3 ms·
There are topics HN has extreme difficulty in discussing rationally, one of my prime disappointments with the site.[1] There are also clues in HN's moderation,
by dredmorbius 2y ago
There are topics HN has extreme difficulty in discussing rationally, one of my prime disappointments with the site.[1]
There are also clues in HN's moderation, in guidelines, by members, by mods, and by automated processes, which may give some hints as to how to best address this.
HN guidelines include a number of elements which might be used against disinformation / denialist comments, specifically:
- Comments should get more thoughtful and substantive, not less, as a topic gets more divisive.
- Please respond to the strongest plausible interpretation of what someone says, not a weaker one that's easier to criticize. Assume good faith.
- Please don't post shallow dismissals.
But most particularly these two:
- Eschew flamebait. Avoid generic tangents. Omit internet tropes.
Virtually all denialist points are in fact generic tangents and Internet tropes. That they're also by all appearances based on shallow (if any) reading of the specific article in question is all the more damning.
- Please don't use Hacker News for political or ideological battle. That tramples curiosity.
I've discussed both these two points in emails with dang, and suspect that applying them has significant traction.
The out-of-the-woodwork aspect of this suggests that accounts are engaged in ideological battle. I'd watched this post as, early in the US/SV day, multiple shallow takes and tropes (most though not all now strongly downvoted or flagged) appeared on this topic. Long before any substative and science-backed discussion appeared. The one apparently substantive denialist take was from a twenty-year-old study which, for some reason, omits much of the past twenty-years alarming increase in flood frequency and severity (now flagged comment: <https://news.ycombinator.com/item?id=41646319 https://news.ycombinator.com/item?id=41646319>). I'm seeing both "green" (recently-created) and some rather familiar accounts expressing denialist views. And yes, I've been tracking some of the less-substantive / useful contributors to HN for a while now. It's a small set but they can effectively steer discussion, especially if posting early to a thread.
Member flags and voting contribute somewhat to post and discussion quality, and here again on contentious topics what often seems to be occurring is a great deal of strongly motivated voting and flagging activity. I may partake of that myself, though I believe, with some basis in underlying science and facts, that my own actions are strongly justified. (I am not in the denialist camp, should that not have been sufficiently clear already.) That's something of a given, though mods can take actions on flags to some extent.
Mods will wade in sometimes, though for the most part their time and attention are quite limited. They may be summoned by email to hn@ycombinator.com. Responses generally occur, though I'm noting these are often many hours delayed of late.
Automated moderation applies to specific sites (most general news sites are modestly penalised by HN, they remain among its most frequent and highly-voted sites), and more specifically for this discussion, the "flamewar detector". This is a heuristic that posts with > ~40 votes with a comments:votes ration > 1 (<https://news.ycombinator.com/item?id=36926236 https://news.ycombinator.com/item?id=36926236>). I've suggested "spiciness" as an alternate name for this. Whilst it does work pretty well in many cases, it seems to unfairly penalise precisely the highly-contentious threads which HN already has difficulty in discussing. I'd argue that that characteristic cuts against HN's prime directive, of stimulating intellectual curiosity (<https://news.ycombinator.com/item?id=26998308 https://news.ycombinator.com/item?id=26998308>) and [T]o make thoughtful comments. Thoughtful in both senses: civil and substantial" (<https://news.ycombinator.com/newswelcome.html https://news.ycombinator.com/newswelcome.html>).
If anyone could suggest ways of distinguishing truly flamey threads from those which are substantive-but-spicey, I'd love to hear them, and I suspect mods might as well. What's most frustrating is that the trolls, flames, disinfo warriors, etc., all manage to frustrate substantive discussion. I find it's best not to engage significantly with them, other than to provide a single succinct rebuttal, e.g., pointing out that an article claiming no increase in high-water events dates back two decades.
I also suspect that dang gets frustrated with the disinfo aspect as well.[2]
One of the best ways to counter thread-derailment, which submitters can undertake, is to post a thoughtful, considered, comment which highlights a key aspect of an article. HN is reactive, and such comments tend to get upvoted and, though drawing critical "no, but" responses, also typically draw "yes, and" responses which further emphasize key points, facts, arguments, and related materials.
Posting such a comment even if apparently after a thread has derailed can also perform surprisingly well. I've adopted this practice to some success. Both are far more effective than complaining about behaviour.
Oh, and email the mods. I'll be doing so after posting this comment, linking it as my broader complaint. I do believe there's some opportunity for improvement here.
________________________________
Notes:
1. A couple of earlier mentions in a recent comment (<https://news.ycombinator.com/item?id=41569514 https://news.ycombinator.com/item?id=41569514>) and eight-year-old post (<https://news.ycombinator.com/item?id=19694881 https://news.ycombinator.com/item?id=19694881>).
2. As evidence, Algolia search of "by:dang frustrating", which turns up quite a mix of comments, though some touch on this element: <https://hn.algolia.com/?dateRange=all&page=0&prefix=false&query=by%3Adang%20frustrating&sort=byPopularity&type=comment https://hn.algolia.com/?dateRange=all&page=0&prefix=false&qu...>.
- fragmede 2y ago> If anyone could suggest ways of distinguishing truly flamey threads from those which are substantive-but-spicey, I'd love to hear them Taking threads that trigger the flamewar detector (simple comments vs points comparison) and then running them (either individual comments or the comments as a whole) through an LLM to score them for flamebait seems obvious so my question is why didn't that go anywhere. If you didn't want to use new technology, you could still just build a classifier based on the usernames involved. It wouldn't have to be perfectly accurate, just enough to flag something for manual review.
- dredmorbius 2y agoThat's one thought I'd had. There's a simpler heuristic which was developed at Microsoft back in the early aughts, based on Usenet, which simply looked at thread topology to make inferences about the type of exchange. I'm pretty sure that was Marc Smith (the name was a very common one, that or Jones was in my memory), see: "You Are Who You Talk To: Detecting Roles in Usenet Newsgroups" (~2005?) <https://www.microsoft.com/en-us/research/wp-content/uploads/2016/02/YouAreWhoYouTalkToHICSS2005.pdf https://www.microsoft.com/en-us/research/wp-content/uploads/...> (PDF) (Date based on citations and recent data (2004) mentioned in article. Why aren't academic articles internally dated?) NB: That's one paper I've found, not necessarily the most applicable, though it points in the right direction. I'll update (edit or reply) if I find something more appropriate.