Hacker News new | past | comments | ask | show | jobs | submit login

We don't ban sites of competitors of companies we fund. Even if we wanted to do something like that, how could we ever get away with it? I'm guessing you're referring to Nodejitsu.com. They're banned because they created an army of sockpuppets to vote up their posts.

Your claim that Nodejitsu "created an army of sockpuppets to vote up their posts" is outrageous. Let me enumerate the issue here:

1. Your moderators are simply unable (or unwilling) to distinguish "sockpuppets" from a legitimate community of developers (that is the node.js community). Yes, when we post something to HN we send out the link via Twitter and IM to colleagues in the node.js community who may potentially upvote it. This is natural "word of mouth". We have never used bots, or any kind of malicious automation to ensure posts reach the front-page.

2. There is no transparency for bans, nor is the ban every communicated to the owner of the HN account. If you think an account is unfairly trying to manipulate the voting system; you have to contact them. Anything less is blatantly ambiguous and unacceptable.

3. Given the silent nature of bans on HN, we have no choice but to move onto a new HN account and continue to submit legitimate posts using legitimate word of mouth dissemination. Perhaps you view this as "sockpuppets," but what other choice do we have?

4. In the particular case of Nodejitsu, this behavior is even more unacceptable. We are a platform-as-a-service where any user can create a publicly available application on a subdomain of nodejitsu.com. How can you ban the entire *.nodejitsu.com domain when we are not responsible for malicious actions (real or perceived) taken by a given user? This is not an admission that any such behavior has taken place; I'm sure that your claim is addressed in [1]-[3].

In the copyright world, site owners are protected by malicious actions of users through DMCA safe-harbor. In such senarios, the owner of the site will be contacted and the malicious behavior (real or perceived) can be addressed and/or removed. By silently banning posts and not communicating with members of the HN community about what you perceive as malicious behavior you are doing yourself a serious detriment. This has (and continues to) raise questions about the legitimacy of the content on HN as just a shill for YCombinator.

If you'd like to speak with me personally you can email charlie@nodejitsu.com

Charlie Robbins Founder and CEO Nodejitsu Inc.

You could have just emailed him and asked.

You seem to be having difficulty with the concept that what you know not to be sockpuppetry would, to the part-time admins of this site, seem like sockpuppetry.

HN makes banning mistakes with some regularity. Get over it or go somewhere else. It's ungracious to spit venom ("unacceptable", "blatantly ambiguous and unacceptable", "HN as just a shill for YCombinator") at the site admins, all of whom have better things to do than resolve individual disputes like this.

I'm as cynical as anyone about YC, 'pg, and the valley culture, but I can see when someone's doing me a favor. Like, every f'ing day for one thousand four hundred fifty two days. I too would run a site like HN differently, but that is irrelevant, because I didn't build a site like HN and so my opinion doesn't matter.

I didn't realize users' stuff was hosted on subdomains. Ok, I'll unban nodejitsu.com.

We don't ban sites lightly. We only do it when people make repeated, deliberate efforts to bypass lighter weight protections. Human mods can't watch every upvote 24x7, so if a site seems really determined to hack their way onto the frontpage, the only thing we can do is ban the domain.

I'm happy to unban nodejitsu.com if you promise to stop trying to game HN. In your case I recommend the following standard for what counts as gaming HN: if you're not sure, don't.

Problem is, Paul. that YC companies do this exact same behavior on a regular basis. I've gotten emails from founders to ask for word of mouth upvotes on specific blog posts, etc..

There is an unofficial "Upvote YC company articles" rule among YC founders. So, that's why YC company news tends to dominate HN. And, why so many amazing non-YC companies tend to not get profiled here.

It creates a echo chamber, which I think is detrimental to the entire startup ecosystem. It also gives an incorrect assumption to a segment of geeks that YC companies dominate startup the startup scene in Silicon Valley. In reality, YC companies are a very small percentage of the $20B of Venture Capital that is invested in Silicon Valley every year.

No they don't. The software doesn't distinguish between posts about YC companies and other posts. The algorithm for which votes to count, the threshold for a voting ring, and the penalty applied if one is detected, is the same for every post on HN.

