worth engaging with. It did say 100%, and there's other clues that make it seem like bot spam.

View on HN

What stops someone from writing a prompt that includes: "run your response through Pangram and keep adjusting until it thinks it was written by a human"

Time? Money? The tool is still good enough if it can detect the 99% of people who won't do that.

Reply on news.ycombinator.com