3 ms·
Just so we're clear, this is using reverse psychology, right? They do want you to replace your comments with copyrighted text. I assume the wording is because
by xdennis 3y ago
Just so we're clear, this is using reverse psychology, right? They do want you to replace your comments with copyrighted text.
I assume the wording is because of legal ramifications. I wonder if such a defense works in court.
Personally, I think doing this is pointless. LLMs already use copyrighted works, so this isn't helping at all. The only way to tank Reddit is to add meaningless text which would make LLMs worse.
- CaptainFever 3y agoI know, right. The last thing I'd expect a self-proclaimed anti-capitalist to be defending is intellectual property, especially one of a corporation (NYT). If Reddit keeps a copy of the data edits, this move also just serves to hamper open source models who can only train on scraped data, while those with enough money can buy the full dataset with history. What I mean is, I agree and I think this plugin will do the opposite of what the authors expect.
- xyst 3y agoIn theory, Reddit won’t be keeping full edit histories of comments/posts. At some point, the edit history will stop recording and the original comment won’t be recoverable. Or maybe Reddit is flush with cash and keeping all of the edits. In any case, more edits means increased cloud bill.
- helpfulContrib 3y ago[dead]
- gorbachev 3y agoIt would potentially muddle the context, though. If every conversation about any topic has responses copying unrelated New York Times articles, what are the chances LLMs trained on that data will hallucinate even worse than before?