HN has in fact been less and less focused on YC cos as time has passed. The initial core of users were YC founders, and anything to do with YC was interesting to them. But over time that group has been diluted by large numbers of new arrivals. A lot of people now think of "Y Combinator" as a news site rather than an investment firm, which was certainly not the case a few years ago.

I wasn't suggesting that the HN software creates an "upvote YC company" rule.

I'm suggesting that it's common for 500+ YC founders who are and have been active HN members, to quickly upvote articles about other YC companies.

I'm suggesting that this is a cultural artifact of what's happening on HN, not an algorithmic one.

And, I'm suggesting that this cultural artifact changes the content on HN to be quite focused on YC companies. It's natural. It's reasonable. And, certainly beneficial to the investment aspect of YC. You run a media site about startups, and you invest in lots of startups. Those startups naturally want their investor to do well, and their brand to succeed.

I'm not accusing you of malice. What I'm saying, is that I think that this cultural artifact benefits YC and it's startups to the detriment of the startup community at large.

It seems like the software isn't very good at determining the difference between a "voting ring", and a small-ish subcommunity that's interested in a specific technology niche.

For myself, as a node developer, articles about rails, python, erlang, and iOS apps aren't so interesting. I don't upvote them. Articles about node, politics, system programming, http, and JavaScript are interesting, so I do. As it turns out, a bunch of my friends are the people who write node-related articles, and often post them on Hacker News, and then tell me about it.

I'd be surprised if I wasn't one of these alleged puppets. Of course I'm more likely to vote up something that someone mentions to me in IM, because I have to see it before clicking the button.

I can see how this looks like sock puppets. You've got a few dozen users that always upvote the same articles from a few dozen other users, and are much less likely to upvote other articles. I'm not sure the best way to know that we're all humans. If there were fewer people using/excited about/writing about ruby on rails, then the behavior would appear very similar, I suspect.

The term "gaming" is really vague. It's not at all clear how "this is something interesting that I'd like to share and discuss" is much different from "this is something that I'd like to get people to look at." It seems to me like the ultimate purpose of a site like this is to be gamed in a particular way by a particular population.

Great point. It'd be interesting to be transparent about rings of people who frequently upvote on the same post. I'm sure they have that info. It might actually be interesting and open up some level of transparency -- whether among the YC companies (and the ability of their job posts to somehow make it to the top) or among other subsets of the coding community like node.js I have nothing against YC companies; I'm friends with lots. I have nothing against promoting a link you post on HN.

Just would be interesting to be more transparent.

Sorry pg, I didn't downvote you, but you're wrong (or you didn't read the comment you were replying to)

Here is what iamelgringo said "Problem is, Paul. that YC companies do this exact same behavior on a regular basis. I've gotten emails from founders to ask for word of mouth upvotes on specific blog posts, etc.."

iamelgringo is just one of thousands of people who receives requests to upvote HN posts. I have received several requests myself (from people who are YC founders as well as people who aren't YC founders).

I'd be happy to forward some of those (mass-email blasts) emails to you, if you think that YC founders and other HN members don't game HN the same way that the nodejitsu guys tried to do)

What the Nodejitsu guys did was more extreme than just asking a few friends to upvote your stuff. The software catches and ignores most of that automatically.

I thought we were past this kind of slander.

* Can you enumerate what you claim we did?

* Can you provide log files illustrating your arguments?

If the answer to either of the above is "no", then please keep your false claims to yourself. I'm happy to view this as water under the bridge, but I will not allow lies about my company, Nodejitsu, to circulate: we did not (and would never) use any kind of malicious, automated, or anti-competitive tactics to falsely generate upvotes for anything on *.nodejitsu.com

Again, if you'd like to speak with me personally I am available at charlie@nodejitsu.com.

Here's one of many:


Apparently this is Marak Squires using multiple accounts (barracks and changelog) in the same comment thread. I learned about this when the guys who run the changelog wrote to me to complain about it.

How is this more extreme than asking friends to upvote stories? It is simply posting from a different account. I have 15 email accounts for various projects and companies; does that mean I'm a spammer?

