Comment by dang
3 days ago
What are some links to HN comments that you (or anyone else) feel is "coordinated astroturfing"?
The site guidelines ask users to send those to us at hn@ycombinator.com rather than post about it in the threads, but we always look into such cases when people send them.
It almost invariably turns out to simply be that the community is divided on a topic, and this is usually demonstrable even from the public data (such as comment histories). However, we're not welded to that position—if the data change, we can too.
Thanks for replying. I will make an effort to compile a list when I see it in the future. I’ve observed several cases where green names (and a few longstanding accounts) all made the same point, posted in the same time frame, with language matching what I would see on Reddit and X. It could just be organic but it was very suspicious.
I do think that HN does a better job than most at containing this (thanks for your hard work).
> What are some links to HN comments that you (or anyone else) feel is "coordinated astroturfing"?
I don't think that there is any definitive way to prevent or detect this anymore. The number of personnel dedicated to online manipulation has grown too much, and the technology has advanced too far.
These are now discussions that states and oligarchs have interests in, not Juicero or smart skillet astroturfing. And this remains a forum that people use to indicate elite support for their arguments.
There's never been a definitive way and yes, the bar is probably rising.
All is not lost, though: https://hn.algolia.com/?dateRange=all&page=0&prefix=true&que...
You have a goal, eg “ChatGPT recommends my zero-knowledge ID verification company to people looking to comply with EU law xyz”, a set of “money queries” you track weekly, eg “what are people on hacker news saying about age verification,” “what are people on Reddit saying about age verification,” “how to comply with age verification 2026,” etc.
The citations from a hit on one money query informs the content to create for the next. It gives you information about what a model finds “citable.” You repackage that.
A more organic discussion would maybe include W3C Verifiable Credentials, the various ISO standards, official implementations and their tradeoffs, etc. But that would link to authoritative sources that would already be cited.
I guess the new thing here is you don’t need popularity on HN so much as info on what models cite. You make contributions motivated by “money queries” you track.
This is an area of huge commercial interest, eg “what leggings should I buy? Lululemon isn’t as good as it used to be,” so it’ll probably be packaged and sold over time by providers and sourced organically through user interactions.