HN Leaders — What are the most upvoted users of Hacker News commenting on? | Hacker News Reader

HN Leaders — What are the most upvoted users of Hacker News commenting on?

hn-leaders.herokuapp.com

74 pointsdrpancake61 comments

What I enjoy about HN is that for most part we ignore who is saying something and instead focus on the content. This is in contrast to a site like Twitter, where the number of followers and blue checks make a massive difference to the number of people who will see your thoughts. This ends up creating a system that feels fairer. I don’t need to waste time cultivating an audience and farming clout, for instance. I can simply lurk and comment on the rare threads where I know what I’m talking about. I don’t care about my karma, and no one else cares about my karma either.

The exception to this is links to personal blogs, where links to domains of well known people immediately rise to the top. This is a simple question of time - it takes time to evaluate full articles and people are more likely to invest that time if the name seems familiar.

This is my take too. The asynchronous nature of HN takes some time to get used to, but it organically brings diversity to every discussion. I've seen insightful comments that make me want to follow a user, but in the end I'm glad I can't because it'd reduce diversity.

Other social networks focus too much on following influencers, which naturally narrows your horizon to whatever the influencer thinks. People who follow influencers to criticize them are still narrowed to the direct opposite of what the influencer thinks. It's the monetization of bugs in our psyche.

I just find that it's more a reflection on how soon you post on a topic, rather than how much insight you provide. A fantastically insightful comment #100, is not going to get a fraction of the upvotes of a mediocre comment #10 on a post. That's because of the nonlinear exposure. If you post early, you will get more reads which leads to more upvotes which will puts your comment near the top. Almost all my comments with more than say 20 upvotes, are "early" rather than "great".

What I'd like to see are average upvotes of buried comments. Comments that were highly rated based on their content rather than their exposure.

Sadly, I've gotten more upvotes on a nasty quip towards someone than I have on any technical subject I've put effort into writing on.

I notice this too, although I don’t think it’s quite as bad on HN as on the internet in general. I dislike the “sick burn” comments and wish they weren’t so popular.

I've seen multiple people post similar things on this thread. I actually have the opposite experience. Most of my most upvoted comments tend to be multiple paragraphs with many citations

I guess the difference is I don't make these types of posts on software-engineering heavy discussions. I've been a Wikipedia editor for a few years and I have a lot of niche interests in botany, evolutionary biology, anthropology, microbiology, etc.

There are certain non-software/hardware topics HN seems to really like (phages, the gut microbiome, and nootropics come to mind). It seems that making a long comment with a bunch of citations related to one of those topics is more likely to get you upvotes. Probably because its something a lot of people will be interested in but you're much less likely to be disagreed with given the lack of community expertise around the topics

Maybe my posts wouldn't do as well on a botany forum haha

A decade ago any comment that was less than a couple paragraphs would be almost instantly down voted. Sadly that is the case today. Threads are a lot noisier, and I even see some bad joke reply threads appearing here and there. A dramatic shift in culture.

Similarly, discussion is heavily skewed toward the first 5-10 comments on a post. If I think something is really interesting but it already has 100 comments, I usually don’t bother posting as I probably won’t get replies (which is the interesting part of commenting IMO). The sweet spot is a new article that I find very interesting with less than 25 comments. Then I often want to comment because I know it’ll appear close enough to the top that others will reply and discuss :)

On the contrary, the sorting algorithm accounts for both votes and time. Whenever I manage to provide good thoughts from a fresh perspective on "old" (by your account) discussions my comments have stayed at the top despite being late.

Actually, I haven't really found this to be true. The HN algorithm does some sort of thing where it spotlights new comments at the top of the stack to see how well they perform before pushing them down. There have been a number of times when I was very clearly late to the conversation but my comments apparently got enough upvotes when they were spotlighted to stay on top.

It feels to me like the comment sorting is actually dictated by a rough calculation of impressions / votes rather than just raw votes, which is why this sort of thing can happen.

Completely agree. I think to really show the best comments in a thread, comments should be weighted higher based on how recent they were made. At the same time though, you don't want to weight them too highly and flood the top comments with new, less notable comments because that would seriously deteriorate the quality of the top comments section.

I find HN strikes a really good balance here, occasionally putting new comments at the very top, but not enough to flood the top comments. It's much better than Reddit at least. HN also seems to give additional weight based on the length of the comment which filters out many of the joke/quippy comments and gives preference to comments which add a lot to the discussion.

I agree this effect exists, but much less than a comparative site like Reddit. One thing HN does well is that a new comment appears at the top of its sub-thread and stays there if it gets enough upvotes.

This has happened a couple of times with my comments. Once there were at least 200 comments on a hot button issue but my comment rose to the top anyway.

I didn't know that. But that I guess makes it even worse for me: I have had many highly upvoted comments, but none that weren't early. So in conclsion then, all my highly rated comments were highly rated because they were early...

A gem from this article: "Minecraft would probably never have happened, if Notch started by asking in the forums if he should use Java or not."

Most things wouldn't happen if the creator asked ahead of time what they should do. This is the dark side of peer review and relying on peers. It leads to a kind of mental / creative vetocracy situation where ideas are killed in their most fragile state by casual comments.

