0810: "Constructive"

[Ghona]

Unfortunately, all the spammers have to do is outnumber the humans and upvote themselves. And many spammers possess millions of IP addresses, thereby guaranteeing themselves positive votes. Any vote-based system is susceptible to this sort of manipulation.

Which assumes that there exists an automated way for the spambots to ID other spambots.

At which point you take that automated system and use it to ID spambots.

[HonoreDB]

Before there are any real comments, the captcha is populated by one or two sample constructive comments, created by the blogger, and a bunch of markov-chain manglings of those comments.

[Poochy]

Problem: What if the YouTube scenario occurs and people start using sockpuppets to repeatedly vote comments up or down based on whether they agree or disagree with the opinion of the poster?

Remember, in the minds of about 1/3rd (in my experience) of the American population, a <any positive adjective> comment is one that expresses an opinion they agree with, and a <any negative adjective> comment is one they disagree with.

Conclusion: People are a problem.

[quietus]

So, change the wording, which of these comments are useless / spam.

[melladh]

Almost as bad a tease as shibboleet

[BlitzGirl]

But what about all the people who won't be able to join the community because they're terrible at making helpful and constructive comments?

They are subhuman and should not be allowed to communicate with people like Randall. Easy.

These individuals would fall into the same category as "those who do not know Shibboleet" and so would be doomed to a never-ending phone call with level one tech support. They just aren't as smart as the rest of us (and Randall). We are the Elite.

[waldir]

A good tease indeed, but unfortunately wouldn't work:

* Obvious flaw #1: Spam-bots approve spam-bots.
* Obvious flaw #2: Human (a) argues with human (b). human (a) declines genuine constructive comments of human (b) because they disagree with them/ out of spite.

it also gets rid of people without the patience to wait for a second user to confirm that they are real.

There's also:

just because the bots are capable of making automated constructive and helpful comments, doesn’t mean they’ll be used for that once they pass the entrance exam.

but that might be good in a way: they'd still have enhanced current Turing test results

[waldir]

By the way, there seems to be a few misled comments here regarding reCaptcha:

all the spammers have to do is outnumber the humans and upvote themselves.

ReCaptch is vulnerable to this as well.

well, no it's not. From moot wins, Time Inc. loses:

Luis von Ahn, the project lead of reCAPTCHA goes on to say: «about the “penis attack”. We serve over 400 million CAPTCHAs per week, so submitting 200k CAPTCHAS with the word penis doesn’t even come close to poisoning our database — we serve each word to multiple random users, and we require them to be correct on the other word, so to get any traction with this attack, they would have had to submit at least 100 times more CAPTCHAs. And even if they did this, we have many other measures against it. That attack simply doesn’t work.»

besides, as phlip said, as they integrate the ORC-unreadable words, the "penis attack" becomes... well, moot.

[Klear]

Unfortunately, all the spammers have to do is outnumber the humans and upvote themselves. And many spammers possess millions of IP addresses, thereby guaranteeing themselves positive votes. Any vote-based system is susceptible to this sort of manipulation.

Which assumes that there exists an automated way for the spambots to ID other spambots.

At which point you take that automated system and use it to ID spambots.

The spambots need only to approve everybody else.

Also, this comic is brilliant.

[BAReFOOt]

I’m sorry, but unfortunately, the idea in the comic can’t work.
Since spammers will also be the ones rating those comments. And of course they will rate their own up. But even if not, they will be the vast majority.
So in the end, spam will be rated as highly constructive, and that’s the end of your little idea.

But 1. don’t be sad, cause 2. I already solved that problem.
The problem is, that you are making the same wrong assumption, that Wikipedia and every other forum on the net makes: You assume that there is one global view on the world. One global right and wrong. One global reality.
I’m sorry. There isn’t.

Even if the state of two human brains is exactly the same, they at least still stand at different positions at the same time, or at the same position in different times. Which means that from a physical standpoint, reality is already relative. Now of course, as far as we know, physical rules are the same for both of them. But those rules already state, that even time is relative to your speed. But speed itself is only defined, relative to another object.
And all that doesn’t even matter much. Since human brains never are quite the same. What we call e.g. “helpful and constructive”, is entirely relative to our inner model of reality. It has to be compatible with that one, and it has to be information which makes it easier for us to survive, grow, etc. (= win natural selection). But with limited resources, one person’s winning is the other person’s losing. One person’s “helpful” is the other person’s “hurtful”. And really everything that does not fit in one’s own, very very personal, model of reality, is considered “not constructive”.

