Hacker News Reader

Comment by thewebguyd | Hacker News Reader

We also need to stop immediately dismissing certain writing tropes as being automatically LLM generated.

I'm re-reading an older fantasy series, and the amount of em-dashes in the text and similar tropes (you know, normal good writing) would easily be called out as "AI Slop" on this very forum, even though it was written in the early 90s.

View on HN

Look at even older writing and you'll see em-dashes there too: Edgar Allan Poe, Lucy Maud Montgomery, the Brontës, among others.

> “No, reprobate! you are a castaway—be off, or I’ll hurt you seriously! I’ll have you all modelled in wax and clay! and the first who passes the limits I fix shall—I’ll not say what he shall be done to—but, you’ll see! Go, I’m looking at you!”

-- Wuthering Heights (1847), Chapter 2, Emily Brontë.

I wonder what the frequency of AI common words like delve and other words/phrases are in Project Gutenberg, Wikipedia, and other collections.

I think that's fair. Em-dashes are pretty innocuous and I think technically correct regardless. What bothers me is all the filler in AI writing that just makes it longer and harder to read. I'm actually not that bothered by the idea of AI writing, but don't waste my time and patience with all these "not X, but Y" digressions thrown in for flavour. Get to the bloody point.

I suspect the filler comes from RHLF and the amount of web marketing content that was ingested. "Not X, but Y" is all over marketing copy, as is the triad of concepts "By doing X, Y, and Z...." because it mimics TED-talk style rhetoric, it gets rated highly because people think it sounds profound. So optimizing for what appears to be helpful ends up just creating a bunch of corporate-speak sounding garbage.

Plain, direct language gets selected against during post-training leaving us with slop.

Reply on news.ycombinator.com