3 ms·
This content is oddly similar to what one can read at [1]. So much so that it feels like plagiarism. There's no reason LLMs don't also do this with code by the
by jraph 1y ago
This content is oddly similar to what one can read at [1]. So much so that it feels like plagiarism.
There's no reason LLMs don't also do this with code by the way.
[1] https://emojipedia.org/seahorse-emoji-mandela-effect https://emojipedia.org/seahorse-emoji-mandela-effect
- ruszki 1y agoWhat if it’s the other way around? Your linked article is based on AI slop? Nowadays, you cannot be sure.
- lsaferite 1y agoHaving read the above conversation excerpt and the page you linked... how do you get to it feeling like plagiarism? Given a constrained set of information here, there's only so many ways to present the information. They roughly discuss the same data points, but the writing is different in both. Is this disallowed?
- jraph 1y agoTake the "seahorse sticker or emoji-style image" part. It's the exact same words. It is quite specific phrasing. There's no other result on the web for this. It's not only this part, the whole thing feels very similar, just with more confidence and in a sublty different order. When the set of ideas and the phrasing are this similar without citing the source, I call this plagiarism. Now, it could be that the emojipedia page is AI slop, but then the AI would seem to have invented this specific phrasing, which I would find unlikely.
- astrange 1y agoIt could also be that the UI shows citations but they don't survive copy-paste.
- Grimblewald 1y agonot only that, but, if a LLM can actually complete a code task it's a good sign that that problem has a public code repository solving it (usually in a way that is better than what the LLM offers)