“Luckily”, most people are “cattle” for most topics. They blindly believe and follow the hive mind “reality”. Never questioning it. Never even thinking that you could question it. Never building their own sense of “right” and “wrong”. Which means their “right” is actually someone else’s “right”. And “right” just means “is useful to help me win natural selection”. So they usually protect something that hurts them.

Now to answer the core question: What kind of system could make it impossible for spammers to disturb anyone? (Bonus question: And how can we kill off the forced “one true reality”?)
Well, nature already had a perfectly good system, before all those new forms of communication appeared. It’s called “network of trust”! You trust your friends and real leaders. They trust their friends and real leaders. And so on. Because you learned to know them and they knew you, and you worked in a team, because that made it much likelier for you to win in natural selection.
And nobody trusts a spammer. At least not after the first spamming. His bad reputation would always hurry ahead of him. Which means that nobody would want to spam others. Because he would have a very hard time finding anyone to even talk to after that.
Also in that network, reality is relative. There is no global “truth”. Truth is, what you and your trusted ones deem to be true. And that’s it.

You just have to implement this on the Internet. Which means, that the rating for each and every comment and in fact the whole view on reality, has to be by either you, someone you trust, or someone they trust, and so on. Just like for every element in HTML, every style information for it is defined, by cascading down from the most specific one to the least specific one.

So for forum spam, you could do three things:
2. Create comments with defined topics,
3. rate comments in specific topics, and
2. add/remove people you trust for specific topics.

And since those people would themselves have people they trust, and themselves add ratings on topics you trust them to be competent, down to the big leaders that create and aggregate a lot of ratings and that a lot of people trust, there would be a very personal view for everyone, without having to rate every single comment yourself. (In fact you could just leave all rating to others, if you trusted them.)

Yes that means you have to get to know other people, to be able to trust them. Sorry if you don’t like that, but that’s a core element of being human. If you start to lose natural selection (or in this case get lots of bullshit and spam to read), you will quickly appreciate its usefulness.

But imagine how a spammer could possibly infiltrate such a system. It would be impossible.
He would have to get lots and lots of people to trust him. And then send out the spam.
But that would mean that people would instantly remove him from their network of trust.
Also they would become a lot more wary about adding random people to their network. Which means that just creating a new user name and trying to get people to trust him again, would not only be a lot of work again, but also be a lot harder, if not impossible.

And those who would blindly believe a hive mind that is also a source of spam, and not remove him… those would deserve what they get. They would be those who lose natural selection. And that would be a good thing. And they would not have the right to complain. Since they could just learn to distrust spammers. Or be more wary in general. Or find some real friends. Duh. If they can’t do that, then nobody’s gonna be sorry for them. Or maybe someone will sacrifice himself, to protect them… if he thinks they are valuable to him. Just like in RL.

I think it’s very nice and elegant. Which maybe is, because humans did it since the dawn of times, and it got perfected in all that time.

[Steroid]

Problem: What if the YouTube scenario occurs and people start using sockpuppets to repeatedly vote comments up or down based on whether they agree or disagree with the opinion of the poster?

Remember, in the minds of about 1/3rd (in my experience) of the American population, a <any positive adjective> comment is one that expresses an opinion they agree with, and a <any negative adjective> comment is one they disagree with.

Conclusion: People are a problem.

Perhaps it is not people who are a problem, but the system which refuses to cater to people.

Rather than try to eliminate spammers, phishers, and trolls, a better mission would be to establish a system where the people who just want to have a conversation are required to respond to spam; where their bank accounts are left wide open to phishers; where the less reason put into a comment, the more insightful and wise it is treated. A system where all shall take and none shall give. A system where we not only have our cake and eat it too, but eat it twice, have it, give it away, cut it up to decorate another cake, use it to start a bakery, and keep it around to impress the cake connoisseurs. The paradise of the future is not here; it is 4chan.

[MadLogician]

Great comic. Funny, and thinking about whether it could work is thought-provoking.

'accomplishing the mission' would be the equivalent of creating strong AI, so I doubt it can be done.

[NixonsGhost]

Isn't this what the Gawker sites already do?

[SirMustapha]

It's a shame that Randall only has these "brilliant" ideas that only work in his cartoon land.