Vetocracy is a real thing. It plays an important (devastating) role in scientific progress, even there is a saying about advancement at funeral-based time steps, making a similar point:

Science advances funeral by funeral. Science advances one funeral at a time. Science progresses funeral by funeral. Knowledge advances funeral by funeral. [1]

The quote is attributed to economist Paul A. Samuelson[1].

[1] https://quoteinvestigator.com/2017/09/25/progress/

- Wikipedia page for Vetocracy:

https://en.m.wikipedia.org/wiki/Vetocracy

- Notable prior HN coverage of topic:

https://news.ycombinator.com/item?id=29625988

Perhaps is there a better term to describe vetocracy in science? Wikipedia here just refer to it's application to democratic institutions.

Likely better to generally the pattern than narrow it, since more surface area it covers, more likely people are to want to understand it.

For example, in my opinion, pattern aligns to phrase "not in my back yard" — where ultimately, like in science, if a subject matter is vetoed as being off limits, scientists just go somewhere it is not:

https://en.m.wikipedia.org/wiki/NIMBY

Vetocracy and NIMBY under the same conceptual umbrella? I need to think about it, I see the semantic fields have non-intersecting areas that can be contradicting. E.g. how saying no to externalising pollutants to rural areas is a vetocracy?

I suspect this is true but not necessarily for the reasons you suspect; that a person who is asking what to do is often not the type of person who does, and the second type is really what you need to get things moving.

I really like this. One peculiarity about HN culture I found is that commenters will often comment something tangental to the subject of the post, so I often find myself reading really insightful comments on threads I otherwise have zero interest in. This is cool because it surfaces all those cool sometimes totally tangental comments.

> One peculiarity about HN culture I found is that commenters will often comment something tangental to the subject of the post, so I often find myself reading really insightful comments on threads I otherwise have zero interest in.

Four days ago I was reading a thread I did have interest in about hiring freezes at Google and Facebook, and there was a tangential thread, which was even more insightful, between the Linux maintainer for ext4 (the Linux filesystem probably 99% of Android devices use, and a high percentage of Linux servers too), and another Googler/Xoogler about whether the Google storage engineering team could make non-incremental contributions, and which contributions were worthwhile and which weren't ( https://news.ycombinator.com/item?id=32323734 ).

That's why I read HN comments. You never know when they'll go on a very interesting tangent. For very loose values of tangent.

Also, the "{person involved} here! {Comment full of insider information}."

One could probably compile a book of great comments in this vein that have appeared on HN over the years.

Five plus years ago I wrote a chrome extension that was like an RSS feed for people on HN (still works).

Basically, you followed users and it would create a "news feed" which would show posts from users and comments (with links to context). It would also highlight interesting comments from those users on pages I visit. Similarly, I could block users. Let's me follow people I find interesting. Looks very similar and with some tweaks could be public.

Very interesting, but giving them the title ‘leaders’ seems a bit.. overbearing?

Do not get me wrong, there is some very interesting stuff there but half of gathering chunks of upvotes is not rocking the boat and saying anything that goes against the echo chamber.

You'll be downvoted for saying this.

EDIT: just as I'm being downvoted. The echo chamber cannot be called out.

EDIT: the guideline of not discussing downvotes hardly applies here!!

HN guidelines

>Please don't comment about the voting on comments. It never does any good, and it makes boring reading.

edit

>the guideline of not discussing downvotes hardly applies here!!

Why? Your entire post was "You'll be downvoted for saying this."

I think they meant it doesn't apply on this entire page. Which seems reasonble, since the whole thread is about posts ranked by commenter's karma, which is a direct reflection of votes. How could "not discussing downvotes" be an appropriate guideline in a discussion specifically about comment votes?

It's named after the karma leaderboard page: https://news.ycombinator.com/leaders

> not rocking the boat and saying anything that goes against the echo chamber

I think there may be something in that.

My most-upvoted comments have been (a) comments to the root article, not sub-comments; and (b) generally rather short and 'flip'. Apart from that, I can't find any pattern in which comments get upvoted and which don't.

I don't (I hope) post echo-chamber comments; if everyone already agrees with me, there's no point in posting.

I post what I think are insightful comments, that stick on 1 forever; and comments that I know are controversial, that scoop 20 upvotes in 10 minutes. I also occasionally post vacuous, "witty" one-liners that, to my surprise, garner lots of upvotes.

I don't post for karma; I post for sub-comments, i.e. I want to discuss stuff. I don't like long comment threads on HN though; so I never post more than one or two replies to my own thread.

Would be cool if time were taken into account, e.g. define "leaders" as most upvoted users this day / month / year.

I'm conflicted about this, in theory good comments and insight can come from anyone. In practice, reading some comments on at best waste your time, and at worse make you dumber.

Information is like food for your brain. But with information, you can't tell what's nourishing, what's junk food, or what will make you sick until after you've consumed it.

Interesting idea. Kinda cool to explore topics based on other peoples comments.

I wish "average karma" was still a metric on this site.

I think it helped provide a gauge to help people decide between leaving valuable comments vs. just leaving as many comments as possible.

