The first of today’s set of posts with allusions to song titles in the header. This is offered largely without comment, except to note that – regrettably – there’s no data here for the number of failed Kickstarter projects per category. Found via This isn’t Happiness.
Compsec maven Brian Krebs rakes over the findings of University of California, San Diego research report into the online market for what I like to call meatpuppets: cheap human labour-on-the-web that gets leveraged for bypassing the security systems that are supposed to stop automated spammers.
“The availability of this on-demand, for-hire contract market to do just about anything you can think of means it’s very easy for people to innovate around new scams,” said Stefan Savage, a UCSD computer science professor and co-author of the study.
The UCSD team examined almost seven years worth of data from freelancer.com, a popular marketplace for those looking for work. They found that 65-70 percent of the 84,000+ jobs offered for bidding during that time appeared to be for legitimate work such online content creation and Web programming. The remainder centered around four classes of what they termed “dirty” jobs, such as account registration and verification, social network linking (buying friends and followers), search engine optimization, and ad posting and bulk mailing.
“Though not widely appreciated, today there are vibrant markets for such abuse-oriented services,’” the researchers wrote. “In a matter of minutes, one can buy a thousand phone-verified Gmail accounts for $300, or a thousand Facebook ‘friends’ for $26 – all provided using extensive manual labor.”
The evolving marketplace is best illustrated by the market for services that mass-solve CAPTCHAs — those agglomerations of squiggly numbers and letters that webmail providers and forums frequently require users to input before approving new accounts. The researchers found that the market for CAPTCHA-solving was fostered on freelancer, but quickly expanded into custom markets when the model proved profitable on a large scale. Today, there are plenty of commercial services that pay pennies per day to low-wage workers in India and Eastern Europe to solve these puzzles for people wanting to create huge numbers of accounts at one time.
It’s interesting to see massive crowds of human labour getting rolled quite effectively into these vast and largely automated systems: the darkside equivalent of Amazon’s Mechanical Turk, with a smattering of Matrix metaphors on the side. But those digital peons are just trying to make a living, and when you look at the prices being charged for Twitter followers by the thousand and factor in the significant cut being taken by the service aggregators, you realise that they’re probably not making much more than sweatshop wages. Which means that until the massive differential in income between developed and developing nations gets narrower, web security procedures will always be subject to this sort of outsourced brute-forcing. Shorter version: spam ain’t going anywhere anytime soon.
The irony of having blocked five fake Twitter accounts in the time it took me to write this post is palpable. Death, taxes, noise*, spam.
[ * Anyone who’s worked in the recording or music industries will tell you that noise is the third certainty of life. As, I suspect, will anyone who has lived in a block of flats. ]
But it makes for an attention-grabbing skiffy-tastic headline, AMIRITEZ? The actual story here is rather less OMFG: University of Pennsylvania have obtained the first neurobiological evidence in support of the theory of episodic memory.
“Theories of episodic memory suggest that when I remember an event, I retrieve its earlier context and make it part of my present context,” Kahana said. “When I remember my grandmother, for example, I pull back all sorts of associations of a different time and place in my life; I’m also remembering living in Detroit and her Hungarian cooking. It’s like mental time travel. I jump back in time to the past, but I’m still grounded in the present.”
Jumping back in time to perceptions of the past while still grounded in the present? Strikes me that rewatching old home movies is at least as good a metaphor as time travel, but I’ll grant you that a lot less people would have reported it if it were pitched that way.
Neuroscience is still a fairly new scientific frontier, and while the last decade has seen the arrival of amazing new tools (and enhancements of existing ones), I believe it’s fair to say that these methods are still pretty crude, and the interpretations of results somewhat speculative. But even so, it’s interesting to see these early phases of our attempts to measure something as inherently intangible as the mind:
The memory experiment consisted of patients memorizing lists of 15 unrelated words. After seeing a list of the words in sequence, the subjects were distracted by doing simple arithmetic problems. They were then asked to recall as many words as they could in any order. Their implanted electrodes measured their brain activity at each step, and each subject read and recalled dozens of lists to ensure reliable data.
“By examining the patterns of brain activity recorded from the implanted electrodes,” Manning said, “we can measure when the brain’s activity is similar to a previously recorded pattern. When a patient recalls a word, their brain activity is similar to when they studied the same word. In addition, the patterns at recall contained traces of other words that were studied prior to the recalled word.”
“What seems to be happening is that when patients recall a word, they bring back not only the thoughts associated with the word itself but also remnants of thoughts associated with other words they studied nearby in time,” he said.
The findings provide a brain-based explanation of a memory phenomenon that people experience every day.
“This is why two friends you met at different points in your life can become linked in your memory,” Kahana said. “Along your autobiographical timeline, contextual associations will exist at every time scale, from experiences that take place over the course of years to experiences that take place over the course of minutes, like studying words on a list.”
I should probably go and register that domain now, shouldn’t I? If you were gonna make an algorithm for checking that transhuman-or-not status of the species, though, you might wanna refer to Kyle Munkittrick’s transhuman checklist, which consists of the following points:
Munkittrick suggests that “[i]ndividually, each of these conditions are necessary but not sufficient for transhumanism to have been attained”; that jars slightly with my own comprehension of the term, which has always assumed that the “trans” in transhuman implies the transitional phase on the route to becoming posthuman, which is what I’d say we’d be once Kyle’s checklist is complete. Semantic carping aside, however, it’s a solid and non-sensationalist list, well worth a read.
Speaking of non-sensationalist pieces on transhumanism, here’s an unusually subdued post from Michael Anissimov which is either indicative of a massive change of outlook or a rhetorical gambit that has yet to be revealed as such: Why “Transhumanism” is Unnecessary. Having been following Anissimov for many years now, I suspect the latter is the case, but hey – this is the internet, and all bets are off.
Via the dashing and debonair Ryan Oakley, researchers at MIT have managed to get a computer to do what most computer users never do, namely Read The Frackin’ Manual. And guess what – the computer’s performance at the task at hand improved hugely! The task in question was… playing Civilisation.
But the task isn’t the point, you see; this is about teaching machines to comprehend input in a linguistic fashion:
The MIT Computer Science and Artificial Intelligence lab has a computer that now plays Civilization all by itself — and it wins nearly 80% of the time. Those are better stats than most of us could brag about, but the real win here is the fact that instruction manuals don’t explain how to win a game, just how to play it.
The results may be game-oriented, but the real purpose for the experiment was to get a computer to do more than process words as data — and to actually process them as language. In this case, the computer read instructions on how to play a rather complex game, then proceeded to not only play that game, but to play it very well.
If you take the same process and replace gaming with something more real-world applicable, like medicine or automotive tech, you could have a computer that’s able to act as more than just a reference tool. A lot more.
If I’m grokking it right, this is the opposite of the approach embodied by IBM’s Watson, which is essentially a search engine on steroids; I’m reminded again of the Chomsky/Norvig debate, and MIT’s approach here looks to be much more in the Chomsky direction. I suspect some sort of synthesis of the two approaches will bring the best results in the long run.