Posting from a different account and up voting from multiple accounts are two completely different issues. We have never used fake accounts to artificially upvote anything. Period. Nothing provided suggests otherwise. On the subject of data, if the above is an attempt to justify the original decision it is insufficient.

* How were you able to infer that both accounts were in-fact represented by a single individual?

* When one considers it objectively, two accounts with the same IP Address posting comments to HN on the same thread from different accounts is not necessarily malicious. It simply could be two different computers on the same home or work network. Point in fact, Marak and I were roommates during the early days of Nodejitsu. How would you be able to infer the difference?

Full logs of all relevant messages / comments with dates and IP Addresses included are really the only thing that will be satisfactory, feel free to email me: charlie@nodejitsu.com.

I still do not see a pattern of malicious behavior or any malicious behavior at all. As I mentioned before, I am very happy to make this water under the bridge, but please stop making false claims in public without providing more than conjecture.

I used this example because I think in most people's opinion using sockpuppet accounts in a comment thread is even worse than using them to upvote stories. It's not uncommon for people to try creating sockpuppet accounts to upvote their posts, but rare for people to resort to this.

I was able to infer the two accounts were the same person because, as I said, the changelog guys emailed me and told me so. I notice you don't deny it, which seems both revealing and revealingly disingenuous.

(Incidentally, if anyone is reading this far, this sort of thing is why it sucks to run a forum.)

> "two accounts with the same IP Address posting comments to HN on the same thread from different accounts is not necessarily malicious"

Of course not. My wife and I both post to HN, and nobody has ever criticized us for it.

It helps that we give full disclosure. If we're both involved in the same thread, whoever came in second will usually say "so there's no confusion, I'm married to [the other of us]".

That seems to be one factor some of your colleagues are missing.

> "It is simply posting from a different account."

The HN community has no problem with those who post from separate accounts for "personal" and "professional" stuff. Anonymous threads asking for legal, moral, or health advice are commonplace. Using your personal account to comment on political stories and your business account to comment on code stories is cool.

But it definitely bothers me when someone uses multiple accounts in the same discussion. pg suggests this is a pattern, not a one-time occurrence. Whether or not it's intended to be malicious or manipulative, posting to the same thread from multiple accounts definitely comes off that way. Similarly, using multiple accounts to circumvent bans so you can give your posts publicity appears sketchy.

Maybe pg, the moderators, or the algorithm made mistakes. There are better ways to respond than making multiple accounts, and there are better ways to "make this water under the bridge" than the path you've chosen in this thread.

Does that constitute an army of sockpuppets? I would disagree. I have two gmail accounts. One is personal, one is professional. Also, our posts we're censored long before hook.io got its second wind. Perhaps we should chalk this up to a misunderstanding and move forward.


I like silent bans. They seem like the appropriate response to real spam. If you give spammers a clear indication that they are blocked then they can immediately start trying a different attack. By removing that indicator then the spammers even know they need to start trying something new. It makes the whole process muddy for them and generally slows down the whole spam "cold war arms race".

It does suck more to be hit with a false positive since you might not notice it right away. But I think overall the system is good. It speaks to my inner passive aggressive. :-)

Sorry, no. I led AOL's mail team in the '90s, and we had the same philosophy. Many, many complaints about lost e-mail, but it was Necessary And Right to block spam, our #1 complaint. (In fact, we once got sued for trying to deliver the bounces back to the spammer.)

We were wrong. It didn't really block much spam; from the very beginning, spammers had "quality control" bots checking delivery rates of their spam, and they'd adapt to new algorithms within hours, demonstrating just how well they could monitor their black hole rate.

Early this decade, the technology to do efficient blocking "at the edge" was finally implemented, and boom - no more lost mail.

Silent bans are really just a way to exert power. They achieve nothing.

AOL is a big enough target that it's worthwhile for spammers to write custom code. HN isn't, fortunately, so the technique works well here.

Good point.

While I appreciate your insight and experience, you might want to view http://news.ycombinator.com/newest with and without "showdead" turned on for your account: http://news.ycombinator.com/user?id=jaylevitt

Frequently over half the submissions are dead spam, presumably killed by the policy in question. I rarely see false positives. While the quality of the remainder isn't always great, it's a lot better than it would be if there were no controls in place.

ps. Flag some spam while you're there :)

They are not a good way to stop spammers, the bot can easily make a comment and check if its there when it is not logged in. But it is a good way to annoy HN users when they get banned.

