Thanks again Dan for being so patient and building one of the best online communities ever to exist.
https://www.newyorker.com/news/letter-from-silicon-valley/th...
https://news.ycombinator.com/item?id=7494093
https://news.ycombinator.com/item?id=7494440
bonus credit to DonHopkins https://news.ycombinator.com/item?id=18513120
Edit: @Doreen comment was not meant in any way derogatory, my apologies. Much respect for you, I’m on your Patreon and have learned so much from you [1]. I will be more thoughtful before replying.
Re the second bit: there aren't any accounts whose upvotes count for more, but if accounts upvote too many bad* comments and/or get involved in voting rings, we sometimes make their votes not count anymore.
* By "bad" I mean bad relative to HN's intended purpose as defined here: https://news.ycombinator.com/newsguidelines.html. Relative to that, "bad" means snark, flamewar, ideological battle, etc. — all the things that zap intellectual curiosity.
> is agreeing with comments killed by down vote really a red flag?
On the contrary, that's a good contribution and we hope everyone will do it when good comments (that don't break the site guideline) have been unfairly downvoted.
https://hn.algolia.com/?dateRange=all&page=0&prefix=true&sor...
p.s. It's always the good users who worry about these things!
The ongoing thread remembering Bob Lee is filled with that quality: https://news.ycombinator.com/item?id=35457341.
In terms of moderator action: we might downweight ChatGPT topics (for oar against) if they seem repetitive rather than significant new information (https://hn.algolia.com/?dateRange=all&page=0&prefix=false&so...). But we don't downweight posts that are critical of YC companies—or rather, we do so less than we would downweight similar threads on other topics. See https://hn.algolia.com/?dateRange=all&page=0&prefix=false&qu....
Are you sure there aren't abuses from your portfolio companies managers/employees to flag negative stories? I imagine Sam, for example, knows exactly what he has to do to get ChatGPT criticism guided off the stage.
Edit: for example, do you know what happened with this story? https://news.ycombinator.com/item?id=35245626
This is a very interesting/important topic. This was a new topic. It was really hot in the first hour, and just got smashed off the front page.
I wrote about this here: https://news.ycombinator.com/item?id=21868928 (Dec 2019)
Quite sure. That is, there may be managers/employers of $companies trying to flag things, but being a YC portfolio company doesn't make that any easier. And yes I'm sure that Sam can't do that. (I also know that he wouldn't try, but that's a separate point.)
Re the FAQ: it doesn't give a detailed explanation (we can't do that without publishing our code) but it summarizes the factors comprehensively. If you want to know more I need to see a specific link. Speaking of which:
Re https://news.ycombinator.com/item?id=35245626: it was on HN's front page for 4 hours, and at some point was downweighted by a mod. I haven't checked about why, but I think most likely it was just our general approach of downweighting opinion pieces on popular topics. Keep in mind that the LLM tsunami is an insanely popular topic—by far the biggest in years—and if we weren't downweighting follow-ups a la https://hn.algolia.com/?dateRange=all&page=0&prefix=true&que..., it would saturate the front page every day.
Actually we tend to not do that moderation on randomwalker posts (https://news.ycombinator.com/user?id=randomwalker) - because they're basically always excellent. But a certain amount of randomness is inescapable and randomwalker posts do great on HN most lot of the time. If we made the wrong call in this case, so much the worse for us and I'm genuinely sorry.
In general downvotes seem like a relatively poor feedback mechanism because there's no shared agreement on how they should be used. This [1], perhaps ironically flagged, post offered feedback on why people downvote, and it's just all over the place. Even if there are guidelines, people will be people. At least with something like clear adjectives, the percent of 'intended' feedback would be higher.
Thanks for confirming this. There was some speculation last year about partial shadow bans for voting,* and it's good to hear an authoritative answer.
* https://news.ycombinator.com/item?id=30317059
When this happens to an HN account, is it permanent or can it be reversed if the account stops upvoting "bad" comments? If it's permanent, affected users would like to know. Evaluating comments and determining whether they should be voted on can take a long time, and the affected users could save a lot of time if they knew that their votes would never count again.