If he were still at NASA, probably someday he would be able to say "MISSION FUCKING ACCOMPLISHED" for real, which is far more rewarding.

[gordysc]

This reminds me of http://img513.imageshack.us/img513/1952 ... teruq1.jpg for some reason

[Monika]

But what about all the people who won't be able to join the community because they're terrible at making helpful and constructive comments?

Well they will have to stay outside.

The real problem is the sociopaths will still get in. You know, those who do not have emotions, but understand them, can fake them and manipulate them in other people.

(Sorry, watched too much House MD last week, there was this episode with the sociopath.)

[monicaclaire]

Ha! Spam? An unavoidance of the internet. What we need is people who use the internet to be a little more OBJECTIVE! Of course, this is coming from someone who sometimes feels like she is "giving in" when buying household items or even food.

I approve of this idea. Bonus points if they have to prove they can type/spell correctly

Oh, wouldn't that be nice. On an unrelated note to the comic, but related to your post. I recieved a Math project where in the first page it specified that the project needed to be typed and spelled correctly. It then went on to the second page of the project description, and no futher than the second line, it proceeded to spell passenger passanger. :/

[dougw]

Am I the only one who thinks the third panel doesn't scan very well? Here's how I would edit it: "But what will you do when spammers train their bots to automatically make constructive and helpful comments?"

[Monika]

Spellchequing is overated.

[ysth]

Before there are any real comments, the captcha is populated by one or two sample constructive comments, created by the blogger, and a bunch of markov-chain manglings of those comments.

I wonder if anyone but me thought of "I'm in Marsport Without Hilda"

[myrcutio]

If the spambots result in modding their own coded bretheren "constructive" to such a degree that it passes the bar for accepted commentary, perhaps we shouldn't fight it. It's just the language of the future, and we should learn to speak it. I for one welcome our 14 inch enlarged penis botting overlords with a dancing smiley and free iPad.

[Zylon]

Um, guys? The fundamental premise of today's strip is wrong. AIs aren't cracking CAPTCHAs these days... people are.

[Pterosaur]

Oh wait, time for a funny outcome of this situation. To catch the checks if a user is posting constructive comments, more spambots are created to confirm that other spambots are being constructive, with more spambots to confirm those spambots, creating a Ponzi scheme-like tower of spambots on every website until the entire blogosphere implodes under the DDoS of CAPTCHAs

Perhaps this will result in the creation of an emergent intelligence, a kind of hive-mind. Hopefully it will mutate and evolve into a more benign and intelligent form.

Skynet?

[xnick]

It would have been funny to need to type a captcha before reading this particular comic

Is anybody else afraid to comment here and get negative votes, now?

--
Now, the flaws mentioned earlier in this thread would be solved if the comment rating were automated by some super-bot, and not dependant on users...

[Jatopian]

Korea has a system in place in website with a lot of traffic that makes anyone who puts down a comment to confirm their identity with an ID number which they get from the Korea Communications Commission before the comment is allowed to be posted.

When I heard about this, I was very surprised we weren't just talking about the North. Guess we should just cut that peninsula off and let it sink into the ocean. It's a lost cause.

[skeltoac]

This system uses feedback to develop a sense of trust in those whose comments are approved by others in the community. However, its sense of trust also depends on one factor that has not been noticed by those who believe the spammers can win by outnumbering the humans:

There is at least one human in the system who is a trusted authority.

The founder can delegate this to any trusted entity--if they choose a spammer then they deserve what they get--but this is the root authority that imbues all other community members with the right to assess spam. Without input from this authority, the system would never get off the ground.

Also, you can't just vote on other comments; you must leave a comment. The value of your assessment derives from the assessed value of your comment. Thus spammers' assessments would never count for much.

There is another feedback loop that must be accounted for: complaints from the users. When the authority learns from the users that the spam system sucks, either something gets done about it or the community devolves.

Spammers don't like it when they ruin a community, however, because it weakens the host on which they rely parasitically.

It's a good system if your users are willing and able to perform the required assessments.

[hatten]

It would have been funny to need to type a captcha before reading this particular comic

Is anybody else afraid to comment here and get negative votes, now?

--
Now, the flaws mentioned earlier in this thread would be solved if the comment rating were automated by some super-bot, and not dependant on users...

-1

[qukuss]

On top of the afore mentioned flaws with the idea, this would be even more annoying then recaptcha. Since there are only 2 choices you would have to ask at least 5 comments to make the spam protection even mildly resistant to brute force attempts. I would be far too annoying to read an article and rate 5 comments.

