I’ve noticed over the past week or so that the spammers have a new trick up their sleeves. Within the last week I’ve gotten invites to iLike.com, information pills IMVU, glaucoma and myYearbook from people I don’t know, viagra sent to an email address that I don’t really use (it’s forwarded to the same place like all the rest, but I don’t give it to anyone).
I’ve had to fight spammers quite a bit on LyricWiki.org, and I’m beginning to realize a little bit more about why things work the way they work. As far as I can tell, the state-of-the-art in spamming is that tech-criminals build up bot-nets and then sell them as spamming machines. They use the zombies to attack popular technology in ways that uses other people’s web-servers to send out spam. This way, they can use the reputation of these servers to assure higher delivery-rates and they can count on the people running the servers to try to keep their reputation w/spam-filters as high as possible.
For a little more background for the uninformed: a bot-net is a vast array of hacked computers (zombies) that can be controlled remotely. Basically these are just everyday people who have been infected and are none-the-wiser. Years ago when your computer got infected, you generally got viruses that caused a ton of popups and eventually you sought help to remove the viruses. But with today’s bot-nets, the infected user generally has no knowledge of the problem and therefore doesn’t clean off their computer. When the bot-herder (who runs the bot-net) wants to do something, they use Trojan Horses which they’ve installed on the computer to send updates with what the computer should silently do.
For instance, I run MediaWiki on LyricWiki.org, and many bots have been trained to vandalize pages with random letters (I’m assuming it’s random… it might actually be a tracking-code) which they later come back and check for. If the wiki is not well-patrolled, then they come back and spam these pages with links. This way, they don’t have to reveal what product they are promoting unless they know it is some small wiki potentially with low resources – this prevents them from being tracked down by huge companies and reported to authorities. An added bonus of the bot-net approach is that each computer has a different IP address, so it’s hard to block all of them.
In this new flavor of spam, it appears bot-nets are signing up for profiles at social networking sites, and sending out invites to victims. This is a great way to use other sites’ reputable servers to send out spam that is highly likely to get delivered and also to make it through contextual spam filtering (since they look like any other invite).
This creates an interesting conflict for the sites who are being used to send the spam: on the one hand, these bots are out promoting them for free, getting new users to sign up out of curiosity (“Do I know this person? The name sounds vaguely familiar…”). On the other hand, these are ill-gotten users, and the spam that’s being sent out probably moves their servers on to more and more blacklists. Both options are a mixed-bag, and in the end I feel that it’s always best in business to do the right thing without immolating yourself. You didn’t earn these new users, so just take a stand and try to solve the spamming issue if you can. Aye, there’s the rub: often, a startup’s most rare asset is time. How much time should a company devote to trying to fix a problem like this? They could be out promoting their site, adding new features, or fixing bugs. They’re always understaffed, and there is always more work to be done.
This is a hard problem to deal with since you’re either protecting strangers from a bunch of spam that’s coming from your servers (which you really had nothing to do with), or you’re adding features for your users. A tough call to make. Hopefully some of these companies can co-operate to come up with a technical solution that they can share amongst each other to make it practicable for them all to implement it. The three companies whose servers spammed me aren’t even direct competitors – one is chat (IMVU), one is youth social-networking (myYearbook), and one is music-focused (iLike).
I’ve emailed a friend at one of the companies and explained the situation. It will be interesting to see how they respond.
PS: Please don’t comment about just adding a CAPTCHA. Those things are horribly useless against talented programmers and have an inherent “economic” flaw. I’ll probably write more about it later, but to put it simply, every time I see a site using “ReCAPTCHA” in a place where they should have actual decent Turing-test security, I cringe. It doesn’t do that!