Re "certain users": I don't see it that way! I'd prefer this to be a place where anyone can post about anything, and if their comment is insightful, that is what matters.
Re "US only": last I checked, US users were only 50% of the community. It may be less than that now.
Re brigading: it's definitely not allowed, and we've worked a lot on trying to stop it, but it's a hard problem.
Re non-technical discussions: I couldn't agree more, and we work hard to encourage that. Even to the point of various secret agendas.
Re sections or categories: no, I don't think that's in HN's DNA. For better or worse, this site is organized around a single front page that everyone sees the same way. Past explanations here if anyone is interested: https://hn.algolia.com/?dateRange=all&page=0&prefix=true&que...
Re politics: it's not possible to ban that altogether, nor would it fit HN's mandate of intellectual curiosity to do so. For past explanations see https://hn.algolia.com/?dateRange=all&page=0&prefix=false&so.... If there are questions I haven't answered there, I'd be happy to answer them.
For the philologist, from around 1780, The chapter of accidents: a comedy by Sophia Lee.
> Jacob: Dang it, doan't I zeay, I'll tell thee present...
> dang it, I thought at first I ha' zwounded...
https://archive.org/details/chapteraccident00leegoog/page/n4...
...But I also immediately think of Gary Larson.
There's no easy way to distinguish between "hacker" and "non hacker" stories because different people evaluate these expressions completely differently.
My friend, have you noticed in these few years that I believe in particular attention to words? :)
It was pretty easy: Etymonline, which is part of my few core references as "dictionary of meaning" (as opposed to "dictionary of use"), gives the first clues at https://www.etymonline.com/search?q=dang . Then I looked for the original text and checked the details.
dang explains most of the privileges / thresholds in various comments over the years:
Karma > 30 enables flagging and vouching privileges on posts and comments. <https://news.ycombinator.com/item?id=14697607> & <https://news.ycombinator.com/item?id=24134548>.
Karma > 500 enables downvotes on comments. <https://news.ycombinator.com/item?id=10224653> (Posts cannot be downvoted.)
Sufficiently negative karma causes an account's comments to be autokilled. Threshold not specified. <https://news.ycombinator.com/item?id=16409493> (Profiles can also be banned specifically by moderators, e.g., <https://news.ycombinator.com/item?id=35426426>.)
The 100 profiles with the highest karma are listed on the leaders page: <https://news.ycombinator.com/leaders>. Currently the 100th-ranked profile has 40,270 karma. There's no special privilege other than being listed, no member lounge or secret handshake.
(You can figure out most of this by searching for dang's comments with appropriate keywords, e.g., "by:dang karma threshold", which turns up most of these. I'd also looked for "vouch", "downvote", and "flag" specifically. <https://hn.algolia.com/?dateRange=all&page=0&prefix=true&que...>)
A List of Hacker News's Undocumented Features and Behaviors - https://news.ycombinator.com/item?id=33076053 - Oct 2022 (68 comments)
A List of Hacker News's Undocumented Features and Behaviors - https://news.ycombinator.com/item?id=30459276 - Feb 2022 (64 comments)
A List of Hacker News's Undocumented Features and Behaviors (2018-20) - https://news.ycombinator.com/item?id=26866482 - April 2021 (255 comments)
A List of Hacker News's Undocumented Features and Behaviors (2018) - https://news.ycombinator.com/item?id=23439437 - June 2020 (266 comments)
A List of Hacker News's Undocumented Features and Behaviors - https://news.ycombinator.com/item?id=20292361 - June 2019 (25 comments)
A List of Hacker News's Undocumented Features and Behaviors - https://news.ycombinator.com/item?id=19212822 - Feb 2019 (183 comments)
Hacker News's Undocumented Features and Behaviors - https://news.ycombinator.com/item?id=16437973 - Feb 2018 (391 comments)
https://news.ycombinator.com/item?id=35326541 certainly posted, and would have shown up on both /newest and /asknew. But it wouldn't have ranked on /ask (or not for long) because the upvotes on it were mostly dropped by our anti-voting-ring software. It looks like that was a false positive. I'm sorry! We don't know how to write anti-abuse code that doesn't have false positives.