[neoliminal]

So that's what M.F.A. stands for.

[kanraga]

It's a shame that Randall only has these "brilliant" ideas that only work in his cartoon land.

If he were still at NASA, probably someday he would be able to say "MISSION FUCKING ACCOMPLISHED" for real, which is far more rewarding.

Troll harder, please. Your work of recent is too desperate and not up to your old standards of bitingly ironic disapproval.

Or is this the way you've always been?

[neoliminal]

It's a shame that Randall only has these "brilliant" ideas that only work in his cartoon land.

If he were still at NASA, probably someday he would be able to say "MISSION FUCKING ACCOMPLISHED" for real, which is far more rewarding.

Troll harder, please. Your work of recent is too desperate and not up to your old standards of bitingly ironic disapproval.

Or is this the way you've always been?

+1

[chrth]

It's a shame that Randall only has these "brilliant" ideas that only work in his cartoon land.

If he were still at NASA, probably someday he would be able to say "MISSION FUCKING ACCOMPLISHED" for real, which is far more rewarding.

Troll harder, please. Your work of recent is too desperate and not up to your old standards of bitingly ironic disapproval.

Or is this the way you've always been?

Now now, be nice. Maybe someday someone will actually click one of the links in his (her?) signature making it all worthwhile for him (her?).

[jc]

... just supposing the mission does accomplished, then just because the bots are capable of making automated constructive and helpful comments, doesn’t mean they’ll be used for that once they pass the entrance exam.

Heh. I've been working on a recognition problem related to this. I have a web site that has been hammered by bots (from around 1400 addresses so far) that try to use my web form to insert URLs and advertising material into my data. I've always done a bit of sanity testing on the input, but after a few successes on their part, I decided to investigate more. One thing I found was that in about half the cases, an attempt to insert spam was immediately preceded by a completely valid-looking request from the same address, which my code handled normally. In about 1/4 of the cases, the first successful attempt to call one of my CGI programs was immediately followed by a burst of 3 or more similar calls with the form params full of URLs and/or ad phrases (mostly pharm spam).

I long ago gave up on the CAPTCHA approach, because the available packages all see to fail fairly quickly, and I haven't learned enough to know how to do it myself. So instead, I have a blacklist of known spam-injection sites, plus a lot more checking of the input.

The most curious aspect of this is that most of the attacks include the string "2rand[0,1,1]", which gets around 300,000 hits on google right now. This includes a fair number of questions of the "WTF is this?" sort, but so far I haven't found any explanation for why someone would be using that peculiar string. It doesn't do anything at all in any language I know, and that seems true for a lot of other people, too. But it leads me to use the phrase "2rand spam injection" to describe that subset of the problem. There are sites all over that have this string inserted in otherwise normal-looking text, but nobody seems to know what the motive is. There are a few hits for "2rand" that are discussions of South African currency, but I don't think it has anything to do with that.

[bfollinprm]

It would work, as long as the probability P(H) of a given user in the forum being A useful human is given by P(H) \ge (K-\epsilon K+\epsilon)/(1+K), where K is the mean ratio of the rate of spam bots rating comments to the rate of human ratings, and \epsilon is the percentage of humans whose comment rating opinion approximates that of a spam bot.

As k \rightarrow \infty, the probability must approach 1. In other words, it would never work.

[jc]

Korea has a system in place in website with a lot of traffic that makes anyone who puts down a comment to confirm their identity with an ID number which they get from the Korea Communications Commission before the comment is allowed to be posted.

When I heard about this, I was very surprised we weren't just talking about the North. Guess we should just cut that peninsula off and let it sink into the ocean. It's a lost cause.

Nah; they'll discover soon enough that all the spammers have to do is send a (fairly cheap) bribe that gets them a file of all the valid ID numbers, from an unnamed source inside the appropriate government department. This will be quickly followed by terabytes of spam, all from "valid" Korean sources, which will be quickly followed by the rest of the world blacklisting all Korean addresses.

Actually, the real motive is probably an easy way for the Korean government to learn who is making each post to any forum. Fighting spam is merely an excuse to take aim at the general population for politically unacceptable words. Sorta like here in the US, invoking terrorists or child molesters is the standard excuse for total monitoring of the US population. And in other countries, a popular local bogeyman is used for the same purpose.

