I checked the spam filter this morning and came across a new comment on an old post.
It’s was a reasonable comment. Not an amazing contribution to the discussion, but not completely nothing, either. And I would have approved it—except that the url supplied by the commenter was a spam link. Or, maybe not spam, just some business that had no connection to anything that ever appears on the blog. In any case, I shot the comment into oblivion. I don’t want to be hosting or encouraging spam—at least, not for free!
We get this kind of comment from time to time, and it always makes me wonder: When people do this, are they coming to the blog with an intent to spam (or, one could say, to advertise their wares) and then they write some minimal comment in the hope that it gets approved? Or are they coming to the blog to write a comment, and then they figure they might as well get some benefit out of it so they throw in the spam link? I have no idea.
Occasionally I’ve clicked on links tied to specific comments only to find that they don’t lead anywhere. Perhaps those are accidental links (entered wrong or copy/paste from some previous activity)? In any case, I guess the “anonymous” comments are the opposite of these – no intent to gain anything from the comment. What I have yet to see is an anonymous comment with an active link to a webpage – any takers?
I’ve seen (in crypto chats) a separate channel/room provided for spammers. Oddly, this arrangement seems to be respected by the spammers, so it must work for both sides.
You could try a blog once a week advertised as open to spam, and explaining your intent. Everyone else can ignore those, and perhaps you’ll get less spam on the other posts. Maybe even do some kind of analysis on the spam.
My guess is it’s the former. I’d even go so far as to conjecture that they used an LLM to compose a relevant-sounding comment, and then appended the spam link. So, probably not a great loss to block it.
Adede:
Some comments are clearly computer-generated, either from the chatbot or from some more primitive technologies. In the above post, I’m talking about comments that seem to have been hand-written.
With LLM being as good as it is, the chatbot text is gonna be more and more hard to recognize.
+1 I would actually be very surprised if this type of comment was not LLM-generated.
Years ago I started seeing a version of this which was definitely mass spam. The comment would be two-three generic sentences to the effect of “I liked your blog post”. They don’t even have a URL in the comment itself, rather they use the “Website” field which your blog (and presumably every wordpress blog) offers as part of a comment. Presumably the goal is to affect google search ranking by getting the linked website “mentioned” in many places.
I often report these and get no response from authors. I delete them from my own blog but I assume lots of people get flattered, since the comments are just generic praise. I guess with LLMs you can now produce more tailored messages for this type of spam, which is the version you encountered.
Anon:
Yeah,. those super-generic comments don’t count at all. But, also, as noted in my earlier comment, in the above post I’m not talking about comments generated by chatbots or other computer programs. Those are pretty obviously spotted. I’m talking about comments that I’m pretty sure were hand-written. It doesn’t really matter since I send them to the spam folder too; I just wonder whether the person was actually trying to contribute to the discussion. Spammers are people too, and maybe sometimes they feel they have something to say!
Haha, sorry for missing your point then. My general assumption is that there are obvious chatbots and less obvious chatbots, and maybe the less obvious ones like to hang out on stats blogs (chatbots are people too)
This post by Anonymous is an example of what I was referring to. It shows up as a link, but when I click on it there is no page to be found. What makes the blog show there is a link – what something entered in that field? If so, then was it intentional or a mistake?
You can put anything in the link field, it will show up when you hover over it.
Except not “Hi Dale!” apparently. Then either the blog or my browser must do some kind of filtering.
I prefer “Hello world”
Most browsers will show you the destination of a link when you hover over it. If you hover over the anonymous link above, you should see on the bottom left “https://example-spam-website”. Clearly this is not a valid URL – there is no domain name.
This hovering technique was very important 10-20 years ago, because browsers were not very secure. Opening a malicious website could crash your browser, hijack it, or worse. So the best practice was to always check where links go to before clicking them, or ideally never click on links at all and instead manually type in the ones for known legit websites (this avoids typosquatting). I still have this habit and it bothers me when I see my mother clicking random unknown links. Zero day exploits still exist, even though nowadays they are not used for low-effort prank links.
OK, but is a person listing “some business that had no connection to anything that ever appears on the blog” as their website necessarily spam?
I mean, I think there’s probably a continuum here. At one end are business websites like amazon.com, which are spam even if the poster happens to work Amazon. Likewise any pimps that post here probably shouldn’t link to hotgirlsinyourareanow.com (not a real website, I hope). At the other end is a purely private website like dangermouse.net. In between lie:
* university faculty/grad student websites, which are personal but necessarily include extensive discussion of one’s professional work;
* a blog or the website of a few-person company like johndcook.com, which is arguably professional but Cook is ultimately selling himself so the professional content *is* personal; and
* a much larger company has recently grown but not professionalized to match its scale, such that it gives the company webmaster a little too much freedom.
I’m not sure where (say) a LinkedIn or Wikipedia user page would fall.
Somewhere along the line, a poster’s connection to a website attenuates such that a link would be spam. But we all probably draw that boundary slightly differently. Maybe they and you just disagree about whether they’re distinct enough from the website for the link to be spamming?
I wonder if this isn’t some computer security issue. A low skill way to get your links past spam filters is to write a browser extension which you trick people into installing (either by modifying some legit extension or the usual encouragement to click) which then appends your spam links to any text fields above a certain length their browser posts. Blogs might be a happy accident with emails representing the real target.
Either that or maybe they harvest comments from related posts and sometimes it looks like it applies to the post in question.