Tuesday, August 22, 2006

Googlezon

If I'd written the post below in 2003 and turned it into a really good video, it might look something like this:

Googlezon

Thursday, August 17, 2006

Vox Machina

Everyday, actually two or three times a day, particularly savvy bits of spam seep through the multiple filters between the bad guys in the Ukraine and my eyeballs. I use the Ukraine here as a proxy for all of the spam making machines in the world. My mind is but a self-reflexive series of heuristics; maybe I'll tell you more about this particular proxy in another post.

Anywho, these messages get past the filters on both the stanfordalumni servers and the gmail servers. Judging from the sheer number of messages which land directly in my spam folder and never cross my retinas, these filters generally do a pretty good job. However, particularly slippery missives seem to slide through even the tightest of machine nets, though the spamtastic patterns are easy enough for a human to catch.

First, the sender's name is always a standout. Yesterday I received messages from Bella Broussard, Connie Dougherty and Christie DuBouis. My real world mail correspondents tend not to adopt the nomenclature of burlesque stage stars. The other obvious identifiers of these messages are the wildly poetic text and the inevitable attachment which I dare not open.

For example, Bella Broussard sent the following, under the subject heading "Methane Publicity":

"He backed them out of their stalls and began harnessing. Don't put it off too long or I may not be there. Said ‘twas goin’ to blow the lighthouse out to sea, or somethin’ like that. His mental speculations were engaged with matters much more personal and intimate."

Google tells me that the sentences are mismatched pieces of text lifted from Joseph C. Lincoln's 1924 novel Rugged Water. The novel is publicly available on the Internet. I assume that someone, somewhere in the Ukraine no doubt, created a 'bot to scour the web for long passages of text which match a certain format--say dialogue or lots of commas or something and then wrote a script to tie these strings together into passages that are random and poetic enough to be indistinguishable from a legitimate human-to-human email. Attach and advert for a 419 scam or worse yet, attach some malware, and the bad guys are off to the races. Some fraction of fools will respond and the spammers make enough money to cover the cost of sending the messages and a handy profit to boot. But this has been going on for ages. What I find interesting about vox machina, the voice of the machine, is that machine-generated text is everywhere, and it’s sounding more and more human. It’s all over the ‘net, intermingled with earnest blogs, corporate marketing messages, online petitions and truly hideous Myspace pages. With time, as the programmers get smarter and the costs of producing spam, link farms, and new and as yet unimagined types of machinespeak fall lower and lower, it will be completely indistinguishable from human language.

Here's an example. Today I was doing a little research on the Yorùbá god, Changó. Again, my mind is but a self-reflexive series of heuristics. I might get back to my fascination with the Orishas another day. Anywho, I went to my favorite source of all things, Wikipedia, where the encyclopedia entries are compiled and edited by the people, for the people, in a sort of open-source epistemology of the bored and altruistic. As I was reading the Changó entry, which was of course cobbled together by a gang of at least superficially unknown who evers from where ever, I came across this phrase:

"In further apatkis Shango goes in search of Aganju, his father..."

Apatkis? What's that? I typed "define: apatkis" into Google, my other favorite source of all things, and it returned no results. Nothing. Impossible! So I kicked out to a wider search of simply "apatkis" which returned a scant 210 results. I'm morally opposed to browsing past the first page of any Google search, but the results returned on that first page were telling enough. Three of them included the phrase "In further apatkis Shango goes in search of Aganju, his father..." and the other seven spewed the wild poetics of the machine with phrases like "a 350 wedding place cards clocked and operated by its apatkis newspapers" and "Massiel reached its apatkis of 90 mph over the franco-manitoban Gran Pacific."

So here’s the trouble:

I still don’t know what “apatkis” means and my primary source for all the information in the world, outside of the infinite loop of my own mind, is both the source of the question and the answer. Moreover, both the question and the answer may in fact be the nonsensical rant of the machine or the first few words of the dawn of a new breed of intelligence. I can’t tell the difference, is Wikipedia vox populi or vox machina? Four thousand years from now when archeologists are digging through the detritus of the present era, will they be able to discern Moby Dick from the machine? Maybe the machine will be the one doing the digging!