Hacker News Reader

Comment by Tiberium | Hacker News Reader

Okay, I have a test for you. It's a bit undocumented, but I'm pretty sure HN is using Pangram or a similar tool to auto flag AI written comments. Can you copy the entire text from the linked article and reply with it here, and see what happens to your reply, would it get flagged or not? I think this will show you what reality we live in.

View on HN

Pangram is just some BS man can you please stop

Can you actually just do the test I explained? You seem to be basing your evaluation of Pangram on some old experiences. It's a very reliable service, used by research conferences and universities.

It's baffling that people would rather call BS on me than try the actual tool being talked about.

It is not a reliable service. The CEO likes to say that it failed 1 time out of 10,000. In Panagrams own research the failure rate was shown at 2%. That's a pretty massive difference.

How about you do this, write your own words on a topic, but use AI-isms in it. A few 'not X but Y' an a few EM dashes and watch it say you're text is AI written.

Writing using AI stylisms has nothing to do if a text is actually written by AI or not. The tests are mostly bullshit because just adding some informalisms in the text corpus tends to get it counted as non-AI. It's not detecting what you think it is.

Who cares about Pangram? Why do you care about Pangram? Why not leverage our own intelligence? How does this post read to you? To me, it does not read like it was written by AI.

Reply on news.ycombinator.com