Any "security" measure that can be defeated by a simple bribe to a low-paid agency worker should always be assumed to be not aimed at actual security (whatever that may mean), but as an excuse to monitor the general population for politically unacceptable words and thoughts. If you apply this inference, a lot of "security theater" becomes obvious and understandable.

[SirMustapha]

Troll harder, please. Your work of recent is too desperate and not up to your old standards of bitingly ironic disapproval.

You're lying.

If by "Troll harder" you mean riticise the comic itself (because someone who dares speak negatively of LORD RANDALL must obviously be trolling), well, that is a tough one because the comic is a non-joke -- it's a half-arsed idea that would earn, at best, a throwaway Twitter post from an actually talented person. Really, the only thing that makes this comic stand-out somehow is the "MISSION. FUCKING. ACCOMPLISHED." (the FUCKING. is very, very important), and it seems so much like Randall trying to make himself look smarter than he is that it's pretty pathetic. If the comic actually ended the way the alt-text ended ("... oh!"), it would seem less like self-fellatio and more like an attempt at comedy, but would show even more just how lame the whole thing is.

One thing I'll give to Randall, though: the comic is so jarringly anti-White-Knighting that it almost makes it seem like Randall is listening.

[Nyerguds]

Unfortunately, all the spammers have to do is outnumber the humans and upvote themselves. And many spammers possess millions of IP addresses, thereby guaranteeing themselves positive votes. Any vote-based system is susceptible to this sort of manipulation.

Yeah, I realized that pretty quickly too... just having bots up-voting all posts destroys the system.

[djkjr]

Am I the only one that read this and thought "He is attacking the people on the fora who criticize his recent work constantly!" ?

Your comment, lost but not forgotten.
No, you're not the only one.
I often wonder why individuals criticize the writer of a webcomic, especially in the forum for that comic. It's not the first time I've seen it.
I often wonder, also, with all their great criticism’s, what have they created that they can be so critical?

[kanraga]

You're lying.

If by "Troll harder" you mean riticise the comic itself (because someone who dares speak negatively of LORD RANDALL must obviously be trolling), well, that is a tough one because the comic is a non-joke -- it's a half-arsed idea that would earn, at best, a throwaway Twitter post from an actually talented person. Really, the only thing that makes this comic stand-out somehow is the "MISSION. FUCKING. ACCOMPLISHED." (the FUCKING. is very, very important), and it seems so much like Randall trying to make himself look smarter than he is that it's pretty pathetic. If the comic actually ended the way the alt-text ended ("... oh!"), it would seem less like self-fellatio and more like an attempt at comedy, but would show even more just how lame the whole thing is.

No, bro. By "Troll harder", I mean "Be a more original troll! Write something that can't be so easily deconstructed into the following!"
"
1. This idea sucks.
2. Randall is a white-knighting douche.
3. Randall is unintelligent/overidealistic/boring/pretentious/(have I missed anything?)
4. By virtue of 1. and 2. or any of 3., Randall sucks.
5. Insert optional further ad hominem.
6. Also, anyone who disagrees is an xkcd fanboi.
"

Vary it up a bit, you know? I mean, even Randall has his fun comics once in a while. I recommend not spreading yourself so thin - don't feel obligated to comment on every comic that you have the slightest objection to. Save your shitstorming abilities for the ones you think are truly balls-to-the-wall-horrific. It's more refreshing that way, you'll be more effective, and people might actually listen!

[asdfzxc]

I seem to have a different idea of how this works than you guys:

• Guy A tries to comment.
• Guy A is presented a handful of randomly selected posts.
• Guy A disagrees with Guy B, who makes about 1% of the site's total posts.
• Guy A has a small chance of getting a post by Guy B.
• Guy A downvotes any Guy B posts he gets on general principle.
• Much like Recaptcha, some posts are confirmed to be "good" or "bad". About half the posts given are "confirmed".
• If Guy A correctly identifies the "confirmed" posts as good or bad, the comment passes. Otherwise, reset.
• If the post passes it's added to the list of possible comments.

• Spammer A tries to comment.
• Spammer A is presented with the above test.
• If the spammer somehow manages to pass, his post enters the pool. (Note: Actual spambots are stopped cold here)
• If the post is shit, it gets downvoted into oblivion.
• Important part: The odds of a sockpuppet getting the original spammer's post are ridiculously low. In order to save it from the wrath of other votes, you would basically need to DDOS the site with sockpuppets.