16 ms·
Looks like ML-generated garbage with multiple accounts driving "traffic". HN is heavily infested with ML-generated posts and what appears to be account botnets
by voytec 2y ago
Looks like ML-generated garbage with multiple accounts driving "traffic". HN is heavily infested with ML-generated posts and what appears to be account botnets voting or commenting with LLM-generated output.
Setting "showdead" to "yes" in HN profile settings is the red pill.
I recently triggered[0] some LLM-based robot, by mentioning poor air quality in India in scope of reported cancer rates. You can read the first paragraph of all these "[dead]" responses and you'll get the pattern.
[0] https://news.ycombinator.com/item?id=41215564 https://news.ycombinator.com/item?id=41215564 (HN red pill - you need showdead=yes setting in user profile to see this spam)
- monkeyfun 2y agoWow, thanks for posting that, it was seriously impressive and surreal. I mean, even if it was entirely some random Indian poster with no automated assistance, which is still more than plausible, that speaks to a pretty shocking amount of dedication to shit things up that I usually only see on imageboards. Then add in the obvious signs he tried to use AI at at least some stages, and that's where it gets actually wild to me, because if the posts were even slightly: - more coherent - less irrelevant - less duplicated - timed a bit further apart I can totally see most undiscerning readers never assuming that was all one person or arguably less than one person with automation.