Apres CAPTCHA, le deluge
CAPTCHAs -- the technology used to stop spammers from using computers to sign up for email accounts and the like -- are thoroughly broken; spammers and researchers are finding better and better ways to get computers to recognize the word-soup. Computerworld's Steven J. Vaughan-Nichols has a good feature on what happens now:
According to Nagle, waxing sarcastic, "Several commercial products are now available to overcome those little obstacles to bulk posting. A tool called CL Auto Posting Tool is one such product. It not only posts to Craigslist automatically, it has built-in strategies to overcome each Craigslist anti-spam mechanism." It's not the only one. There are, he added, "other desktop software products [such as] AdBomber and Ad Master. For spammers preferring a service-oriented approach, there's ItsYourPost." The result? "The defenses of Craigslist have been overrun. Some categories on Craigslist have become over 90 per cent spam. The personals sections were the first to go, then the services categories, and more recently, the job postings."Link (via /.)
the latest
latest episodes
Those captchas have gotten so funky that I can't even read them half the time.
Umm why can't cragslist hire moderators? Really how many could it possibly take to look at every ad before it's posted and approve it.
cant have that, then people would freak out and claim their freedom of speech was being denied under the brutal reign of the dictatorial oppressors :p
I think a lot of the world programming problems will be solved VERY quickly if they were placed in the way of spam.
"Some categories on Craigslist have become over 90 per cent spam"
How could you tell?
Yer Maw, that's exactly what reCAPTCHA does, instead of presenting randomly selected letter/number sequences to solve, reCAPTCHA serves words from book scanning projects that failed OCR, thus lowering the cost of book digitization and preservation at the Internet Archive. http://recaptcha.net/
To echo #1, I am a native English speaker who can see 20/20 with glasses and I sometimes have to try different captchas a dozen times to find one I can repeat correctly. Some of the case-sensitive ones use fonts that hardly differentiate between lowercase and caps! It was a useful and interesting tool, but it's had its day.
If the spammers can use AI, why can't we? Are we that far away from scripts that can equate "vi.Agra" with "Viagra?" How hard is it for a well-programmed piece of software to figure out that "MAKE $$$MONEY NOW" is not a typical headline for bulletin board postings?
Btw, #4 made me giggle at this thought:
"To post this comment, please type the cure for cancer into the text box below and click submit."
reCAPTCHA should just use the CAPTCHA bots to figure out the failed OCR text.
Ths ppl shld ll b sht. Craigslist is one of the best boards on the internet. They will ruin it. When the revolution comes these f**kers will be up against the wall with the rest of the evil capitalists. Not like the ones with a soul. The ones putting a dollar sign on every damn thing on the planet.
It's obviously pointless in the long run to try to eliminate Spam. It's the Spammers themselves we have to eliminate. With extreme prejudice.
Related to #7:
http://www.wellingtongrey.net/miscellanea/archive/2008-04-07-what-hath-captcha-wrought.html
I like #9&10
If it means I no longer have to spend five minutes staring at Rapidshare's CAPTCHAs trying to determine whether the warped sketch I'm looking at is a stylized cat or dog, I'll be happy.
Gabrieel (#8), because reCaptcha only knows the "answer" to one of the two words you're seeing so they'd have no way of knowing the captcha cracker actually got it correct. Though I suppose you could always have some sort of cracker battle royale to see if they all agree :)
I think the Asirra system is promising. http://research.microsoft.com/asirra/ It asks users to identify pictures of cats and dogs, something that is very easy for a person to identify but it is difficult for a computer to tell the difference.
Additionally you can adopt any of the animals in the photographs. The images are pulled from petfinder that finds owners for homeless pets. Seams like a win-win, easier captchas for us and fewer animals get put down.
Why are people not celebrating? I mean, one of the biggest tricks in AI was solved, does it really amtter if it was by the bad guys. Now why when I buy a US$300 scanner it comes with a crappy OCR software that can't tell a vertical line from a lowercase "l" printed black in a white paper?
Though this problem will inevitably be solved, it does illustrate the one superiority that classified ads have over Craigslist: human editors who can reject ads that violate certain rules before publication.
Any form of prior review for Craigslist would be rejected by the community, no matter how useful it would be against spam, and Craigslist would lose out to an online competitor.
Likewise, print publications would never accept many of the legitimate ads that Craigslist posts without charging more money than they're worth.
That distinction won't go anywhere soon. All Craigslist can do is escalate the arms race against spammers.