blogospheroid comments on Stupid Questions December 2014 - Less Wrong

16 Post author: Gondolinian 08 December 2014 03:39PM

You are viewing a comment permalink. View the original post to see all comments and the full post content.

Comments (341)

You are viewing a single comment's thread. Show more comments above.

Comment author: Lumifer 16 December 2014 01:34:54AM *  1 point [-]

You seem to want to build a massive sledgehammer-wielding mech to solve the problem of fruit flies on a banana.

So the attacker expends a not inconsiderable amount of effort to build his sockpuppet army and achieves sky-high karma on a forum. And..? It's not like you can sell karma or even gain respect for your posts from other than newbies. What would be the point?

Not to mention that there is a lot of empirical evidence out there -- formal reputation systems on forums go back at least as far as early Slashdot and y'know? they kinda work. They don't achieve anything spectacular, but they also tend not have massive failure modes. Once the sockpuppet general gains the attention of an admin or at least a moderator, his army is useless.

You want to write a library which will attempt to identify sockpuppets through some kind of multifactor analysis? Sure, that would be a nice thing to have -- as long as it's reasonable about things. One of the problems with automated defense mechanisms is that they can be often used as DOS tools if the admin is not careful.

If a large enough percentage of outside user content is "bad"

That still actually is the case for Twitter and FB.