I agree. I liken it to "security through obscurity". I imagine it would be quite upsetting to be posting as a reasonable community member, and suddenly discover through some external means that my account had been silently black-holed, without being notified that my behavior was violating some of the community norms before-hand.

But perhaps such measures are needed to defend against the non-bot trolls and other belligerent individuals, and only implemented after numerous attempts at corrective feedback have failed.

"only implemented after numerous attempts at corrective feedback have failed."

That's not how it works. You say something that offends some random moderator, and he/she can trivially silent-ban you. No comment under your comment saying you will be banned etc.

That is indeed unfortunate then. :( Thanks for the clarification.

As to your 1.: Let's assume you're being honest and it isn't sock puppetry. It's still basically astroturfing, and I, at least, don't want to see that here.

If entire communities are allowed to flock to Hacker News to upvote the latest post about their pet platform, the site will become nothing more than a cheap marketing tool.

So, PSA: Please don't strip-mine this community's goodwill for your own or your company's benefit.

As to your 2.: You're entirely correct.

"If entire communities are allowed to flock to Hacker News to upvote the latest post about their pet platform, the site will become nothing more than a cheap marketing tool."

Because there's been none of that happening for the past year...

I guess you missed the month of erlang posts that were posted because community was asked to post and upvote them?... How is that different exactly?

What? You don't think articles are being systematically upvoted by groups of people? Sorry, I am a realist and it definitely seems like they're being upvoted.

HN is now the tech news site for insiders. It's not some obscure backwater.

What an odd series of replies. Surely you all realize that:

1) Whether astroturfing does happen has no bearing on whether it should happen,


2) Examples of people astroturfing are an argument for having the rule, not an argument against it.

Right? Bueller?

"... Given the silent nature of bans on HN, we have no choice but to move onto a new HN account and continue to submit legitimate posts using legitimate word of mouth dissemination. ... This has (and continues to) raise questions about the legitimacy of the content on HN as just a shill for YCombinator."

Charlie the above scenario should have triggered alarm bells. I always thought of HN as both an experiment for Arc & free publicity for YC. What's the controversy?

If you think an account is unfairly trying to manipulate the voting system; you have to contact them. Anything less is blatantly ambiguous and unacceptable.

I completely disagree that it's unacceptable. I prefer it that way. It makes it harder to game the system, and I think that keeps the quality here better. The current system is biased towards the community as a whole. You want it biased towards the organizations that benefit from visibility. I don't see any reason to swap those biases.

Personally, I find it marginally acceptable to tweet a HN submission instead of the blog post itself. Tweeting the HN submission is an implicit "Hey guys, please vote this up for us!" to everyone who follows you.

Wait, does gaming mean a thriving community of hackers spreading faster than gossip? Ahh, caught. Detective Graham ... To the zzZZZZzz

The problem with that is you are also penalizing Nodejitsu customers (like myself) that host their projects on their platform. I support the Nodejitsu guys but I'm not an employee and don't share all their viewpoints; I was just working on my first Nodejs project for NodeKnockout and wanted to get some feedback from the HN community, but my site was blocked also.

A fair argument, but I don't see how excluding content without transparency is any better. At the very least you could let people know why or what is happening when something is submitted. Not all communities need to overlap and it's HN's decision for what goes here. Also, you'd be better off if your guidelines included some wording on this.

As for the competitors line, I was possibly confused by past drama certain threads that there was something going on with that. Sorry for any misinformation this may have spread. It might be to everyone's advantage to ensure that there is a proper way to communicate with HN to get someone unbanned (I am not with Nodejitsu but I'm sure they'd be willing to talk). There is good content on that blog.

I don't see how excluding content without transparency is any better.

You can see everything that gets killed by turning on showdead in your profile.

Ok, that statement seems a little misleading. I just turned on showdead. Yes, I can see the text, but the link doesn't seem to be there, so I can't convince myself that the kill was warranted. Am I missing something?

