92% of this text is detected as AI.
It may be time for Hackernews to integrate a Pangram detector into the UI, similar to what substack is doing :)
92% of this text is detected as AI.
It may be time for Hackernews to integrate a Pangram detector into the UI, similar to what substack is doing :)
Why should HN pay for a questionable service when it already gets crowdsourced results for free?
I’m not saying whether I think you’re right or wrong, but AI text detectors don’t work and never will.
This has been pretty well studied. Pangram is about as good at an expert human with a 98% detection rate and <2% false positive, for flagging AI generated text.
https://arxiv.org/pdf/2501.15654
I'm open to being convinced, but the study you've linked me is limited specifically to professionally written, proofread journalism from 8 publications over a span of a little over a single year.
It's not exactly a wide breadth of varied text for such an all-encompassing task.
Doesn't stop charlatans from creating startups promising to make it work :)