LowEndBox has been live since February 1, 2008.  In those 16 years, we’ve had 6,264 posts published (including this one) with a total of 129,351 approved comments.

As far as I can determine, we did not install an anti-spam plugin until September of 2016.  That was before my time working on LEB, so I don’t know if we didn’t have one before then, or if at some point the stats were reset.

I say “approved” comments because the total number of submitted comments is far larger.  We get a good amount of spam, like anyone who runs a WordPress blog.

Early on, it was little more than something like this:

Great post!  Really enjoyed it!  BTW, you can save big on hair tonic at (some url)

Then it became more sophisticated, with an attempt to make comments that “lure you in”.  For example, on our post on Copilot’s terms of service this week, we got this spam:

This is fascinating.  Microsoft says Copilot is for “Entertainment Purposes Only”?  Just look at the Microsoft Copilot terms of use.  Wow.  I wouldn’t have thought this was the case, but you lay out the evidence.  Coincidentally, I was thinking about this situation also and wrote a post at (some url)

In this case, the script has lifted parts of the title (in bold), dropping sentences under a word count threshold, and then splicing it into a block of text, Madlibs-style.

Or on this one on our post about Zypher, the homemade AI companion:

I can’t believe that GitHub has announced that going forward, your data on GitHub will be used to train Microsoft’s Copilot LLM.  That is not something I expected to see.  I wonder what the underlying motivation is?  It’s often hard to tell.  Deep thinking required!  Maybe AI can tell us the answer?  I have a theory (linked to some url).

That bold part is lifted directly out of the Excerpt for that article.

The next frontier will be something where an AI reads the article, feeds it to an LLM, and pumps out a believable comment with a link somewhere in it.  I think this is still computationally expensive for spammers.  They’re obviously not going to pay Claude Opus API fees for every blog they target.  Their “business model” is to hit thousands (if not millions) of blogs.  But with smaller self-hosted models, they could eliminate the API cost.  After all, the comment doesn’t have to be perfect.  However, I think the time to download the article text, feed it to an LLM, and get the response is still 10-15 seconds per blog and that is too long.

So…what is your guess for how many spam comments we’ve received since September 2016?  Make your guess and then vote to see the actual number.

raindog308

raindog308 is a longtime community LETizen, technical writer, and self-described techno polymath. With deep roots in the *nix world, he has a passion for systems both modern and vintage, ranging from Unix, Perl, Python, and Golang to shell scripting and mainframe-era operating systems like MVS. He’s equally comfortable with relational database systems, having spent years working with Oracle, PostgreSQL, and MySQL.

As an avid user of LowEndBox providers, raindog308 runs an empire of LEBs, from tiny boxes for VPNs, to mid-sized instances for application hosting, and heavyweight servers for data storage and complex databases. He brings both technical rigor and real-world experience to every piece he writes.

Beyond the command line, raindog308 has a life-long love of German Shepherd Dogs, high-quality knives, target shooting, theology, tabletop RPGs, playing guitar, and hiking in deep, quiet forests.

His goal with every article is to help users, from beginners to seasoned sysadmins, get more value, performance, and enjoyment out of their infrastructure.

You can find him daily in the forums at LowEndTalk under the handle @raindog308.