For example, if I quote a GenAI response –even in criticism– (See what I did, there?), it can get flagged, and result in a shadowban (has happened to me -lesson learned).
But a good use for LLMs, is as a copyeditor. They do a great job. Some unedited stuff is so bad, I'd rather read slop, any day.
The problem is, what's the threshold? If they just fix a few typos and misspellings, that's fine, but what if they offer more substantial changes? How much text must change, before we can legit dismiss as "slop"?
Also, what if there's a significant GenAI component, but the article really is something that we want on the HN frontpage, because of its content?
For example, if I quote a GenAI response –even in criticism– (See what I did, there?), it can get flagged, and result in a shadowban (has happened to me -lesson learned).
But a good use for LLMs, is as a copyeditor. They do a great job. Some unedited stuff is so bad, I'd rather read slop, any day.
The problem is, what's the threshold? If they just fix a few typos and misspellings, that's fine, but what if they offer more substantial changes? How much text must change, before we can legit dismiss as "slop"?
Also, what if there's a significant GenAI component, but the article really is something that we want on the HN frontpage, because of its content?