X
Innovation

You may say I'm a dreamer...

Up at 4am today in something of a tizz, having had one of those nightmarish dreams that seems real on the point of waking but dissolves away as you regain yourself. But this was a good one – it had me for minutes.
Written by Rupert Goodwins, Contributor

Up at 4am today in something of a tizz, having had one of those nightmarish dreams that seems real on the point of waking but dissolves away as you regain yourself. But this was a good one – it had me for minutes. In going over it, I slowly realised that most, if not all, of the raw material is actually happening and the way the dream fitted it all together had a bizarre but undeniable logic.

What's more, I think it may turn true.

In my dream, one of my slightly neurotic friends had come to me with a worrying story. Someone was stalking her on the Internet, she said, and sending her scary stuff. There were pictures of her from her past she didn't know existed. Others of road junctions in the US where one road had her name, the other of someone she'd only half-remembered meeting. Sets of apparently random text, but that included in the slurry of novel extracts and synthetic English some real details of her life.

The story was mostly worrying because it made me begin to question her sanity. In real life, I've seen one person I knew begin to slip into paranoia: it's that old hippy question, how do you tell a friend that their brain is fried? Back in the dream, I was going through the same fears, but I agreed to go and see some of the material on her computer at first hand.

I went, and looked, and it really was there. But you can fake anything on a computer, right? My friend was in a mess – she thought I suspected her sanity, and was beginning to wonder herself. The files were convincing, though bu I deliberately didn't tell her everything I recognised – what if she thought I was behind it all? But the emails and attachments came from all over, there was no way to trace the headers. If it was a stalker, they'd gone to the trouble of setting up a network of zombie PCs controlled by bots – but is that so much trouble, these days? The tools are all on the web. But if it was her doing this to herself in some fugue state – it was a class job. She wasn't that technical. Was she?

So many questions.

The dream cut forward. I was back at home, by myself, late at night, going through her evidence, looking for clues. I checked my email – and there was an image of me from years ago, sitting on someone's sofa, wearing a shirt that had long since become moth fodder. Then came more, clips and snippets of my life woven into the same bizarre, surreal mix. Some of it even mentioned my fearful friend.

I don't know how you'd react in real life if this happened, but the dream Rupert got badly spooked. Whose sanity was shaken now? So I did what she'd done. I went to my most technically savvy, stone-cold sane pal and poured it onto his shoulders.

I did this while we were walking down a London street on a grey, wet afternoon. We were going past one of those big glass-fronted pubs full of screeching yahs as he began to sketch out some ideas, when he nudged me and pointed. There, on a huge flat-screen TV inside the bankers' watering hole, was a picture of me....

And then I woke up at 4am, thinking "what the...?". As you would.

It's now 5am as I write this, and the modern miracle of caffeine has restored my phlegm. I've worked out what my dream was telling me, and why it matters.

The zombie nets and spambots are out there, of course, vast ad-hoc networks of compromised computers used to flood the Net with unblockable emails, denial of service attacks and so on. This, we know. But that's a remarkably simple-minded use of a very powerful resource.

Elsewhere, scientists are using nearly identical networks of computers - this time, with their users' knowledge and permission – as supercomputers, doing insane feats of data mining and pattern matching. Seti@home was just the first; I'm folding proteins via BOINC even as we speak.

But the one thing the spammers really need is to get past filters – of all sorts – and plant their message in the heads of their victims. Now, the Internet is a remarkable compendium of individual information – again, this we know. Most famously, Google threatened not to speak to us for a year after we'd published information about its CEO and his family that we'd, er, Googled from the Net.

Put all this together, and you have the next stage in spam technology. Zombie computers that mine their host computers for personal information and swap it between themselves until one has enough to synthesise a message that looks utterly compelling to its target. Others that use facial recognition techniques, snapshots swiped from hard disks and Google image search to come up with a best-fit name-to-face match – again, to build compelling spam, perhaps by including local pictures they've found of the remote target.

Then they can start to play on relationships that they know exist by looking at email conversations and comparing them across computers – hence that street corner in some small US town, accessible by location search tools. And of course, those emails contain lots of personal details that can be woven into the spam text body.

None of the above has to work 100 percent. Far from it. As soon as it works well enough – and with spammers, the economics of production is so insanely efficient that they can afford to make products that are five-nines faulty – it becomes worth doing.

I started on the second cup, and stared down the barrel of a Dickian dystopia coming true. The spammers didn't even have to make the thing work at all – and already they'd spammed me in my dreams. How many filters did that get past?

Be sure to let me know if my brain's fried.

(Oh, and the TV picture in the pub? I had to be up at 4:30 anyway to broadcast to this and other nations on what's going on with Galileo, the European satellite navigation network. That's another really interesting story, this time alarmingly true, and one I'll get onto after cup number three.)

Editorial standards