Oh yes - it's probably the worst part of the job. But I think it's necessary for community because people respond differently when they're getting personal attention.
It's still my intention to build a canonical set of explanations for each common question and then mostly refer people to those. I've inched toward that over the years via HN Search links to my post moderation comments (which I know can get a bit annoying).
Even then, though, the mechanism will just be standard comments in ordinary threads, because that's how conversation takes place here, and people will always want to have personal conversation with the mods.
Users also tend to respond better when they get a detailed explanation of specifically how their post(s) broke the guidelines. Unfortunately, that sort of explanation is super expensive to produce—in time, energy, and stress. I don't have what it takes to do it in every case, which is a pity, because it tends to work. Here's an example from the other day: https://news.ycombinator.com/item?id=35403143.
Long term, we need better feedback mechanisms to let people know that they've broken the rules, and which rules, and for how long they might be in the bad dog box.
I'd be happy to take the penalty off your account - you've been a valuable and valued contributor for many years. Unfortunately I'm still seeing flamewar comments in your feed. I know that people have strong and valid reasons for posting that kind of thing but we just can't have that on HN, regardless of how right someone is or how legitimately they feel. It leads directly to this place burning itself to a crisp and preventing that from happening, or at least trying to stave it off, is our #1 job.
What I tell people in this position when they email us is that if they want to build up a track record of using HN as intended, they'd be welcome to email again after a while and we can take another look and hopefully remove the rate limit.
Can you link me to that?
> The world is filled with people/organizations who do the right thing almost all the time, but then use that clout to do a bad thing when it really matters.
That's a good point! but it's also an irrefutable charge. In fact, someone who behaved perfectly forever would be no less accusable of this. Btw I'm certainly not saying we behave perfectly—but we do take care to moderate HN less, not more, when YC-related interests are part of a story (https://hn.algolia.com/?dateRange=all&page=0&prefix=false&qu...). That's for reasons of self-interest as much as anything else. It wouldn't make sense to risk the global optimum for local gains.
> It sounds like all they need to do is ask one mod to take care of it, and it goes away pretty quickly.
People are going to feel like that's happening no matter what we do, but FWIW, we don't do that. We do downweight submissions as part of moderation practices that have been established for years, but a YC person doesn't have any more clout over that than you do, if you happen to email us and ask us to take a look at a particular thread (pro or con). And we always answer questions about what happened when people ask.
Btw if you feel like that randomwalker article is still relevant and can support a discussion of something specific and interesting—that is, not yet-another-generic-AI thread—go ahead and repost it and let me know, and I'll put it in the second-chance pool (https://news.ycombinator.com/pool, explained at https://news.ycombinator.com/item?id=26998308), so it will get a random placement on HN's front page at least for a while (how long depends on how the community reacts).
https://web.archive.org/web/20080115223854/http://ycombinato...
Dred's HN CSS Madhackery: <https://pastebin.com/gLXiqKyd>
Dred's HN CSS Madhackery -- Dark Mode: <https://pastebin.com/6PF3dCXH>
Recent local tweak highlights mod comments as well:
/* Highlight mod(s) */
.comhead .hnuser[href="user?id=dang"],
.comhead a:link.hnuser[href="user?id=dang"] {
color: #ff6600;
color: #dd5500;
}
The other HN superpower, available to anyone with an email account regardless of karma, is to email the mods. That's how I started interacting with dang, and still do generally a few times per week/month, mostly for mundane stuff (titles, link disambiguation), occasionally to vouch killed stories, or to point out spammish or hostile behaviour (rarely on the last, flags really do most of the work here).
(I'm familiar with ... all or nearly all of it, but a one-stop-shop for HN features and behaviours is quite useful.)
The points= parameter is quite useful, and one I'd either not known or forgotten. E.g., posts with >100 votes: <https://news.ycombinator.com/over?points=100>
You've unfortunately been posting some of these yourself - e.g. https://news.ycombinator.com/item?id=35429902. Can you please not? It's not what this site is for, and destroys what it is for, and we have to ban accounts that keep doing it.
If you'd please review https://news.ycombinator.com/newsguidelines.html and stick to the rules when posting here, we'd appreciate it.