I’ve noticed hundreds of new spam comments here in the last few days…
The only spam detection I have is Akismet, since it’s built in to public WordPress. It works well though. The first few of this “new generation” spam made it to the blog comments, but only because the spammers were stupid enough to forget their links. Normally the comment writer’s name is a link to the spam website, but this time they forgot. But that was two days ago…
Now I’ve seen some more intelligently generated spam. They often reply to existing comments (which is new), or post new comments as usual. But these new spam comments aren’t only generic compliments and gibberish. For example, there was this one:
This was so helpful and easy! Do you have any articles on rehab? (Emphasis mine.)
It’s still a generic compliment, but notice the last word? That’s a keyword relevant to this blog. It was pushed to a post about atheism, not addiction, recovery or drugs. But still… This new generation spam must be using an API specific to the targeted platform, interrogating blogs and generating comments in response to existing comments, presumably so that followers can get notified that their comments have responses. They’re also using keywords relevant to the target blogs, which make the spam look almost like real comments.
And then they fuck it up by being adverts for insurance (in this case), just like the 50 others it pushed around the same time. At least there’s that… They can never avoid detection completely, unless they forget to put their ad link in the comment writer’s name. Do that and the spam becomes useless because it is no longer an advert.
This irks me though. It means that there are programmers out there, people like me but without integrity, writing code that attempts to get around spam detection algorithms. These are people who deserve to be unemployed. Granted, they’re not exactly skilled programmers and their “intelligent” spam is not as intelligent as it could be… I can think of a way of easily circumventing the detection of the spam links, and would find it trivial to generate spam “smart” enough to get past Akismet. But I won’t do that or write how I would do it because I am not an arsehole.
Edit: One day, and over a hundred (detected and removed) spam comments later, and I think I was giving the spam devs too much credit. That one comment with an apparent keyword must have been a fluke. I’m almost disappointed, because it probably wouldn’t be difficult to interrogate blogs and generate content including keywords scraped from their tag clouds. But clearly spam developers aren’t too bright after all. And that’s a good thing.