I don't want to say don't waste your time since the first half is human written. Questions for the authors: did y'all just get tired of writing and said "fuck it let's have the LLM finish the rest"? Or did one of you use LLM to write the last half and the other used their own words?
While I am not without my critiques of the article (such as the lack of any measurements), I still find the ideas of interest (such as AST based chunking).
try feeding it content already in the corpus - niche religious texts are a good one where they've tried to special case out the mainstream things like the King James but missed other ones. They claim it's 100% AI.
duhhhhh1212 · · focus · HN ↗
I don't want to say don't waste your time since the first half is human written. Questions for the authors: did y'all just get tired of writing and said "fuck it let's have the LLM finish the rest"? Or did one of you use LLM to write the last half and the other used their own words?
verdverm · · focus · HN ↗
1. detectors are unreliable, humans are "writing like ai" now (first we shape our tools, then they shape us)
2. more people don't care as long as the content is quality
3. there is a spectrum of ai-human writing and how people make use of the tools, some good, some NS;NT
<a href="https://news.ycombinator.com/item?id=47089907">https://news.ycombinator.com/item?id=47089907
---
While I am not without my critiques of the article (such as the lack of any measurements), I still find the ideas of interest (such as AST based chunking).
duhhhhh1212 · · focus · HN ↗
If you can prove pangram is unreliable I will stop using it. By prove, I mean show their false positives and false negative rates are made up.
> more people don’t care as long as the content is quality
That’s your opinion. If you want to read AI generated content then there are wonderful sites that can produce that within seconds.
hactually · · focus · HN ↗
Garbage app In my experience.