Early comments aligned with the echo chamber are the most upvoted. Little to do with 'value'.

Yes, but at least it prevented some people from leaving followup comments with little value.

e.g. - that feeling of having to respond to the person that responded to you, rather than giving the community a chance to respond.

worked for me at least.

I'm not saying that this is necessarily good, but it is kind of expected.

People click on threads which are interesting to them, according to their topic. If you're interested in a topic, you most likely view it positively, or as you would put it, you're part of it's echo-chamber. If you then see a positive, reaffirming comment, you'll probably like it.

Doesn't seem that bad to me, especially because I frequently see that the most upvoted reply to the top comment are often contrarian, starting discussions etc.

Echo chamber is such an overused word. In other words, you define that people are part of an echo chamber because you personally don't agree with the majority, which makes you a bit of a contrarian, reactionary hater ;) (exaggerating a bit here)

Isn't the "karma" calculated from both upvotes on your comments and your submissions? If so this could easily end up being a pretty bad metric for deciding which users leave valuable comments - a couple of front page stories will send your score through the roof.

Additionally, I think simple 1-scoring comments don't contribute to your karma - it's when they get upvoted that they count. So spamming a bunch of useless stuff won't help, and to be honest people are likely to see and start downvoting/flagging you.

It's probably not worth caring about the points you get, as long as you're not saying a bunch of abusive stuff and getting flagged or shadowbanned :D

Edit: confirmed my suspicion re "1 point" comments not adding 1 point to your karma: I was on 8700 before hitting "reply" and I'm still on 8700 immediately after.

The "karma" value seems to climb with number of comments, only it's calculated with a lag. I think that's why your "immediately after" check showed no change. Now that it's been a couple weeks, can you recall if you hit 8701 in the half-a-day or so after posting the above?

At this point, a lot of the leader list are accounts who have posted a lot over a very long period. I don't give them any special weight, and in some cases I give them less weight due to how much they post.

I think this is neat, but i don't honestly care what the comments of these random (to me) people are, nearly as much as I care what posts they're commenting on.

An almost guaranteed way to get a highly upvoted submission on Hacker News is to submit something about Hacker News...

This may reduce the quality of the posts as those with less try to emulate those with more.

"leaders"? More like the inner sanctum of the echo chamber.

EDIT: If you look at historical influential figures most of them were controversial.

That's not surprising... "leading" society means leading it to a new place, not keeping everything as it is.

I doubt I count as anywhere close to the echo chamber (#65 right now).

On a side note, I wonder how many people on HN have actually met someone they recognize from HN?

Probably pretty rare since so many people don’t put detailed or full info of who they are. Or am I off here?

As someone with that much karma, have you became friends with any one on HN? I realize it is exceedingly difficult to do without private messages and such.

Not really. I think you hit it on the head as there is actually no contact on HN except replies. I know a couple of handles that I either really like or don't, but have no personal knowledge of them. I do admit that I've never been able to go to one of the meetups people have done.

The only person I know I have met was really secondary to HN. I met Bryan Cantrill (bcantrill) at Systems We Love in Minnesota. I sent him a t-shirt from the college I worked at.

Thank you for that, BTW -- that shirt had barely hit the drawer before my teenager stole it from me (curse of having kids the same size!) and I know that it has ended up in his rotation. It was great meeting in person in Minnesota, and thanks again for the shirt!

Glad to hear it. The three wolves shirt hasn't been printed in a while (COVID killed team sports). It makes me chuckle a kid out there is wearing a shirt from a TCU on a ND reservation. We must be big time.

People may be surprised how many are not in the echo chamber, or peer echo chambers.

patio11 showed up as a speaker at an event, and the only reason I'd heard of him at the time was HN.

I might be confused, but, is leadership in a social network same as (degree) centrality in this network? The leaderboard could be just a result of the centrality of the topics. What happens when other people touch these topics?

Love to follow what's interesting people are sharing

When I see highly upvoted users, I tend to recognize a lot of them as people posting wild and unfounded speculation, drama, conspiracy theories, etc. I'm not terribly interested in those views.

The illuminati made you post this.

> ...people posting wild and unfounded speculation, drama, conspiracy theories, etc.

One percent[1] of crazy-arse conspiracy theories will probably turn out to be true: Conspiracies do exist. But since you don't know which one percent that is, dismissing them all is still the rational thing to do: If one percent[1] are true, that still means ninety-nine percent[2] of them are total tinfoil-hattery.

___

[1]: Or two percent... Or 0.1, or 0.01.

[2]: Or, correspondingly to [1], 98 %. Or 99.9, or 99.99.

Is the domain herokuapp.com commonly use for botnets/phishing etc? None of my work networks allow it.

The top voted comment on a post is considered the "best" opinion or thought that the thousands of viewers & commenters had in response to the subject matter.

Sometimes it is a bit disappointing that some bland platitude or parroted opinion of the week is rated as the most insightful thought our collective brains could muster.

I upvoted this in the hope it gets to the top and becomes deeply ironic.

Time for a new hobby maybe?

Certainly for me, it is.

Reply on news.ycombinator.com