|
|
|
|
|
|
by inanutshellus
689 days ago
|
|
|
I'd guess it's here because their algorithms determined high-odds of generating undetectable-as-bot/relevant content. Looking at the two examples, they primarily stand out because there are two almost-identical ones, not because they're so obviously bot-generated that I'd go out of my way to flag'm. So I'd say they're doing a pretty good job already. And quickly you start to question your sanity. Like... is this post (https://news.ycombinator.com/item?id=41076632) bot-written? it's written a bit oddly, could it be a bot? Who's to know? The famous "firehose of misinformation" is hitting new high scores in the game of modern civilization. |
|
|
i know that this is probably a very bad approach, but until now it worked out - if a text contains grammatical errors i automatically assume that it was written by a human, as AI-content always sounds so polished, boilerplate-y and corporate. but you can probably tell your AI to integrate errors, ignore capitalization and miss out on some punctuations.
I asked ChatGPT to reply to your comment and to include some errors, leave out some punctuation etc., that's the result:
oh, i see what you're saying. yeah, it's really hard to tell if something is bot-written these days. like, even reading your comment, i start to question if i'm talking to a human or not. the examples you mentioned do stand out, but more because they're so similar, not necessarily bot-like. and you're right, the "firehose of misinformation" is just getting worse and worse. it's a strange new world we're living in.
i think we're cooked.