I’ll moon you if you’re not careful

To be honest, I probably won’t, unless it is by accident: I appear to have lost a small amount of weight recently and my trousers keep falling down. Between that and a belt that is held on by one wafer thin thread, there’s a reasonable chance that I will show off my three piece suite to commuters at least once this week so I’d better be sure to check my underwear for holes each morning. Mrs Cobras has reminded me on several occasions that I need to a) go into the loft and fetch my old “slim Cobra” clothing and b) buy a new belt. I have ignored both these suggestions and prefer to live dangerously.

Yes, yes, let's all ignore the snake's footprints TOGETHER

On the subject of moons, this article of mine gets more spam than any other article on this blog. And not by a small margin, either, by a vast margin. In fact, before I started adding all the IP addresses to the block list, I was pressing the spam button on 30 or 40 a day. All to the same post. What is it about moons? There does not appear to be anything odd about that post that makes it more attractive to spammers, but clearly, the artificial “intelligence” that drives the spambots are attracted to it like The Cobras to red wine. You know what they say, Em-Oh-Oh-En, that spells artificial intelligence¹!

Fortunately for me, I moderate everything by hand and I’m one of those human being things so I can tell that this:

“I like the valuable dirt you prepare for in your articles. I force bookmark your weblog and control again here frequently. I am unequivocally certain I will learn varied up to date shit right here! Righteousness fluke in place of the next!”

… is about as likely to be a real person as the delicious home-made biscuits I’m enjoying with coffee. Although, let’s be fair, it deserves at least one gold star for “varied up-to-date-shit” and “righteousness fluke”.

The worst thing is that this is a good one and yet it is still utterly appalling. The spammers only have to get past the automated filters of most blogs and judging by what I see on the Internet those filters are hardly reinforced concrete doors, they’re more like bead curtains. Some automated filters are not bad, but they work off fuzzy recognition based on a database of examples and thus tend to play catch-up rather than being pro-active: when the spammers find some genius new sentence structure that evades them, they’re cooking for a few days. It’s an arms race with no sign of finish in sight. Manual filtering, of course, catches ‘em all, because automated software is utterly crap at context. Furthermore the english is, at best, questionable so it’s dead easy for a real person to recognise them. Even bad english spoken by real humans is recognisable as human by humans.

Whilst these things may be rubbish at context and believability now, they won’t always be. There are some substantial developments taking place in the field of AI right now and it is inevitable that one day, machine intelligence will equal or surpass our own. Then spamming will become really interesting: most of the content of the Internet will become artificial agents all trying to have relevant, sensible conversations with each other in the vague hope that they’ll get a chance to slap their link somewhere where it’ll get clicked on. I am sure it’ll make for cracking content, but what will be real? What won’t be? What is real at that point?

Meh, it’s Monday. This is all too complicated. I drew a moon. With a snake on it. It’s great.
–
¹ One gold star if you can get the reference 🙂