ASCII by Jason Scott

Jason Scott's Weblog

A Human Moment in Spam (Update: Nope, Still Machine) —

On the post I did about Floppy Disks, that’s gotten a lot of attention in its own right, there was a comment the weblog software marked as suspicious, and possible spam.  The content itself was rather innocuous and in theme:

I will always keep a few floppies and a 3.5 inch drive. Can’t install older versions of XP or 2003 without them to load custom drivers at setup.

I also have the 5.25 inch drive and disks (just in case).

In 25 years some government secret on a floppy will need to be read and I’ll have the last drive and hardware to read it. Then I’ll be rich. Mark my words.

But yes, it was definitely spam – the website attached to the user, their e-mail address, and the name of the user were all references to a specific fruit that appears to be the new thing garnering the need for attention and weblink juice to get its message of “please fucking buy me” out to as maximum an audience as possible. It is, without question, spam.

There’s a range of commentary I make where I am probably not saying something new to a lot of people, and I get a lot of response of “well duh”, but I still feel the need to mention it, for historical context later.

Spam, as you know, is the idea of unsolicited commercial speech being thrust upon audiences without their consent, assent, knowledge or awareness. I’m going to leave it at that general definition for the purposes of what I’m talking about.  Originally a reference to endless messages on Usenet (and emulating the Monty Python Spam Sketch), it expanded out to e-mail as e-mail became more prevalent. From e-mail, where it got the most hand-wringing and programmatic work to divest it away, it has followed, at various paces, to every single platform in which any amount of people congregate. Spam has cropped up in twitter, facebook, online games, IRC chat, Flickr, wikis, and places infinite.

Someone solved the spam problem a while ago: go after the tiny handful of banks that the spammers take payment from.  But, knowing how to solve something and solving it are two different plateaus on a very large landscape with dangers and silliness fraught between.

So instead, let me focus on what I’m observing here, in this specific situation, and why I think it needs to be made into a historical note.

A person made this. For me.

It was weird to have machines scripting spam in the 1990s beyond the standard “and take this pre-written pitch and shove it in as many newsgroups/e-mail addresses as possible”. You wouldn’t have it, say, look at the writing of a newsgroup, re-arrange the last major topics talked about, plagarize items from a year or two previous, and compose a “new” message that turned out to be links to spam. If Usenet had survived into the later 2000s, it would have. And that’s what I’ve been seeing – scripts that knock on my doors with composed text taken from the weblog itself. So there’s been advancement there.

But this… I checked. This is a valid comment. They’re saying something reasonable and clear, something that would be a perfectly fine comment about Floppy disks, their perspective to them, and what they might have said if they were some pleasing retro-nerd happy I was talking about 5.25″ floppies. It was composed for real, for this weblog posting.

This is an enormous human cost, compared to just scripting once and blowing it out across a thousand million weblogs.  It means, and I knew it was coming to this, that hiring humans to do things so bone dull has become enough of a commodity that spammers are employing this to get their messages injected into the conversation. If my software didn’t show me the spammy e-mail and website attached to the posting, I’d have approved it.

I don’t like it at all.  I don’t know where this is going to lead. But it’s worth noting, for the future.

Update: It turns out this was a spam script, which took one of the comments on a Reddit link to my entry and then refashioned it into my commenting structure. Brilliance. So, we’re still in machine land. I am actually rather relieved. There are some things human beings just shouldn’t be doing.

Categorised as: computer history

Comments are disabled on this post


  1. The software on Caltrops is completely custom. There isn’t a copy of it running anywhere else. A gentleman named Stew wrote it almost ten years ago. One day, several years ago, some spam appeared. Just your usual links for hardschlonger. HOT. I changed the name of the PHP file and that kept the spammer away for a few months. But then he changed HIS script.

    There’s no logins or captcha on Caltrops, so for a few years there was this quiet cat and mouse between myself and this dude who was getting some submarginal benefit to reworking his URLs that were gonna display to, at most, 300 people, albeit amazingly handsome ones.

    He hasn’t posted in the longest time. Sometimes I wonder what became of him. I wonder if he gave up. If he fired off a crooked smile to his PC one evening and just let me win. Or… if he’s actually still out there in the ether, looking for single-instance installs of custom code to hit. But then I remember that all spammers eventually get green syphilis and die from it, so there you go.

  2. Anon says:

    The text of the comment came for a human, but it wasn’t written for you… It was a spambot that borrows text taken verbatim from the comments from a Digg submission that linked to your post. Gotta give the assholes credit, they’re getting clever.

    Of course, someone *could’ve* used Mechanical Turk or Fivrr or paid a homeless guy to write something original, and that certainly happens for high-profile and/or difficult spam targets, but that costs real money.

  3. Adam Bradley says:

    Speaking of MTurk, I recall reading a while ago that a sizable percentage of HITs are for spam or similarly-questionable tasks (e.g. “Make a Twitter account and follow this idiot who paid me $50 for 1000 followers”). I imagine for a few cents you could get someone to write 5 or so relevant comments and link them to your cheap Canadian pharmacy or whatever, but that’s probably too expensive for comment spam–I don’t expect these guys have a high enough hit rate to justify even that cost.

    Using Digg as a source of spam comments is clever, in a vile sort of way.

  4. Aside from Anon’s discovery of the true origin, consider also that this could be a virus of some sort that attaches these URLs to posts made by a legitimate user. That would be pretty clever, too.