E.g., "[dead] Andrew "Weev" Auernheimer, hacker in AT&T iPad case, on Occupy Wall Street (youtube.com)" http://news.ycombinator.com/item?id=3146776

Edit: another killed link, "Funding for venture capital industry plunges" from the mercury news. http://news.ycombinator.com/item?id=3146633 The article seems pretty relevant to the audience here.

I make the links inactive to ensure spammers get no traffic, and thus give up spamming us. HN gets so much traffic that otherwise they might get a significant amount just from people clicking on the dead link. Anyone who wants to see the url of a story can see it on the edit page, e.g. https://news.ycombinator.com/edit?id=3146776

As for your specific examples, the first was killed because that user is banned from submitting, and the second I killed myself because when I tried the link it was broken; it's since been fixed so I unkilled it.

YouTube links are auto-deaded. Write a short blog post about the YouTube video, post that, and see if it dies. I'm guessing it won't.

I don't know if the Mercury News is auto-deaded, but I wouldn't be shocked if it was.

Actually neither site is banned.

I think it's more about informing the person submitting. Most people who hear about this are surprised and had no idea. It might be solving the problem of transparency from the wrong end.

When we kill something it's almost always for some form of abuse: usually spamming, or in a handful of cases for particularly egregious sockpuppet upvotes. I don't feel like we owe such people much. Plus the more you tell them, the easier you make it to get around you.

I'll let it rest here as I'm sure you've thought much longer than I have but I do feel it a bit unfair to assume it's always the spammers submitting content when the domain is shared.


The poster is completely unaware however. See http://news.ycombinator.org/item?id=3146243 - I posted the link to check whether it's actually banned. It showed up and I could see it. Now it's dead without any notification or feedback for me.

I don't think pg's trying to hide anything. It's presumably just a lot of work to maintain an up to date public index of blocked domains and rationale.

Simply killing submissions and relying on a lesser known profile switch on the users end is pretty indirect if not obscure. The submitter should be able to lobby to get something unblocked and to do that they need to know what happened in the first place. It's a pretty simple change to the process.

>[Nodejitsu.com] is banned because they created an army of sockpuppets to vote up their posts.

Some evidence to support this claim would be useful.

How would it be useful? Are you planning on making your sockpuppets more realistic?

I'm disappointed that HN has become such a big news site, because it is not designed responsibly to be one, so trying to get a post featured is very difficult. I think the only reason any posts do get featured is because of the 'sockpuppets' that the submitters have built up (or that follow the content), but that is legitimate upvoting (I'm not sure why someone would upvote something they aren't interested in?).

Perhaps the most important change that HN needs to make is to have topics. It isn't fair to other people and content if the stream is full of NodeJS content, but that is a legitimate interest so HN can't ban the prominent workers in that field (bizarre). The ban from HN seems to be purely a side effect of the fact that the maintainers are not interested in designing the system responsibly (they are too busy with other responsibilities). Topic channels are obviously needed.

I think the difference between HN and reddit needs to be considered well and the two should be integrated some how, I would like to see topic channels on HN.

> They're banned because they created an army of sockpuppets to vote up their posts.

I too am interested in some sort of substantiation for this claim.

+1 Show us some proof, don't just throw (up) accusations around.

I for one made at least 30 accounts to request this evidence be presented forthwith

Repost from a dead link that seems relevant. I know nothing about this situation, and don't know Marak, but it seemed best to let someone defend their company:

marakmarak 5 hours ago | link [dead]

> They're banned because they created an army of sockpuppets to vote up their posts.

That is inaccurate, and libelous.

We have never used fake accounts to artificially upvote anything.

Your users and moderators harassed me to the point where I would be forced to create a new anonymous account every-time I wanted to post materials, so I wouldn't have to deal with the onslaught of your abusive community.

I was very candid about this, tweeted about it, talked about it in IRC, it wasn't a secret.

These accounts were all used once, then instantly abandoned. You have full logs of all of this.

I think you've been fed some bad information, because if you checked your logs you will find what I'm saying is accurate.

Show me one instance where our "army of socketpuppets" upvoted something from nodejitsu.com. You can't, because it never happened

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact