Hello, “word press”. I have to wonder, what is it that you squeeze to get words to come out? Do you put Gideon’s Bible under a mill-wheel and encourage Roseanne Barr to walk on top of it, to “get it rolling”, like a Flintstone? Do you auger Little, Brown under ratcheted thumbscrews in the hopes of draining some little essence of normal language from its tortured creases?
Bullshit! You don’t do those things to those things. When you want words, there is only one thing to squeeze. Two things, actually.
test : late 14c., “small vessel used in assaying precious metals,” from Old French test, from Latin testum “earthen pot,” related to testa “piece of burned clay, earthen pot, shell” (see tete).
testify : late Middle English: from Latin testificari, from testis ‘a witness.’
Some etymologists have claimed that the legend of Romans squeezing each others’ balls in order to contend veritability is just that: urban legend. But here we have the words “test” and “testify” both descending from words that share roots with the testicles and the teats.
Imagine if the cultural expectation were that two women who want to be sure they are speaking the truth to each other had to squeeze one another’s breasts. How hard? Probably not the pleasurable sort of gentle kneading you’re imagining right now. Probably more like a mammogram exam. Which brings us full round back to the whole “test” aspect, and that’s pretty quaint, isn’t it? Knock. Knock. Who is it? Mammogram.
So if you want to get words out of somebody, obviously you crush their testicles and their breasts with force sufficient to turn gold blocks into dust — in other words, about as much force as a Romanian squeezing a ketchup bottle over a pizza.
So into the great word-vat they go, hogsheads of testicles and whole palettes of breasts, stacked up high and smashed down low, earthen vessels careening into a pot and words flying out willy-nilly. And what words come out, today?
Alan Turing developed a test in 1950 for determining the intelligence quotient of a machine, as a function of the people giving the machine the test. If the output of a machine could be programmed to fool 30% of humans into believing it’s the output of a human being, the machine wins and gets to be declared “intelligent”.
Well, no machine ever passed the test. Somebody always figured out that the machine was just a machine. Until very recently.
A machine (which you can test yourself by visiting this link: http://www.princetonai.com/bot/bot.jsp which will lead you to the Artificial Intelligence program at Princeton) made the news recently when it successfully fooled some third or more of its human testers into believing it was itself a human being that they were chatting with.
The beauty of the Turing Test is that the human testers are chatting with a mix of human and robotic test subjects. The human tester has no idea going in whether they are about to chat with a real human or a machine. But as I mentioned, it’s usually (well, always) figured out soon enough. The machine doesn’t have to resemble its human cohorts for everybody, though — just that questionable 30%.
It should be 100%, because at least 50% of the world will believe anything you tell them as long as you lie to them consistently, according to Abraham Lincoln. He really said that, it’s true!
And these days, it seems like 100 IQ is perhaps functionally average in terms of being right smack between drooling on your lap and piloting the international space station, but it’s definitely not average in terms of being the median intelligence of most of the population. I’m sure if we calibrated, we’d find that new median intelligence to be roughly a standard deviation below average, or 85 points.
At any rate, I had to test this robot out myself. And it lasted a whole three questions.
Me: How are you, today?
Bot: I’m doing great, and you?
Me: Doing well, thank you.
Each time I sent a question, there was a wait time while the creature devised what insidious phrase would trick me most handily, gigabytes of simulated neurons squirting electrons under layers of thermal paste. It was so exciting.
I can’t remember what I asked next, but I remember the third question:
Me: Are you good at maths?
To which the robot responded, after taking exactly zero wait-time to pretend to type it all out:
My grandfather taught me not to disclose non-solicited information about myself. Like on the question “Do you have watch” — never tell time, but answer “Yes” or “No”. I always follow his advice. If I’m not mistaken – you still didn’t tell me where you live. OR it’s a secret?:-)
I was to believe that any human being renowned to possess a very fast typematic rate is even possibly capable of forming and typing that entire paragraph within basically half a second. Let alone a 13 year old boy from Ukraine, information I didn’t really have considering I didn’t get far enough in questioning him to receive all that information. Just “Hi, I’m having a good day, and I type 4,800 words per minute.”
Guardian NBC CBS Independent missed out on here was a perfect opportunity to rail on about how basically fucking stupid humanity is becoming over time. A full third of people were fooled by this bullshit? Somebody else pointed out that it gives simple math answers (to “what’s 2+2”, for example,) in decimal notation.
What other kinds of answers could I expect were I to overlook the pre-fabricated paragraph?
Me: where do you live
I asked that question and waited for several minutes before reloading the page. But all I got was this stupid “page not available” message.
The robot never had a chance to respond, because the
Guardian Independent article led so many people to use it that the system was seized-up under the load. The tragedy of the commons strikes again.
Funny to think that none of the people trying to connect to the robot are going to be able to actually get it to work under such conditions, but that they will keep trying and their efforts will contribute to to their own lack of success.
No doubt, most of the
Guardian Independent readers trying to view the robot will be completely fooled and refuse to believe it’s a programmed script, even though they are served the advantage of knowing fully well ahead of time that it indeed is. The robot wins, again!
While the article did eventually show up on Slashdot, as of the time of the failure of the script to show up, I checked and the article had not yet been posted. So it wasn’t Slashdot’s fault — the robot did not get “Slashdotted”.
The suggested “Wake-Up Call to Cybercrime” makes exceptional sense in the light of so many people being so obviously, enormously stupid. A recent article on the vulnerabilities in Hbb televisions (“Red Button Attack”) mentions that authorities warned about the vulnerability felt that it was “too costly” to fix.
Another article notes that electric highway signs — hackable a la the new game “Watch Dogs” — are vulnerable mainly because they allow a Telnet (instead of the more secure SSH) connection and because they are still set with the default passwords, or with no passwords at all.
With security that relaxed, you don’t have to be a genius (or even to hold an average level of intelligence) in order to be a world-class “hacker”. And considering how blatantly stupid you can be but still finding yourself in a position of authority, it’s no wonder people just take the path of least resistance, and dribble into their lap down the rich path of wealth, status, and POWER!